View Full Version : Speed Graphic Focal Plane Shutter
rdenney
28-May-2010, 23:50
I just finished a fairly major overhaul of the Speed Graphic that I bought this week. This one had been rode hard and put up wet.
In my complete overhaul of the focal plane shutter, I followed the procedure in the factory service manual (a link to which I found on Graflex.org) and did it all--rewound the curtain, diassembled and cleaned the retarding regulator, and so on.
During reassembly, I wound the spring 8 turns as suggested by the manual. When I tested shutter speeds, though, the slower speeds were over half a stop slow and the high speeds weren't. The curtain closed reliably from O, so I think the tension was what it was supposed to be. But I still had to add about another 8 turns to get the slow shutter speeds in range. From 125 down, the shutter is within a small fraction of a stop.
But I can't get the 250 and up even on the right planet. Given that the slit is narrower, the fast speeds should be fast if the slow speeds are correct, and they really aren't. On the unretarded side, the speeds are nominally 1000, 250, and 50. 50 is dead on, 250 is about 190. 1000 is maybe 400. On the retarded side, the speeds should be 500, 125, and 30. 30 and 125 are perfect, 500 is about 250.
Given the design of the shutter, I'm at a loss as to why the fast speeds are dragging. My ears are certainly not complaining about the sound of the shutter--it is now quite crisp. I can live with it the way it is--I have the same useful speeds as any given Copal 3 (long as I don't need slow speeds)--but I'm curious. Anyone have any ideas?
Rick "who was unable to make the camera square and has to be content with locking the front standard down with a slight swing" Denney
Bill_1856
29-May-2010, 06:11
Accurate speeds of the Graflex focal plane shutters have always been a problem.
Do you know about "Packing?" For the highest speeds, the shutter is "packed" by winding it and releasing it at the highest tension setting four or five times before taking the shot where it counts.
rdenney
29-May-2010, 06:42
Accurate speeds of the Graflex focal plane shutters have always been a problem.
Do you know about "Packing?" For the highest speeds, the shutter is "packed" by winding it and releasing it at the highest tension setting four or five times before taking the shot where it counts.
That is standad practice with older shutters, of course. But in this case, I must have cleaned it well--I so no variation in speed over half a dozen complete windings.
Rick "wondering about the little phototransistor tester at those speeds, now" Denney
BetterSense
29-May-2010, 09:31
The phototransistor should be able to read the focal plane shutter pretty well at those speeds, if you put it fairly close to the shutter curtain. I measured my shutter and found that full speed was more like 1/500. I didn't measure the rest of them in detail because I was only using the FP shutter at the highest speed then. I went ahead and assumed the rest of the speeds were a stop slow as well but based on your information maybe I should not.
BrianShaw
29-May-2010, 09:48
I trust technology (even technology, like a shutter speed tester, that I don't own), but when I had this question about my camera I put a roll film back on the camera and shot a roll of B&W at a series of pics of a blank wall at equivalent exposure and examined the film for density differences (I'm such a techno-phobe that I don't even have a densitometer; I just looked with my eyes).
hi rick
whenever i have re-speeded the shutter on
my speed i usually just have the shutter barely make it from the 1 1/2 to T ..
remembering the camera at 1/30th probably was held upside down to work against
gravity ... and then all the other speeds have more or less fallen into place.
i have done this with a series d too ... when i used a shutter tester ( at a camera shop )
it seemed to be within 1/3 of a stop ...
good luck !
john
rdenney
30-May-2010, 14:25
hi rick
whenever i have re-speeded the shutter on
my speed i usually just have the shutter barely make it from the 1 1/2 to T ..
remembering the camera at 1/30th probably was held upside down to work against
gravity ... and then all the other speeds have more or less fallen into place.
i have done this with a series d too ... when i used a shutter tester ( at a camera shop )
it seemed to be within 1/3 of a stop ...
good luck !
john
John, your description isn't helping my limited brain. What do you mean by "just barely make it", and "1 1/2"? Speed Graphics were held upside down at 1/30? That just doesn't compute in my brain.
(By the way, I just cleaned and relubricated the Graphex shutter on the lens, and it is within 10% on the slow speeds of 50 and longer. 100 is really 1/65, 200 is really 125, and 400 is about 210. That's with the transistor sensor in the middle of the shutter, but even when it was offset by a third of the diameter, it didn't change much. Easy shutters to service, too.)
Rick "who would appreciate a little more detailed description" Denney
John, your description isn't helping my limited brain. What do you mean by "just barely make it", and "1 1/2"? Speed Graphics were held upside down at 1/30? That just doesn't compute in my brain.
(By the way, I just cleaned and relubricated the Graphex shutter on the lens, and it is within 10% on the slow speeds of 50 and longer. 100 is really 1/65, 200 is really 125, and 400 is about 210. That's with the transistor sensor in the middle of the shutter, but even when it was offset by a third of the diameter, it didn't change much. Easy shutters to service, too.)
Rick "who would appreciate a little more detailed description" Denney
hi rick
sorry about that .....
not sure where i read it ( maybe graflex.org? maybe here? maybe apug? )
the camera was usually held upside down to achieve 1/30th of a second
upside down meaning so the bed is above the lens like a sun shade ...
if the camera was held like this with the shutter traveling up ( against gravity )
it would be a slightly slower speed. whether it is true or not, not sure, but it
sounded good :), the person who posted it seemed a way more knowledgeable about
graflex-stuff than me ... i've been using them and tinkering with them for
a few years, but i don't have anywhere the experience with them as a lot of other folks.
i misspoke when i said 1 1/2" ... on the series D slr the
shutter is labeled differently, the largest shutter opening / slot
in the shutter cloth is 1 1/2" and with the SLR when the shutter
barely travels from the 1 1/2 setting to the T setting (on the shutter)
everything is around in the right place with the rest of the shutter speeds ( kinda-sorta ).
i've used a shutter speed tester and most of the speeds were close enough
that i didn't worry about it ...
when i set the time on the shutter in the speed graphic i have i usually
wind it 10 revolutions and compare and exposed negative between 1/60thS with
the focal plane shutter and 1/60thS from a regular shuttered lens. or i use
a shutter speed tester. it is never "right on" for any of the speeds
but it is close enough that i don't worry about it.
when i say close enough, i mean maybe 1/30thS is 1/40th or 1/125thS is 1/160thS ...
hope that made sense ...
good luck !
- john
Glenn Thoreson
30-May-2010, 17:33
It's not easy to get an accurate reading using a photo transistor on these shutters. The slot in the curtain is too long and you get too much extraneous light, which screws up the reading. The way I do it is to take a junk film holder without slides, and drill a hole about in the middle of the thing. The hole wants to be a size that comes close to the diameter of your sensor. Remove the ground glass from the focus panel and put your drilled film holdler in, as you normally would. Put your sensor up to the hole and run your test. These things are not perfectly on the money, but they should be pretty close. . The initial winding tension should allow the curtain to close slowly but firmly from the O (open) position. Old timers used to crank the initial tension up a little for sports photogerring. The instructions in the manual should be very close to correct. One thing about a single slit curtain is it takes quite a lot of error in setting to make it unusable. It's funny that Graflex never menioned testing procedure on these that I've seen. There was a proceedure for testing them using a phonograph turntable, though. (don't ask) :D
rdenney
30-May-2010, 19:31
It's not easy to get an accurate reading using a photo transistor on these shutters. The slot in the curtain is too long and you get too much extraneous light, which screws up the reading.
Okay, here's what I did to try it again. I built a 3/4"-long "soda straw" made from flat black masking tape wrapped inside out, and then affixed it to make a tunnel visor for the photo transistor. Then, I held it up to the ground glass, and measured the projected image of a lamp through the 127mm Optar. I tested the measurement technique by measuring the Graphax shutter at 25, and measured exactly the same speed I measured with that shutter off the camera (and all the slow speeds on that shutter are well within 10%). I suspect my soda straw is preventing any weirdness from the slit, and I'm not seeing any effect of the ground glass in my control.
Here are the speeds I get:
1000: 1/300
500: 1/125
250: 1/136
125: 1/77
50: 1/45
30: 1/31
The ungoverned speeds (1000, 250, 50), where it seems to me spring tension might have more of an effect, are more wildly off at the high speeds than the governed speeds. That's just the opposite of expectations. The largest curtain opening, both free running and governed, is close to perfect, and I would think a curtain-speed problem would affect that opening as much as the narrower openings.
Tightening the spring beyond a certain point has no effect on speeds. The shutter sounds excellent, and the curtain is supple and looks in excellent condition.
I'm putting this in the "go figure" category unless someone has another suggestion.
By the way, I did test the 30 setting with the camera upside down, because I just couldn't help myself. It ran 10-20% slower than right-side up.
Rick "not unhappy with the available speeds, but who'll need a table to remember them" Denney
I know that I had similar difficulty when testing my Speed. My light source was too big and my sensor was way too far from the FP shutter. Switched to a very focused beam (sounds like you did a similar thing with the soda straw) and moving the sensor closer to the FP to mimic the film plane made a huge difference in my measured speeds. My 1/1000 tests out at 1/500 and 250 and below is close enough for my work. Try using a laser pointer--very little scatter. I also thought about using the flash sync conductors on the FP to time the speed of the curtain and do the math to figure out the accurancy but then I decided to just get out and shoot--way more fun.
On second thought would an engine timing light work?
rick
did you take the body cap off the end of the roller shutter and put a drop of oil
on the end of the rod? ( be careful not to get any on the shutter itself)
that sometimes helps to make things run smoothly.
( i noticed a difference before and after when i did that with my slr ) ..
did you try to get the 1/30th S nailed with the camera upside down
and see where the rest of the speeds ended up ? it might be just enough
of a difference that the other speeds are a 1/2 stop- off instead of a full stop+.
calumet testers work OK with the lens removed. i have a friend who used his industrial-grade
shutter tester at his repair shop and we removed the lens and just held the probe
( series D slr ) i would imagine a speed graphic would test the same way, seeing
his fancy tester was pretty much the same thing as a calumet tester, except
it came with a light source ( 2 -3 feet away) to point the camera at,
and something to rest/sit the camera on while it was being tested.
we didn't have anything to block the light, or anything like that
one of us held the sensor around where the film plane was, and the other one
wound and released the shutter. needing to be someone with 4 arms was the hard part.
john
rdenney
31-May-2010, 16:50
did you take the body cap off the end of the roller shutter and put a drop of oil on the end of the rod? ( be careful not to get any on the shutter itself) that sometimes helps to make things run smoothly. ( i noticed a difference before and after when i did that with my slr ) ..
Yes, I took both caps off, and I also cleaned the ends of the curtain shafts and swabbed out the bushings in those caps before applying a touch of the light grease as the service manual specifies. Same on the shutter mechanism side. The curtain seems to move smoothly enough, but I still haven't tested it (my testing today didn't get past the front shutters).
I wouldn't know where to begin to adjust the 1/30 speed. The governor is not adjustible, and the only way I can think of to adjust the curtain speed is to tighten up that spring even more. But that had no effect--I stopped tightening when doing so stopped having an effect.
The governor is just a spinning centrifugal brake--no adjustment capability. And though the manual provided troubleshooting procedures, it did not provide any means I could find of actually adjusting the speeds. Seems that if everything is right, so are the speeds. Therefore, something must not be right in mine. But there are still some usable speeds, except that I really wish the rear shutter went down to 1/4, which is just a hair fast for manual timing, at least for me.
I have experience with a Calumet tester that I borrowed from Dan Unkefer a few years ago for long enough to test a bunch of Second-World cameras. I found that the light source for it was quite sensitive. I ended up using a battery-power flashlight placed enough distant to just trigger the tester, and shining through the shutter without a lens. The diode tester is a lot more forgiving. I get basically the same results using any incandescent or window light that is sufficient to saturate the diode. Unless the fall delay is substantially longer than the rise delay, the diode tester should be correct if one knows where on the waveform to measure. The Calumet tester does measure fast shutter speeds correctly, but only if the light source is just bright enough to trigger it.
Rick "enjoying the project" Denney
rdenney
31-May-2010, 19:02
It finally occurred to me to check the shutter tester against a proven focal-plane shutter. So, I pulled out my Pentax 6x7, which uses electronic timing for closing the second curtain. The timer is accurate--at speeds of 60 and slower the shutter speed measured within five percent. 1/15 was worst at a measured speed of 1/15.7. But that is unpersuasive--the maximum error is still .003 seconds, which is too much for the fast speeds.
And 125 measured 114, 250 measured 200, and 1000 measured 435. (I forget what 500 was.) I think the wave shape seen from a photodiode through a sound card is just too hard to interpret with sufficient accuracy at speeds of 250 and higher, and perhaps even 125 and higher. Given where most of us live in terms of shutter speeds, this is no weakness at all. The fastest shutter speed I can use with 100-speed film is about 125, unless I happen upon a lens with a useful aperture larger than f/5.6, heh.
Rick "who just realized this project is complete" Denney
Paul Metcalf
1-Jun-2010, 06:44
On second thought would an engine timing light work?and hooked up how?
and hooked up how?
To the ignition coil on the flash connector bracket, of course!
~Joe
Rick, why don't you try what Brian suggested - shoot a series of exposures on a calibrated film and measure the resulting densities? Maybe even take a control series of shots with a known shutter on the same setup for good measure?
Rick, why don't you try what Brian suggested - shoot a series of exposures on a calibrated film and measure the resulting densities? Maybe even take a control series of shots with a known shutter on the same setup for good measure?
That was my plan, as hinted above. Given that I don't do black and white any more, and have no darkroom, profligate experimentation is expensive, so I was trying to answer the obvious questions first.
Rick "and the use of slow film and tessar lenses suggests rare needs for fast shutter speeds" Denney
BrianShaw
1-Jun-2010, 08:25
Rick,
You can do the same with color film... but I know you know that. ;)
I actually prefer to test like this using E6 since I can drop it at the lab and have it back in 3 hours.
But I wouldn't consider one (or even two) roll(s) of film(s) to be "profiligate experimetnation(s)", but thanks for the Word of the Day!
Doh! Just realized the power came from the ignition system. I was thinking strobe light. Whoops.
Jim Graves
1-Jun-2010, 09:24
It finally occurred to me to check the shutter tester against a proven focal-plane shutter. ....... I think the wave shape seen from a photodiode through a sound card is just too hard to interpret with sufficient accuracy at speeds of 250 and higher, and perhaps even 125 and higher. ....... [parts of quote deleted].
Rick ... I came to the same conclusion about the inexpensive speed testers.
I have tried two of the photo transistor shutter testers regularly available on ebay ... one is identical to the build-it-yourself version from plans on the internet [uses a free download program for timing on your computer][$40 range] ... the other a more recently available Chinese manufactured tester with the readout right on the unit [$120 range.]
I tested both on 3 relatively modern leaf shutter LF lenses that had just been professionally timed by California Precision Services here in Sacramento. I warmed up the shutters with at least 20 releases prior to testing and took 10 readings at each speed. The results were very accurate up to 1/100 at which point they began to vary by large percentages [growing larger as the dial speed increased] ... in the 1/250 and 1/400 range they both indicated at least 25% slower than the professional results on all 3 lenses. A small sampling admittedly but very consistent so I would hazard a guess that either the photo transistors or the circuitry limits accuracy at higher sutter speeds.
I'm curious what other members here have experienced with these testers.
Rick ... I came to the same conclusion about the inexpensive speed testers.
I have tried two of the photo transistor shutter testers regularly available on ebay ... one is identical to the build-it-yourself version from plans on the internet [uses a free download program for timing on your computer][$40 range] ... the other a more recently available Chinese manufactured tester with the readout right on the unit [$120 range.]
Struan and I were having this conversation in another thread just yesterday, and what follows is out-loud thinking. The rise time of typical photo-diodes, when properly driven, is in the tens of microseconds--easily fast enough for this application. The problem is the sound card being used as a data sampler. The sound card is designed for audio, and given that all audio lives in the frequency domain, the sound card will filter out direct current. The current passed by the diode (as pulled by the resistor) will be low when there is no light, and high when there is light. Smaller resistors pull more current and more current means more power and a faster rise time, but it also means pushing the diode closer to its power-handling limits. Most of these home-built testers use high resistances to keep the current flow fairly low.
But when the current rises, so does the output voltage of the diode. And that's what the sound-card reads. But even high-end sound cards roll off response below about 30 Hz using something like a capacitor on the input, which filters out DC and also puts a "swoop" in the rising and falling waveforms as the diode is switched. One way to cure that would be to feed high-frequency audio to the diode instead of DC, using an oscillator. The frequency of oscillation would have to be fast enough to be negligible to the shutter speed being tested, but slow enough to be within the frequency response of the sound card and also slow enough to not be canceled out by the rise and fall time of the diode, if that is possible.
Or, a properly driven diode could be read by a device that doesn't filter DC, such as a data acquisition system for the PC, or a storage oscilloscope. These are a lot cheaper than they used to be, but they are still 20+ times that cheaper tester on ebay unless you know how to scrounge, or write your own acquisition software.
I suspect the Calumet tester, which I know will measure high speeds accurately, couples a fast photo-diode to a circuit that triggers a counter on the first positive change of voltage, and again on the first subsequent negative change in voltage. Not hard to do with a diode detector. It then couples that output to a high-speed timer. The Calumet tester will trigger on a single pulse of a fluorescent light, and even the AC pulsing of an incandescent light will mess it up. The detector is sensitive and the light has to be at just the right intensity. If the more expensive tester is like the Calumet tester, then try it again with a battery-powered flashlight (with a real bulb, not LEDs) and arrange the test so that the light is just barely bright enough to trigger it. You might get more accurate results.
Rick "who might give this project to a local ham radio group to see if one of them will whitewash the fence" Denney
Sounds like you need to hook it up to an O-scope instead of a sound card. It would be simple to measure the width of the electrical pulse with that.
Sounds like you need to hook it up to an O-scope instead of a sound card. It would be simple to measure the width of the electrical pulse with that.
Yup. But my scope is not a storage scope, and getting one of those would change the cost equation.
Rick "already subject to wifely scrutiny regarding test equipment" Denney
Jack Dahlgren
7-Jun-2010, 11:20
Or, a properly driven diode could be read by a device that doesn't filter DC, such as a data acquisition system for the PC, or a storage oscilloscope. These are a lot cheaper than they used to be, but they are still 20+ times that cheaper tester on ebay unless you know how to scrounge, or write your own acquisition software.
I suspect the Calumet tester, which I know will measure high speeds accurately, couples a fast photo-diode to a circuit that triggers a counter on the first positive change of voltage, and again on the first subsequent negative change in voltage. Not hard to do with a diode detector. It then couples that output to a high-speed timer. The Calumet tester will trigger on a single pulse of a fluorescent light, and even the AC pulsing of an incandescent light will mess it up. The detector is sensitive and the light has to be at just the right intensity. If the more expensive tester is like the Calumet tester, then try it again with a battery-powered flashlight (with a real bulb, not LEDs) and arrange the test so that the light is just barely bright enough to trigger it. You might get more accurate results.
Rick "who might give this project to a local ham radio group to see if one of them will whitewash the fence" Denney
I just ordered one of these:
http://www.sparkfun.com/datasheets/Sensors/TSL230R-LF-e3.pdf
and am planning to hook it up to an arduino or other microcontroller to log data and calculate exposure. I think it should be possible to catch the front edge of the shutter opening and integrate over the entire open time. Then compare against the peak value in order to get an effective shutter speed regardless of f/stop.
I'll update on progress as soon as it gets here and I can build a circuit.
Once I get that working it can probably be made into a sensitive light meter. With a time series of readings it probably could calculate flash exposure and other stuff as well. Once you have the data there are tons of things you can do with it.
-Jack
I just ordered one of these:
http://www.sparkfun.com/datasheets/Sensors/TSL230R-LF-e3.pdf
and am planning to hook it up to an arduino or other microcontroller to log data and calculate exposure. I think it should be possible to catch the front edge of the shutter opening and integrate over the entire open time. Then compare against the peak value in order to get an effective shutter speed regardless of f/stop.
I'll update on progress as soon as it gets here and I can build a circuit.
Once I get that working it can probably be made into a sensitive light meter. With a time series of readings it probably could calculate flash exposure and other stuff as well. Once you have the data there are tons of things you can do with it.
-Jack
And I just ordered one of these:
http://pdamusician.com/dpscope/index.html
It's more than I paid for my Tektronics portable, but it has storage and plenty of bandwidth for this and for spectral analysis of music. It's no good for the RF stuff (except at audio frequencies), but then I don't need a storage scope for that and the Tektronics scope works fine.
Rick "don't tell my wife" Denney
BetterSense
7-Jun-2010, 15:45
Once I get that working it can probably be made into a sensitive light meter.
I have made several light meters with these chips. They are very easy to implement with the digital output but downright require spectral response filtration.
You can see them on my very embryonic website
chazmiller.com/projects/lightmeters.html
Jack Dahlgren
8-Jun-2010, 00:40
I have made several light meters with these chips. They are very easy to implement with the digital output but downright require spectral response filtration.
You can see them on my very embryonic website
chazmiller.com/projects/lightmeters.html
Interested in sharing the code? I'm heading down the same path and could use a lift.
kidslong
8-Jun-2010, 02:00
I have experience with a Calumet tester that I borrowed from my friend a few years ago
Yup. But my scope is not a storage scope, and getting one of those would change the cost equation.
Rick "already subject to wifely scrutiny regarding test equipment" Denney
Mine isn't a storage scope either. If you have a normal scope, you could turn up the crt persistence knob and it would keep the shutter waveform on the screen long enough for you to measure the pulse (e.g., it took 4+1/2 squares). Adjust the time/division to be appropriate for the shutter speed. Sweep trigger wouldn't be important, but could make it nice and neat for the shutter pulse to start at the beginning of the screen.
I have experience with a Calumet tester that I borrowed from my friend a few years ago
As do I. They are finicky, and to get a clear reading, the light source must be without any pulse at all, and it must also be just bright enough to reach the trigger threshold on the tester. If it is too bright, the readings will be inconsistent, and even the pulsing of an AC-mains-powered incandescent lamp will trigger it untimely. Of course, one still had to convert the reading of tenths of milliseconds (or whatever it was) into fractions of a second, and still had to calculate the exposure error. I ended up using a battery-powered flashlight placed about two feet in front of the bare shutter under test, and I wrote a spreadsheet to turn the numbers into something more useful. Using it was not completely like hitting the EASY button.
They are also expensive now. For that money, I can get stuff that does just as well, shows me the waveform, and provides good toy value for other hobbies as well. But that is even less like hitting the EASY button.
Rick "obviously not always interested in the easy way" Denney
Michael Cienfuegos
8-Jun-2010, 23:14
As do I. They are finicky, and to get a clear reading, the light source must be without any pulse at all, and it must also be just bright enough to reach the trigger threshold on the tester. If it is too bright, the readings will be inconsistent, and even the pulsing of an AC-mains-powered incandescent lamp will trigger it untimely. Of course, one still had to convert the reading of tenths of milliseconds (or whatever it was) into fractions of a second, and still had to calculate the exposure error. I ended up using a battery-powered flashlight placed about two feet in front of the bare shutter under test, and I wrote a spreadsheet to turn the numbers into something more useful. Using it was not completely like hitting the EASY button.
They are also expensive now. For that money, I can get stuff that does just as well, shows me the waveform, and provides good toy value for other hobbies as well. But that is even less like hitting the EASY button.
Rick "obviously not always interested in the easy way" Denney
Hope you can stay out of the dog house with this new equipment coming your way.
More on this subject: I have obtained a DPScope, which is an inexpensive digital oscilloscope that connects to a USB port. It's a funky little kit, not without it's foibles (mostly as a result of its serial/USB converter), but once I made it work, it is good for this application.
The scope will record, as a storage scope, one screen's worth of information, including 200 samples on that screen. The screen is a typical o-scope display, with 20 divisions horizontal and vertical. It is possible to set each division to 0.5 microseconds, resulting in the ability to record an event 1 microsecond long at a sampling rate of 20 million samples/second. The vertical resolution is similar, being able to measure input voltage to 0.5% of scale, with the vertical divisions ranging from 5 millivolts to 1 volt. The bandwith is about 1 MHz, which is fine for this application and for audio measurements.
The circuit that I used is very simple. The positive lead from a AA battery single-cell holder is connected to a 4.7KOhm resistor. That connection is where the positive lead of the scope is connected. The other end of the resistor is connected to the collector of an IR phototransistor, Radio Shack part number 276-145. The negative lead to the scope is also connected at this location. The emitter of the phototransistor is connected to the negative battery terminal, through a switch.
This phototransistor is actually a TIL414, which has a rise time of 8 microseconds and a fall time of 6 microseconds. I wrapped black photo tape around the LED to make a soda straw about 3/4" long.
With the leads connected as above, the voltage measured will be the battery voltage at full light, and zero in dim room light. Full light came from a desk lamp with a 100-watt bulb.
The tester is sensitive enough to measure the 1/400 speed on a Compur 00, without removing the 47mm Super Angulon cells (lens at f/5.6).
I set DPScope to trigger on the rising voltage, with the trigger voltage at about .02 V.
I have attached a screen capture of the scope output for the measurement of the 1/1000 shutter speed on a Speed Graphic 3x4 shutter. Given the fast rise time, the somewhat more leisurely increase in voltage is the time required for the leading and trailing edges of the curtain opening to pass in front of the sensor. The phototransistor may have reached saturation before it was fully exposed. If I measure from the trigger point to the first point at which the voltage starts to drop, even slightly, I get a measurement of 2 milliseconds, or 1/500.
The 1/400 shutter speed on that Compur 00 measured 1/300 using this method. This is exactly what I got using the "sound card shutter tester" I bought from ebay, and plugged into a good E-MU external USB sound card. Thus, I do not believe that the fact that the sound card filters out DC prevents it from making reasonable measurements, even up to the 1/500 shutter speed. I measured speeds of 1/450 for one of my Compur 0 shutters at the marked 1/500 shutter speed using the sound card.
Thus, I do not believe that a sound card will undermine the accuracy of these home-made testers. If there are inaccuracies, they are outside the electronics of the tester.
The only conclusion I can come to is that the size of the curtain opening on the Speed Graphic is based on an assumption of the curtain speed at the full winding load of the spring, which would cause the curtain to move faster than for the next curtain opening. Friction may be preventing this. That's the only way I can explain how a narrower slit for 1/1000 provides nearly the same shutter speed as the more open slit for 1/250.
Rick "figuring only about three of us are interested in this at this level" Denney
Jim Graves
3-Jul-2010, 22:18
So now put some film in it and shoot the same scene at 1/1000 and 1/250 and see if you're right.
BrianShaw
4-Jul-2010, 07:05
Rick "figuring only about three of us are interested in this at this level" Denney
Three and one-quarter. Interesting! I read your post and then went back to sleep. :D
bbjorkum
4-Jul-2010, 07:08
I just attach my Nikon D300 to the back of my Graflex via an adapter, points the camera at an even surface, and compare the histograms from the Nikon shutter with the Graflex shutter. Easy, reliable.
steven_e007
5-Jul-2010, 02:40
Hi,
This is my first post here. Actually, my reason for finding and joining the site is because I have just stripped and rebuilt a Speed graphic. I tested the shutter and found it to be all over the place. I then serviced it as per the original service manual that you can find on the net and reassembled it following the 'just enough spring tension to close the shutter when inverted' calibration procedure.
It works... but I have exactly the same experience that you describe in your original post. Up to about 1/300 s the speed seems to increase as I crank up the tension or use a smaller slit. They don't match the speeds printed on the plate on the camera too well - but no matter.They are quite consistent and I can laminate up a table...
The problem is that after this speed the shutter gets no faster. Speeds with a spring tension of "6" are hardly any faster than "3". The spring certainly feels tighter, the shutter makes more noise, it seems to hit the stops harder...but measures much the same speed.
How do I know my shutter tester is accurate? Hmmm.. it is home built. I mentioned the problem to my 'other half' who (rather mockingly, I thought) said 'I suppose you need to build a tester to test the tester, now'. Well... she was right. I'm right in the middle of testing my tester with the shutter tester tester (!) - but preliminary results seem to indicate that, whilst my shutter tester overestimates the shutter speed by an error margin which increases as the shutter speed increases (about 25% at 1/1000s) - it does work and should certainly be able to distinguish between 1/250s and 1/1000s.
So... looks like the shutter really does go no faster than 1/300 s :-(
Anyway... at least you have a fourth person interested ;-)
You can amend that count to five. I've got to do the same with mine eventually, and I'm riveted by this conversation. So, any of you folks feel like preparing a more in-depth (with PICTURES!) guide to it? ;)
steven_e007
5-Jul-2010, 05:32
You can amend that count to five. I've got to do the same with mine eventually, and I'm riveted by this conversation. So, any of you folks feel like preparing a more in-depth (with PICTURES!) guide to it? ;)
Sure!
Do you want an in depth guide to servicing the shutter?
Or building a shutter speed tester?
Or testing the shutter speed?
Or testing the shutter speed tester?
Hard to believe, sometimes, that the purpose of all this was to take some pictures!! :eek:
Well, no, the purpose was not to take pictures. The purpose was to take pictures with some confidence in the shutter speed. But even that is smoke screen. The real purpose was to play with the camera, and play with the test equipment. Taking pictures is not my only hobby.
To the question of dialing up the spring preload to increase the shutter speed on a Speed Graphic:
Consider the difference in shutter operation between the higher speed range (1000, 250, 50) and the lower speed range (500, 125, 30). The latter achieves one stop slower shutter speeds by severely limiting the speed of shutter travel using a timed escapement. For those same three slits to provide a speed one stop slower, the curtains have to move at half the speed across the film plane.
That suggests to me that an error of stop or more at the highest speed isn't going to be affected much by a change in spring pretension. That change just can't affect the travel speed of the curtain that much. But this is an advantage, and what makes the shutter pretty consistent despite the garage-door-spring design.
I have adjusted my shutter exactly as per the Graflex service manual, including a full teardown, and using lubricants just like those that are recommended. What opened up this line of questioning in the first place was that it didn't seem sensible to me that the middle speed could be accurate while the fastest speed is not, given the fixed size of the curtain openings. I would have thought that if the friction of the system was out-of-spec enough to cause a one-stop error in the fastest speed, then the middle speed would be affected proportionally.
But my conclusion from more careful testing is that the electronic sound-card tester that people use with the plain Radio Shack phototransistor is not inherently inaccurate, even though sound cards filter out DC resulting in a somewhat strange waveform when recording using something like Audacity.
That does not mean there isn't an error caused by the mechanical or optical configuration of the tester. That will be my next line of inquiry. The phototransistor is in a typical T-1-3/4 package, with a +/- 20-degree beamwidth. When shining a desklamp directly into the focal-plane shutter, the light will not be as collimated as it would be as projected by a lens. The desk lamp might be exciting the phototransistor before the edge of the curtain opening really gets to the center of that package, and for some time after the closing edge has passed. I suspect that edge error will be fixed, and only a significant percentage of the shutter speed with the smallest opening (the opening used for 1000 and 500). I need to test again with a collimated light source.
The suggestion of using a DSLR mounted on the back of the camera is fine, but not exactly cheap. Also, it requires some trial and error to get the histograms to line up to determine how much error is in the shutter. Histograms are not calibrated in stops, after all.
Rick "whose mechanical sense is not happy with the notion of results so far being simple friction or wear" Denney
steven_e007
5-Jul-2010, 15:40
Well, no, the purpose was not to take pictures. The purpose was to take pictures with some confidence in the shutter speed. But even that is smoke screen. The real purpose was to play with the camera, and play with the test equipment. Taking pictures is not my only hobby.
Actually, I was refering to myself...
My objective certainly was to finally get the Speed Graphic I bought 10 years ago serviced and working so I could take pictures. I have a nasty habit of getting sidetracked. This time I was determined I wouldn't be...
:rolleyes:
I too serviced the shutter as per the manual and tried as best as I could to lubricate and adjust as suggested. I am reasonable confident my slower speeds are reasonable - up to around 1/200s maybe.
But... my Speed Graphic is an Anniversary. It has the more primitive shutter with 6 spring tensions but no governor. It appears to me that something can not be right if the maximum speed is reached at '3' - about half the spring tension - and further tension does not increase the speed my tester reads - even though it seems obvious the spring tension is increasing.
So... I am now very suspicious of the tester. After spending 2 hours tonight measuring slots in spinning aluminium disks I have almost convinced myself that the shutter tester is indeed flawed.
Reading up on the electronics - I am now very suspicious of the performance of the phototransistor in this application. I do not have access to an oscilloscope at home - and the office for me is 200 miles away. Are you able to measure the actual rise and fall times of your phototransistor? I suspect the configuration, both electronic and optical, may be hamstringing it's performance...
Jim Graves
5-Jul-2010, 16:13
Well, I posted earlier in this thread and may have come off sounding a little sarcastic ... but that certainly wasn't my intent ... when I suggested going out and shooting film to see if the testers were accurate.
I've gone through this with 3 testers ... two really cheap ones and one more expensive one. All 3 were showing the same "accurate on slow speeds and increasingly inaccurate on higher speeds" that are written about on this thread. I had had my shutters professionally tested prior to getting the cheap home use ones ... so I was relatively confident in the actual shutter times recorded by the professional lab. I should also note that these were leaf shutters and not focal plane curtain shutters which may well raise other issues in speed testing.
I finally tested the testers by actually shooting film and decided the home testers were wrong on the higher speeds because the exposures were correct using the professional test speed results and off when using the settings indicated by the cheaper testers.
I then bought the third tester and got the same results ... which Rick suggested might be remedied by changing the light source and lowering the light input intensity ... (I haven't yet tested his suggestion.)
So you can count me as another very interested reader of this thread ... especially because I also use a Speed Graphic with a lot of my barrel lenses.
The question was: Did I actually measure the rise time of the phototransistor. The answer is: No. And the reason is that I don't have a light source that will turn on as fast as the reported rise time for this transistor, which is six microseconds. Given that 1/1000 is 1000 microseconds, I have to believe that anything approaching the specification for this phototransistor is going to present a negligible error due to rise time.
I should also note that in reviewing the specs for all the cheap phototransistors in the Mouser catalog, I only saw two (out of dozens) that had rise times more than a 100 microseconds. This confirms something Struan suggested earlier--the rise time of these devices is not at issue.
And, yes, I have actually done the measurement with an oscilloscope, as shown above. The scope's bandwidth is not that large (about 1 MHz), but it's certainly enough.
Let's think about what the curve should look like, based on how these devices actually work. Near as I can surmise, as light falls on the transistor, it provides the bias voltage for the base of the transistor, which allows the voltage supplied to the collector to be fed to the emitter. Thus, the transistor is acting like a light-controlled switch. There is a minimum and maximum amount of light necessary to allow current to begin flowing and beyond which little more current will flow (I will call this saturation). To truly get the most out of the transistor, the light would have to be adjusted to make use of a large measure of that range, but not to use all of it.
If the transistor never reaches saturation, you will get an S-shaped curve as the shutter opens, revealing the nonlinear response to light (similar to the characteristic curve of film). There will be a falling S-shaped curve as the trailing edge of the shutter covers the sensor.
With diffuse light, the shadow of the curtain edge won't be sharp and the curve will vary not only with the amount of the sensor that is exposed, but also according to the fuzzy edge of light falloff around the edge of the shutter. The sensor has optics, and will accept light from an angle, which will also provide a non-linear effect as the shutter edge passes over the sensor.
The ideal sensor would have zero dimension so that the switching occurred at the maximum speed of the semiconductor. But that's not what we have.
Now, let's do a little math, because I think the geometry of the sensor is the problem, not the electronics. The curtain opening for 1/1000 of a second is 3.29mm according to my precision calipers. For a 3.29mm opening to expose film for 1/1000 of a second, the curtain must move 3.29mm every .001 seconds, or 3290mm/sec. The T-1-3/4 case of the phototransistor is 5mm in diameter. With a nose cone of that diameter, let's assume the sensitivity zone of the sensor is that wide. At 3290mm/sec, it will take 0.00152 seconds for the curtain edge to expose the sensor, by which time the trailing edge of the opening will already also be over the sensor. From the time the sensor first sees any hint of light until the time it no longer sees any hint of light will be the width of the sensor plus the width of the slit. That dimension is 8.29 mm, which should take 2.5 milliseconds.
But we cannot assume that the tube I placed over the sensor is fully effective at restricting non-axial light from reaching the sensor. That tube was 5mm wide and 19mm long, and the case of the transistor consumed perhaps a third of it. Let's say the tube extended in front of the transistor by 12mm. Doing the math, if I held the sensor 10mm behind the shutter, the effective sensor size would actually be 8.3mm. If so, then the traverse distance of the 3.29mm slit would be 11.6mm, which would require 3.5 milliseconds.
Going back to the scope output linked HERE (http://www.largeformatphotography.info/forum/attachment.php?attachmentid=42711&d=1278218052), it appears that the whole curve, from first effect to last effect, is very close to--SHAZZAM!--3.5 milliseconds. I'm therefore now prepared to believe that my shutter is indeed delivering close to 1/1000.
If the transistor does not reach saturation, the resulting curve should be a hump with no flat top. Given the time it takes for the curtain edge to expose the wide sensor, the opening and closing S-curves overlap, making it impossible to reliably determine where the shutter edge passes the center of the sensor. If the transistor DOES reach saturation, it will clip at that saturation point. That clearly happened in my shown example. The sharp corners of that saturation point are misleading--they do not describe anything about the shutter. This misleading effect has clearly ruined my previous interpretations of both the scope and sound card results. I bet I'm not alone.
Now that we can quantify the error, let's see how much it will affect slower shutter speeds. I measured the 1/250 opening at 9.7mm. To achieve a 1/250 shutter speed, the curtains will have to move at 2425mm/second. This is slower because the spring tension is less at this winding of the curtain. The important thing is that the shutter opening is larger than the sensor, so it is now possible to see the full effect of exposing the sensor separately from the full effect of the closing of the sensor. If we measure from the first effect for the opening to the first effect for the closing, and we prevent the transistor from reaching saturation, then even with a sound card, we should be able to get an accurate measurement. Using the same method as above, the traverse distance of the shutter over the sensor is the effective sensor diameter plus the slit opening, which is 8.6 + 9.7 = 18.3mm. At 2425, the full effect on a scope should take 7.5 milliseconds. If we look only at that full effect, we might conclude that the shutter speed is a stop slow (1/250 should be 4 milliseconds). But since we can look at the beginning of the opening effect and the beginning of the closing effect without one affecting the other, we can measure accurately.
The largest slit opening for 1/50 second is 34.3mm. To achieve a shutter speed of 1/50, the shutter will need to move at 1715mm/second. Thus, the shutter design assumes it will lose half its speed between the highest speed and the slowest timed speed when running free (the 1/500, 1/125, and 1/30 series is governed on the Pacemaker shutter, but the governor will still respond linearly to spring tension). The effective sensor size with my setup of 8.3mm is only about a quarter of the curtain opening, and thus can have no larger an effect than around 25% of the marked shutter speed. Any likely interpretation of scope or sound-card curves will yield a usable measurement with better accuracy than that 25% limit.
Thus, it is my conclusion that one cannot accurately measure shutter speeds for focal-plane shutters where the opening slit is smaller than the sensor, without doing the math of calculating the total effect on the sensor.
I also conclude that if a Speed Graphic shutter is in a state of good maintenance and has the spring tension set to provide a proper measurement for the slow speeds, then it will more than likely be fine on the fast speeds.
Now, possible corrective measures for the tester: Clearly, the size of transistor is an issue. I will experiment with mounting the transistor and ten covering it with a piece of black tape into which I have made a pinhole. That is the approach the Calumet tester uses.
I will then adjust the light source so that the light just misses fully saturating the transistor. I think I can do that by adjusting the light source so that the transistor passes maybe 1 volt from the 1.5-volt battery. Also, the further away the light source, the more collimated it will be. Stay tuned.
(For testing leaf shutters, the mechanics are far more complicated. If I can resolve the issues where the math is understandable, then maybe I won't have to understand the math of the leaf shutter to be able to apply it.)
Rick "necessarily getting closer to the Calumet tester design--who'da thought?" Denney
steven_e007
7-Jul-2010, 02:11
OK... I did study electronics at university - but that was 1979 :eek:
I think I am a bit rusty, so bear with me. I'll try and keep this very simple, partly because I am aware that this is a photography forum, not an electronics one - and partly because I don’t remember it so well!
As I understand it - the rise and fall times quoted in the datasheets for phototransistors are theoretical maximums based on the junction capacitance. Whether you get these times in practice depends on how the device is used and how sophisticated your electronics are. In practice, the actual rise and fall times can be much greater. The circuit most DIY shutter tester makers use is a simple battery + resistor + phototransistor. If the resistor is connected to the emitter and the collector to the battery, this is called 'common emitter' mode.. In common emitter mode the voltage at the transistor - resistor junction is low with no light and rises towards the battery voltage when light shines on the transistor.
Now... a transistor is an amplifier. Even though we want to use it as a switch - it has a gain. In common emitter mode this gain effectively amplifies the transistor's own junction capacitance. (It's called Miller effect, if anyone wants to google it). Capacitance is BAD - because in effect we have to charge it up as we try to switch the transistor from ‘off’ to ’on’ - thus stuffing the nice rise time quoted in the spec sheet. We could try some calculations here... but let's not. I have found some info in a design paper on optocouplers that suggests an increase in the rise time of 20 times is not untypical.
A second problem is that the fall time is dependant upon the current flowing in the transistor - and this in turn depends on the light falling on the junction. Essentially, less light means less current which means the internal capacitances charge and discharge more slowly.
My own experiments have already led me to pursue the 'pinhole' idea. The smallest pinhole I could manage to get to work reliably was 1mm. Also, to reduce scatter and increase collimation I have tried to keep the incident light well back and controlled with a shade. Now... all this drastically reduces the amount of light entering the transistor, reducing the current flow & slowing the fall time. Again, taking guestimations from the optocoupler article, a figure of about 4 times seems typical. The rise and fall of my particular phototransistor are given as 9 and 9 microseconds in the datasheet. In this crude common emitter circuit with restricted light input they could actually be more like 200 and 40 microseconds.
Does this matter? Well... it is still pretty fast compared with the shutter. It certainly won't fully account for the errors I have measured with my shutter tester, but 1/1000 second is 1 millisecond. 1 millisecond = 1000 microseconds. So, 240 microseconds is approaching 25% error. No longer negligible.
With a sound card consideration must also be given to what the circuit is driving, including input impedance & capacitance. I haven't given much thought to this - as I am NOT using a sound card – but I suspect this could be another source of error.
My shutter tester consists of a crystal based oscillator which outputs 100KHz, 10KHz and 1KHz. I am lucky enough have access to a very high quality testmeter, regularly calibrated, which measures frequency. It agrees with the oscillator crystal to within 1/10ths of a %. I've built a counter circuit, which is fast enough to count these pulses at 1mHz if required. The oscillator feeds the counter via a logic 'AND' gate - which is also fed by my phototransistor circuit. When the transistor output is high (shutter open) - the counter can count pulses. When the output goes low (shutter closed) the counting stops. From the number of pulses counted at a given frequency, the shutter speed can be calculated very easily. At low shutter speeds it should be extremely accurate. At high speeds I have seen that it is not.
What rdenney has said about the size of the phototransistor relative to the size of the shutter slot is quite valid & must be accounted for. My pinhole is 1mm, so I have in theory reduced this error to a maximum of 25% (1mm + 3mm slot). In a worse case scenario – slow phototransistor & the max possible error from the pinhole size – this still only accounts for 50% error. With my shutter reading much the same set for 1/1000s as 1/300 sec – something else must be going on.
I decided to replace the speed graphic with a spinning aluminium disk. This has a diameter of 200mm. I drilled a 12mm hole near the edge and use aluminium foil strips and tape to make a slot, which I could adjust from 1 to 10 mm. I placed this between my light source & phototransistor, By spinning the disk at a known rate (using the frequency meter to check the speed of rotation) I now had a way to present accurate slot sizes at accurate speeds to the tester. The results were similar to those for the speed graphic shutter. As the slot size decreased & the speed increased, the error increased until I got to a point where it seemed the tester could no longer ‘see’ the slot getting faster or smaller.
It made more sense when I stopped trying to calculate what I should have got and how much error I was actually getting, but instead worked backwards. Knowing the slot speed – I measured how many pulses the slot counted and then worked out from that how big the slot appeared be. It was usually 9mm, regardless of the actual size. This is why my tester seemed to ‘stick’ at 1/300s. Even though the slot got smaller, the tester still saw 9mm.
Looking for reasons, I noticed that the image projected on the sensor was bigger than the spot size due to spreading of the light source and the relatively large distance from the disk to the sensor (10mm). (I was spinning the disk at up to 360 rpm – I didn’t want it to accidentally snag my phototransistor at that speed!!!) In addition, the image of the slot wasn’t sharp. There was a fuzzy halo (penumbra, if you like) around the slot. These effects made the slot seem larger than it was. Problem is, as the soft edge of the light patch starts to cross the phototransistor, it is difficult to say at what point the light will be bright enough to trigger the transistor.
We now have three errors:
1. Slow rise & fall times.
2. Size of sensor / pinhole.
3. Apparent size of slot & scatter.
The trouble is – they all pull in the same direction: The combined effect seemed to be to stop the tester properly reacting to a slot less than 9mm, making the slot seem bigger and hence the shutter seem slower. The errors will be trivial for large slots, but make measuring 1/1000s on a focal plane shutter challenging.
My immediate reaction when I realised how much light spread I had was to think about using laser diode pointers & reducing the transistor shutter distance and so on… but then I thought: “What is the film to shutter distance in a speed graphic?” Bit fiddly to measure as the ground glass and the bellows make it a bit inaccessible – but… it is BIG. I guess maybe around 10mm. So – the apparent slot size problem & scatter applies to the film just as much as our tester. I fired the shutter on 1/1000 whilst holding the winding key so as to stop it mid travel. Holding the camera up to a very bright light & you can see the image on the ground glass – and it appears bigger than the slot complete with a penumbra . So – as it traverses the film, the feathering edge of the slot image will start to expose the film at any given point, this will get brighter as the slot moves across, then tapers off…
This brings us right back to what Jim Graves said. No, I didn’t take his comment sarcastically – I think he hit the nail on the head. We need to find a shutter speed that when tested photographically, gives us the density on the film that we expect for 1/1000s – then measure it. Trying to pin the shutter tester down to measure the speed of a 3mm slot that theoretically represent 1/1000s may not give the correct exposure on the film.
We still have the problem of measuring the curtain speed accurately, even when we know what it should be. My first idea was to use a photodiode instead of a transistor. These have rise & fall times in nanoseconds, not microseconds! I am trying a much better circuit that is specifically designed to be as fast as possible. But this only eliminates one error, it won’t help with the size of the sensor.
But... Maybe there is a better way?
If we used two detectors and placed then a long way apart – this could be as much as 100mm for a speed graphic – then we could just look for the leading edge of the slot. Ignore the slot itself – just detect the front edge crossing the sensor. If both sensors are identical, then it doesn’t matter at all what their shortcomings are – whatever delay is seen on the first is duplicated by the second & cancels out. Same for the audio card or whatever electronics are involved, doesn't matter - as long as we measure the same point on each rising slope. It should therefore be possible to measure the curtain speed VERY accurately. We could go one step further. Place on detector in the centre & one at the edge. By taking a reading, then rotating the unit by 180 degrees & taking another, we could see the difference in speed between the start & the end of the shutter’s run. Similarly, by placing the sensors on one edge and then the other, we can check for equal exposure side to side (none parallel slot test).
So, if we determine what curtain speed we need photographically for a given slot to give a desired exposure, from then on all we need is a 'two-detector curtain speed measurer' to give absolutely accurate measurement up to 1/1000s and with some diagnostic ability thrown in :)
Thanks for the details on the operation of the phototransistor.
So, a question. Has anyone opened up a Calumet tester and looked at it? I would have thought it similar to what Steven describes, with the counter attached to a sensor. I have noted the same sensitivities with that tester that I noted with my home-made tester, regarding the saturation point causing errors and so on.
Yet I have used a borrowed Calumet tester to test focal-plane shutter speeds up to 1/1000 (on medium-format cameras of dubious accuracy) that proved photographically accurate. So, they must be doing something to address this issue, and whatever it is must not cost much (if Calumet charged $100 for it originally, I would be surprised if it cost them more than $10 to make).
One thing they are doing is exposing the sensor through a very small hole. Also, they provide instructions that describe how to minimize the light source coming in to a point just under saturation.
But their trigger points are pretty sensitive. I was unable to get stable readings using even an incandescent bulb power by AC mains. I had to use a DC-powered torch.
The effect of the penumbra on the exposure of the film is hard to predict, depending on the shape of the characteristic curve, the aperture size, and the focal length of the lens. But this effect, if a large percentage of the slit opening (as seems to be the case), should make any measurement impossible, and only film tests could be used to test a shutter. I have this feeling that Graflex service personnel tested the speed of the curtain, rather than the exposure, to determine that it was within tolerance. The service manual just says, "test the shutter speeds for accuracy" with no hint as to what to do if the accuracy is not acceptable, what tools to use, the required curtain speed, or its tolerance.
But, as I said before, variations in speed should have a relatively small effect. The speed has to be cut in half to double the exposure. A 10% or even 25% error in speed won't affect most films significantly, and certainly not the films for which the Speed Graphic was designed.
If the penumbra effect is superimposed on top of the rise-time curve, then that curve will change the shape of the rise but not its length. If we measure from first effect to first effect (when we can measure it), we might still be fine, especially if the first effect reasonably corresponds to where the film starts and stops being usefully exposed.
In any case, this has become an academic exercise at this point--it is unlikely that I will ever have occasion to use 1/1000 on the camera. (Being an academic exercise does not, of course, make it any less interesting or enjoyable, at least for me.) Even in bright sun, the 100-speed film that I use won't ever need a shutter speed slower than 1/500 even when the lens is at f/5.6. I would be an issue for an owner of an f/2.8 Xenotar, perhaps. Probably not an issue for me.
If these testers can test up to 1/500 accurately and with reasonable precision (say, 1/10 stop), then they are useful for all leaf shutters. Leaf shutters, of course, expose the center more than the edges, and so their fastest speeds also vary with the aperture, at least.
My next experiment is to test the tester using a known-good focal-plane shutter. I would have done that this week, but my Pentax 6x7 has film in it.
Rick "not done with this topic just yet" Denney
miss_emma_jade
7-Jul-2010, 06:22
"The only conclusion I can come to is that the size of the curtain opening on the Speed Graphic is based on an assumption of the curtain speed at the full winding load of the spring, which would cause the curtain to move faster than for the next curtain opening. Friction may be preventing this. That's the only way I can explain how a narrower slit for 1/1000 provides nearly the same shutter speed as the more open slit for 1/250."
well i can see mine is like that, even without a shutter tester. its probably somewhere accurate till about 250 then the 500 and 1000 aren't much quicker.must be a way to fix this? can you buy new shutter curtains? or take them out and iron them maybe? i think maybe they have too many ripples and roll up to big, and maybe that creates the friction. my experience with roller shutters is mostly disastrous, but maybe one day....
ill keep reading now.
EdWorkman
7-Jul-2010, 07:53
I have been interested in this thread from the onset, albeit you electronic folks lost me early on- I have a pre-anniv graphic to which i attached a 203 Ektar in a Supermatic shutter. I took some shots on a sunny16 afternoon on TMY. I assumed that the Supermatic would probably go off no faster than 1/200 when set at 400, and also assumed that the FP could be a little slow. I had oiled the FP thru the little holes on the bottom and excercised it. I set the FP by the table at 1/435. Comparison of the negs revealed my surmise as to the relative speeds of the shutters to be correct- the 1/435 is close enough for sunny16 at EI 400, and the leaf shutter is definitely at least a stop slower. I was pleased with the tones and separations i got [ tray Xtol 1:2] as well as the shorter exposure using the FP
Michael Cienfuegos
7-Jul-2010, 09:10
This technical stuff is just fascinating, and I thought that all I had to do was wind it up and let 'er fly. If I ever figure out how to use an o'scope I may try and see how far off my 70+ year old pre-Anny Speed's FP shutter really is. I always thought that half the fun of using one of these old toys was lugging it around and having people amazed that they "Still make film for those old things?"
Keep it coming. I am learning more and more about this old thing than I ever thought I would. It is already held together with baling wire and gaffer's tape. Liquid electrical tape has taken care of a few pinholes in the shutter. Thanks to all for the great info. :)
BetterSense
7-Jul-2010, 10:24
I will ever have occasion to use 1/1000 on the camera.
But some of us do. I use my speed graphic to shoot skateboarding. I only ever use the fastest speed, and set the aperture to the smallest I can get away with on TMY.
If measuring the shutter curtain speed is desired, it seems pretty easy to measure it with a pair of photodiodes, since any penumbra effect would cancel out if you used the same photodiodes...you could just measure the temporal distance between the rise (or the fall) for the two diodes. Measure the width of the slit and you should be good. I bet you find that the curtain accelerates quite a bit as the slit travels across the film plane, though I have never noticed the top of my negatives being thinner.
steven_e007
7-Jul-2010, 15:56
Hi again,
My earlier (long) post about the shortcomings of the phototransistor refer to the very simple common emitter configuration using a battery, bulb and one resistor a recommended in several 'Make your own shutter tester' articles. The arrangement is as simple as you can possibly get and alas we have to accept the shortcommings.
I have no idea what is inside the Calumet tester - but I'll eat my socks if it isn't quite a bit more sophisticated than that. Both the rise and fall times can be improved and the sensititivity can be increased with more sophisticated electronics - although in this context 'more sophisticated' than one resistor might mean a second transistor and a couple more resistors - maybe a little too much for someone who doesn't like do it yourself projects - but well within the scope of many of us here, I'm sure. Next level up in complexity would be an op -amp circuit. 8 legs rather than two or three... it would need some rudimentary soldering skill or a breadboard - but still very simple.
Do we know if the Calumet even uses a transistor? Photodiodes are much faster, but need to be amplified.
It is past my bedtime, now, but if I get time tommorrow I'll see if I can post some of the circuits I have been experimenting with. I'll keep them as simple as possible so as not to put anyone off ;)
In the meantime here is the shutter table from my Anniverary Graphic.
I'm sure there is a big hint in here as to how the shutter timing works - but I haven't got my head round it, yet:
_ A _ B _C _D
1> 10 70 200 500
2> 15 80 240 600
3> 20 90 280 700
4> 25 100 320 800
5> 30 125 360 900
6> 40 150 400 1000
According to Tomas Tomasy's book, the shutter curtain slots are 3mm, 10mm, 35mm and 165mm.
Note the out of sequence numbers for the bigger slots. I'm sure this tells us something - I'm just not sure what it is!
It is past my bedtime, now, but if I get time tommorrow I'll see if I can post some of the circuits I have been experimenting with. I'll keep them as simple as possible so as not to put anyone off ;)
A photodiode/op-amp circuit would put me off not at all. You draw something and I'll order the parts from Mouser and build it.
Rick "who would like to plot those Anniversary Graphic shutter speeds to understand the trends" Denney
steven_e007
9-Jul-2010, 02:30
Ok, here's a few circuits as promised.
I'm aiming this post at photographers who have had a go, or might have a go, at the simple home made sound card shutter tester. Those who have a bit of electronic knowledge and circuit building skill can find all they need on the internet and will probably make a better job of it than me. The idea here is to suggest a couple of improvements on the very crude phototransistor circuit so it works a bit faster, but still keep it as simple as possible. I see Rick is keen to have a go at something a bit more ambitious. Bear with me - I'm mid experiment with my circuit at the moment, I'd rather post something that I've got to work myself, rather than something which ought to work! Give me a few days and I'll report back.
Please excuse the naff drawings. I tried to do them in 'Word' - but got to the point of either throwing the computer out of the window or resorting to paper and pencil :rolleyes:
The basic circuit uses one phototransistor, one resistor, one battery and is wired it into your sound card:
It is know as common emitter mode and suffers from slow rise and fall times, as previously discussed.
An improved version is called the Cascode. (This name comes from 'Cascade to cathode' I believe. Don't ask!)
http://www.freeimagehosting.net/uploads/175d8346fa.jpg
This gives a bit better rise time and much better fall times. Note that it needs just one more resistor and one more NPN transistor. Anyone who can wire up the basic circuit can do this one. The output will be high when there is no light and low when the shutter is open - but for the sound-card application this shouldn't matter.
The next one is a variation, still using just one extra resistor and transistor - but the transistor must be a PNP. If you don't know the difference between NPN and PNP - well, maybe it doesn't matter. There are two basic types of bipolar transistor and they are not interchangeable. Just make sure you use the right one. Component details come later...
http://www.freeimagehosting.net/uploads/16b7971c7d.jpg
The interesting thing here is that the fall time is only improved by a modest amount - but the rise time should be excellent - getting near the value in the phototransistor spec sheet. Since rise time in the very simplest circuit is the biggest issue - this circuit should give the biggest improvement.
Now, these are the simplest possible versions of these circuits. An electronics engineer would be tutting at this point about lack of decoupling and primitive bias, but they should work as well as the phototransistor only circuit.
Just slightly more ambitious then, here is an improved version:
http://www.freeimagehosting.net/uploads/4b7ea43175.jpg
This circuit includes a 0.1 microfarad capacitor which helps stability (and may stop the radio ham next door from broadcasting '10-4 rubber duck' over your computer speakers...) The base of the second transistor is connected to a variable resistor. This circuit can now be 'tweaked' to get the optimum sensitivity. The two extra resistors (values aren't critical) just mean the variable resistor has a bigger usable range. They can be left out - it just means the adjustment with all be around one spot.
Disadvantages of these circuits are that the voltage when the phototransistor is not conducting will be near the battery voltage - but when it is conducting the two transistors together form a voltage divider. They share the voltage between them, if you like. The lowest voltage is therefore about half the battery voltage. This shouldn't be a problem - the voltage swing is still high enough for a sound card. Not so useful to drive other electronics, though, if you intend to plug then into something else.
Components:
The transistors are not very critical. There must be thousands that would work. They just need to be reasonably fast silicon low voltage general purpose transistors. Here are a few that would do the trick:
NPN:
2N2222
2N3904
BC548
2N4401
BC238
PNP:
2N3906
2N3907
BC557
BC212
Whatever you choose, you need the spec sheet to see which leg is which. Don't be fooled into thinking the pin outs on one type of package will be the same for all packages of that type - you have to check the individual transistor number.
Resistors:
I use 5 volts, so RL should be about 4.7K and RB about 1K
If you use a 9 volts battery then maybe double these to 10K and 2.2K
(Resistors go in 'preferred sizes'. I'm sure it makes sense to someone).
The smallest or cheapest will do - nothing is critical.
The variable resisitor is just a linear trimpot or whatever you like. Again not important.
About 47K is right - maybe 100K if you leave the resistors out.
The capacitor isn't critical, just 0.1 microfarads - or try it without.
If it costs more than 30p / 50cents / 0.50Euro for any one component, find something cheaper! It should be possible to get enough parts to build two for less than $1!
I hope someone will give these a go and report back.
Steve.
Ok, here's a few circuits as promised.
Thanks! I probably have all the components on-hand, except the transistors, and these are common enough that I might find them at Radio Shack. I'm sure that I even have a 4.7K pot in the pile somewhere.
But a true HAM operator saying "10-4 Rubber Duck"? Them's fightin' words!
I'm not that fond of running these high unless lit--it makes it hard to set the trigger on the scope. Is there another spot in the circuit where the output will be low unless lit? In my simple circuit, which is not like the usual sound-card tester, I ran + to the signal lead on the scope and to a 4.7K resistor, the other end of the resistor to the ground lead on the scope and also to the collector on the photo-transistor, and the emitter to -. Thus, I was measuring voltage drop across the resistor, which would read zero when no current was flowing (as long as the scope input has very high impedance, which, of course, it does).
I may have time this weekend to play with this; otherwise it will be a couple of weeks. I'm going on the road to gain a little more experience with the TSA.
Rick "still thinking the size of the sensor will be an issue, however" Denney
Dan Fromm
9-Jul-2010, 15:39
Um, Rick, I've been thinking about y'r woes, can't tell from your text -- there's a lot of it -- where you placed the sensor. Should be in the film plane.
I did that recently when trying out a Northeast Instruments shutter speed tester with my little 2x3 Pacemaker Speed. Speeds were 1/30, 39.75 ms; 1/50, 25.02 ms; 1/125, 10.25 ms; 1/250, 5.50 ms; 1/500, 3.63 ms; 1/1000, 1.50 ms. The results are the average of six shots.
The "slow, fast" selector has nearly the desired effect, i.e., at the same slit width fast is more or less appropriately faster than slow.
FWIW, the chart on the back of the tester shows "measured speeds" that are, in the range of interest, a little shorter than 1/nominal.
Funny thing is that when I got the Graphic (ages ago) I did the equivalent of exposing E-6 test strips at every shutter speed and found that the nominal speeds gave best exposure. This with a LunaPro (= LunaSix 3) believed to be in good calibration.
I'm having problems like yours with a 2x3 Graflex; with it, the tension setting has virtually no effect on measured shutter speed, even after much fiddling to get the initial tension as low as compatible with "the shutter closes with authority."
I suspect y'r shutter is the problem, not y'r shutter speed measuring device.
Cheers,
Dan
No, it's not the shutter, or at least it's not just the shutter.
I have built the third of Steven's cascode designs. I used a 4.7K for R1, a 50K linear taper pot, a 2n3906 transistor, and the typical Radio Shack IR phototransistor. I drove it with three batteries for 4.5 volts.
For a control, I measured 1/125, to see if the shutter was basically working. That is the first attached image. As you can see, it is exactly 1/125 (8ms).
The 1/1000 speed that I measured earlier was 3.5 ms from first effect to last effect. With the modified tester it is much less than that.
In addition to the modified circuitry, I made a different snout. It is still a roll of black tape, but I pinched the side to make a very tiny tube of less than 1mm in diameter. I believe this had the majority of the effect. But now the sensor should be fully exposed to light even by the smallest opening.
If I assume rise time is no longer an issue (note that I used a pre-trigger in these measurements so that the first effect wouldn't be missed), then I'm seeing a shutter speed of about 1.5 ms, which is 1/667. That is substantially shorter than the same shutter measured with the previous tester configuration.
Note that the full voltage is never seen. This is because of the voltage divider circuit. But I did turn the sensitivity down from the maximum amount, which should mean that the sensor is not becoming saturated before it is fully exposed. The voltages are different for the different shutter speeds probably because the tester was positioned slightly differently, and with that thin snout, a small change can affect the light reaching the sensor significantly. The position changed because I had to move my hands to capture the images, wind the shutter, etc.
Dan, I am placing the sensor at the film plane, close enough. Easy to do with this shutter.
I've also attached scope traces for the 1/500 and the 1/250 shutter speeds. 1/500 is the same curtain opening as 1/1000 but the curtain speed is governed to a slower speed. Notice that for 1/1000 and 1/250, the rise and fall times are about 0.5 ms, while at 1/500 and 1/125 the rise and fall times are about 1.0 ms. That tells me that the rise and fall as depicted represent the edge of the shutter exposing the sensor, and the times above are not far off what one would expect. The proper measurement should therefore be first effect on open to first effect on close. 1/250 measures to 1/185 (5.4 ms instead of the correct 4 ms). 1/500 measures to 1/350 or so.
Note that the shutter speed error for 1/1000 (33%), 1/500 (30%), and 1/250 (35%) are all just about the same. That's about .4 stop overexposed, if it's real. Interesting that the errors are consistent on a percentage basis. But then when it gets to 1/125, it's perfect.
Now, I need to solder this thing up and put it in a box. I think it's probably good enough for most testing up to 1/250, and for sure up to 1/125, which is about as fast as is needed for me anyway.
Rick "still not exactly sure what I'm seeing is meaningful" Denney
Dag-nabbit!
I just realized that the camera I tested in the previous message was a quarter-plate Speed Graphic, which I have NOT rebuilt.
So, I went back and tested my 4x5 Speed, which I did rebuild. And the results are very heartening!
I've attached four more files, this time in order from 1/125 to 1/1000.
Again, the rise times and fall times are the length expected for the two curtain speeds, so I'm measuring from first effect on open to first effect on close.
For 1/1000, I get 1/735, about 1.27 ms.
1/500 is 2.4 ms (note different scale from 1/1000), which is 1/420. 1/250 is 4.3 ms, or 1/230. 1/125 is right on the money. 1/50 measured 1/52, and 1/30 measured 1/31.
I would expect all of these are within normal tolerance. 1/1000 is 1/3 stop slow, and 1/500 is a quarter stop slow. Everything else is within a tenth of a stop.
Rick "thinking this is probably as good as it gets" Denney
steven_e007
13-Jul-2010, 03:49
Rick, in response to your question about the output voltage being high when the shutter is open or low when the shutter is open - the cascode #1 works one way and the cascode #2 circuits works the other. You could flip either around with another transistor and resistor - but it sounds like you are doing ok without that.
Glad to hear the circuits worked ok. I can't take any credit for them - they are simple, standard designs you'd find in any electronics text book.
Here is my experience:
I wanted to know if I could get Audacity to work with the cascode #2 (didn't want to be accused of posting up stuff I hadn't tested!). Actually, I found the 0.1 mF capacitor made no difference whatever in my case.
I struggled to get sensible readings with the camera. I was at the point of blaming Audacity as being incapable of taking accurate measurements at speed. It is an audio editor after all. With too many variables and not knowing whether my camera was working reliably I took it out of the equation and substituted my spinning aluminium disk with a 3mm slot at the circumference.
I was actually quite amazed at the results. Using the cascode #2 and a 1mm pinhole over the transistor I was able to measure a shutter speed of 0.0004 seconds. That was about the limit - at this speed the 'spikes' from the transistor were displayed as flat ripples, not very distinct but just still possible to pick out. This speed is actually a bit faster than 1/2000s! I was able to measure the rotation of the disk quite accurately so I am fairly sure of my 0.0004 sec. This came out on Audacity as 0.0005s. That is 25% error, but with a 1mm pinhole on a 3mm slit we would expect an error. At 1/600s the error came out at 23%. I'm tempted to say that a 1mm pinhole for a 3mm shutter slot - for my detector / laptop sound card / configuration need a 25% correction. I didn't have time to do enough tests to be certain - but it looked consistent and encouraging. Certainly for me Audacity seemed to work quite well.
Nice to know that - now that I've spent hours building a digital counter and oscillator :mad:
It seems to make it read 1/1000s accurately you need:
1. A pinhole / snout / whatever on the detector - with a bright light to compensate.
2. Slightly better transistor circuit (cascode)
3. Work out a correction to allow for pinhole to shutter slot ratio, fuzzy penumbra, light spread, whatever (In my case 25% - maybe!)
I tried to apply the 'now-working-very-nicely' tester to the camera, but at that point it was pointed out to me by my other half, rather forcefully, that I'd spent 4 hours working on this whilst she had been decorating the living room all by herself.
So? - She doesn't help me renovate my cameras :confused:
Anyway - I think the final Speed Graphic shutter test will have to wait a bit... ;)
it is great to know there are ways to more accurately
time + test these old cameras. i have always just shrugged my shoulders
and said " looks about the same" after i compare 2 sheets of film
(1 negative shot with a behind the lens shutter and 1 negative shot with the FP shutter).
i guess my behind the lens shutters all need a good cla since my speeds are as far off as
my ill-timed FP shutter ;)
great thread!
steven_e007
13-Jul-2010, 11:52
it is great to know there are ways to more accurately
time + test these old cameras. i have always just shrugged my shoulders
and said " looks about the same" after i compare 2 sheets of film
great thread!
I think that is a very valid thing to do - at the end of the exercise it is the actual photographic exposure on the film that really matters - not the theoretically calculated exposure based on a moving slot. So - Photographic testing has to be the benchmark. My problem is that the 'old' anniversary graphic focal plane shutter has 24 seperate shutter speeds to test. So that's a whole box of film to to check each speed just once! :eek:
Dan Fromm
13-Jul-2010, 12:13
Do the equivalent of shooting test strips. One shot with the dark slide 1/4 out, another 1/2 out, and so on.
rdenney
13-Jul-2010, 15:06
Steven, I ended up making an orifice in a block of wood. I made the opening with an .040 drill bit, which is near enough 1mm. I then drilled a 1/4" hold in each side, one to hold the photo-transistor (with black tape around it to block extraneous light), and one to open up the snout end a bit. The orifice ended up being about a quarter inch thick, in front of the transistor. I used a 6-volt lantern to provide the light source, and that was pointed straight into the tester, and moved back enough (with the o-scope in free running mode) to drop the voltage a bit, ensuring that the photo-transistor was not saturating. The position of the light changed between tests, which changed how much light was getting down that orifice. That accounts for the different voltage levels between different speeds.
Of course, my main purpose was to make sure that the shutter didn't need further rebuilding or some other repair, while the camera was still sitting on my work bench. And I wanted to make sure that I at least had probabilities in my favor when committing film to the final test.
Rick "who'll probably do a final test with Fujiroid, which is usefully narrow in addition to being quick" Denney
EdWorkman
13-Jul-2010, 15:30
Except as an academic exercise why test all 24 combinations? I can't believe that there can be that many variables- for example there are four[?] slit widths which are constant no matter what the spring tension. Design of the experiment using statistical methods is beyond my experience, but I'd bet a roll of 620 verichrome that less than say 12 tests would cover it- probly 6 ??. With some darkslide trickery as Dan suggests that's only 3 or 4 sheets of film [if that, maybe]. Then there are those combinations that one probly NEVER would use. Take heart!
steven_e007
14-Jul-2010, 00:49
I don't think it is necessary to test each combination in sequence one at a time. I was just pointing out the complications of the early shutter. One problem, of course, is that because the curtain is a single long curtain with several slots in it - just winding to a different slot always changes the spring tension as well... There are always two variables changing.
Remember, my shutter has had a complete strip down and rebuild by an amateur (me!) so it may or may not now be working well.
A few tests might indicate if it was working correctly - but if you have delusions of using the zone system (!) - what you need is consistency. It doesn't matter if a particular speed is, say, 20% slow - as long as it is ALWAYS 20% slow - so it would be nice to be able to test a particular speed half a dozen times or more to see if it is the same each time.
Also, if the shutter has been used at 1/10 second for the last 20 shots (low tension, shutter cloth all nice and loose...) when you crank it up to 1/500th or 1/1000th - will you get the same speed as when the shutter has been fired a few times on high tension and the shutter is nice and tightly packed? Or visa versa?
Those are tests I'd rather do with a bit of blinking electronics than by shooting and developing film! I think it is at this stage that you either waste a lot of film or sally forth with a shutter that might - or might not, be working properly (and possibly still end up wasting film and photo opportunities!)
Once I've decided the shutter is indeed working properly and is reliable and consistent - yep, I reckon you could then calibrate photographically it with a few sheets of film.
I think that is a very valid thing to do - at the end of the exercise it is the actual photographic exposure on the film that really matters - not the theoretically calculated exposure based on a moving slot. So - Photographic testing has to be the benchmark. My problem is that the 'old' anniversary graphic focal plane shutter has 24 seperate shutter speeds to test. So that's a whole box of film to to check each speed just once! :eek:
i see what you are saying steven .. it sounds like a fun exercise :)
you might consider cutting your film in half and sticking them film holders
that way to save you some $$, it's like doing a test strip, instead of a test print...
with an old camera sometimes the springs aren't as springy and governors don't
slow stuff down like they used to ....
i just say well a 40th is the best the shutter can do at a 60th of a second,
instead of trying to get everything just-perfect. it seems to work OK for me ;)
This repair is beyond me, but I have one of these machines that I bought in fairly decent shape. I was wondering how you go about testing the shutter speed.
steven_e007
15-Jul-2010, 05:32
This repair is beyond me, but I have one of these machines that I bought in fairly decent shape. I was wondering how you go about testing the shutter speed.
How good are you with wiring up simple circuits? ;)
The overall conclusion of this thread might be that the audio program 'Audacity' - combined with a little bit of basic electronics and mounting the phototransistor behind a pinhole and you can get pretty reasonable readings up to 1/1000s.
Very, very minimal wiring and just holding the phototransistor behind the shutter - and you can only get up to about 1/200 sec.
Barry Kirsten
25-Mar-2011, 16:14
I've just come across this very interesting thread whilst searching for information on shutter speed testers.
I'd particularly like to try Steve's Improved Cascode No. 2 circuit, but am puzzled about how typical small signal NPN transistors like the ones suggested can function as photodiodes. I'm no doubt oblivious to the 'bleedin obvious', but if someone can enlighten me I'd be very interested to know.
Also, what typical output voltages should we expect with the resistor values shown and a 4.5 volt supply?
Thanks,
Barry.
Jack Dahlgren
26-Mar-2011, 07:45
I've just come across this very interesting thread whilst searching for information on shutter speed testers.
I'd particularly like to try Steve's Improved Cascode No. 2 circuit, but am puzzled about how typical small signal NPN transistors like the ones suggested can function as photodiodes. I'm no doubt oblivious to the 'bleedin obvious', but if someone can enlighten me I'd be very interested to know.
Also, what typical output voltages should we expect with the resistor values shown and a 4.5 volt supply?
Thanks,
Barry.
Q1 in the schematic is the phototransistor. It is denoted with the three slashes to the left of it. I'd guess that you would get close to your supply voltage out of the circuit, but you can test it.
Barry Kirsten
26-Mar-2011, 13:20
Q1 in the schematic is the phototransistor. It is denoted with the three slashes to the left of it. I'd guess that you would get close to your supply voltage out of the circuit, but you can test it.
Thanks Jack, but I'm puzzled because the types suggested for Q1 are all small signal transistors of the type used in amplifiers and other circuits - I can't see how they can function as light-sensitive devices. I'm obviously missing something here, as it appears from other posts that the circuit works. If anyone can put me straight on this I'd be very happy.
Thanks,
Barry
Jack Dahlgren
26-Mar-2011, 14:34
Thanks Jack, but I'm puzzled because the types suggested for Q1 are all small signal transistors of the type used in amplifiers and other circuits - I can't see how they can function as light-sensitive devices. I'm obviously missing something here, as it appears from other posts that the circuit works. If anyone can put me straight on this I'd be very happy.
Thanks,
Barry
Really? I don't see any mention in the list of what phototransistor is suggested. I think a typical phototransistor should work:
http://parts.digikey.com/1/parts-kws/photo-transistor
And yes, you need a phototransistor or something else to detect the light.
Barry Kirsten
26-Mar-2011, 15:04
Thanks again Jack. Yes, the one you suggested is an NPN phototransistor and probably would do the job.
My confusion arose from the list of suggestions for Q1 given by Steve (see page 6 of this thread) - to quote:
NPN:
2N2222
2N3904
BC548
2N4401
BC238
Thanks for the reassurance that I'm not going mad in my old age. ;)
Baz.
Ok, here's a few circuits as promised.
Those are some pretty bizarre circuits.
In all three cases the phototransistor is just acting as a switch, not controlling the bipolar in any manner.
- Leigh
When I tested shutter speeds, though, the slower speeds were over half a stop slow and the high speeds weren't.
Might I ask how you test the shutter speed?
Might I ask how you test the shutter speed?
Please read the whole thread. Many of its 8 pages go into that topic in great detail.
Rick "asked and answered" Denney
Jay DeFehr
7-Aug-2011, 12:18
With all the very impressive electronics expertise demonstrated here, has no one considered replacing the spring driven shutter with an electric motor driven one? Or, put another way, how difficult would it be to build an electronically driven/controlled F/P shutter with capabilities similar to these Graflex units?
I have 3x4 and 2x3 Graflex slrs, but I've never worked on the shutters, or even tested them, except by exposiing film. Very interesting thread!
With all the very impressive electronics expertise demonstrated here, has no one considered replacing the spring driven shutter with an electric motor driven one? Or, put another way, how difficult would it be to build an electronically driven/controlled F/P shutter with capabilities similar to these Graflex units?
I have 3x4 and 2x3 Graflex slrs, but I've never worked on the shutters, or even tested them, except by exposiing film. Very interesting thread!
The electronics seems more impressive than it actually is. The circuits we designed including very few components that could all be soldered together without any special skill. All the talk resulted in the conclusion that the light source and the interpretation of the graphical output turns out to be more important than the circuit (or the computer software), but there wasn't much complexity in the device.
Now, designing an electric motor to run in place of the spring would be more challenging. And the spring is so freaking simple and reliable that it hardly seems worth the effort.
I had though I'd use my Speed for barrel lenses and the like. But I just ordered a Sinar Copal shutter from KEH, so this experiment may have reached it's logical conclusion for me.
Of course, I'll have to test that Sinar shutter.:)
Rick "thinking Wally's (do a search) time-readout shutter tester is a good commercial solution that works as well as the Calumet tester" Denney
Jay DeFehr
7-Aug-2011, 19:29
Rick,
You're probably right. I was thinking about a project I've had in mind for some time- an 8x10 SLR, for which there is no ready-made FP shutter, which is about 1/2 the design problem. Another problem is moving a big mirror, but I thought I'd try a beam splitter mirror instead of a moving one, and use a mylar view screen to make up for some of the lost brightness there. I thought I could use a pair of small stepper motors to run the shutter, but maybe a spring mechanism would be simpler, and just as effective, but slow speeds are not the strengths of a FP shutter like the Graflex. Maybe I'll mock something up with foamcore; fixed focus, one shutter speed, and see if I can make it work.
I just ordered a Sinar Copal shutter from KEH,
Can that be installed in a Speed Graphic?
Can that be installed in a Speed Graphic?
A Speed Graphic has a shutter. But the activity I wanted to support was using barrel lenses not mounted in a shutter, and selected the Speed Graphic to support that because it also supported the separate activity of a fun, hand-held large-format camera. The Sinar shutter is for the Sinar, which is my main large-format camera.
Rick "being goofy with the Speed" Denney
I looked for the link for the overhaul / service manual at graflex.org and could not find it. Do you have a link?
Thanks in advance.
d.s.
Dan Fromm
13-Aug-2011, 14:26
I looked for the link for the overhaul / service manual at graflex.org and could not find it. Do you have a link?
Thanks in advance.
d.s.
Look on www.southbristolviews.com
In reading the instructions for the FP shutter, it says in para d), ...
With the bottom strut of the curtain 1/4 inch from the upper, inside edge of the frame, ...
This is the "O" opening it's talking about and I know what the struts are but whats the "inside edge of the frame"? and why the bottom strut?
In a later paragraph, j), it says...
Check the distance of the top strut of the open aperture from the top inside edge of the case. If this distance is not 1/4 inch from the bottom of the curtain strut to the upper, inside edge of the frame it will be necessary to blah blah...
Whats the edge of the case, and whats the edge of the frame.
What am I measuring 1/4 inch to?
I sounds like the top strut is to be 1/4 inch above the film gate, and the bottom strut is to be 1/4 inch below the film gate. I'm using "film gate" to mean the opening in the back that the light shines through to the film.
Any insite to this is appreciated. Thanks,
d.s.
Got it together and it works. 30/50 are good, and the rest are as others above in this thread have stated. Now to shoot film.
d.s.
Barry Kirsten
10-Jan-2012, 19:25
I eventually got round to doing this. I used Steve's circuit (below), using a BC107 for Q2 and a TEPT5600 phototransistor, which is sensitive in the visible range, rather than IR which seems to be the case with commonly available phototransistors. I don't know if there's any advantage to using a visible sensitive device, but it seemed right to me at the time. The trimpot is set at roughly centre of its range. Supply is 6 V (4xAA).
I use Sound Studio rather than Audacity, only because I already have it for digitising vinyl recordings; it's pretty much the same. A typical trace is shown for my 8/120 Angulon at 1/50 sec (Compur 0 shutter).
Rather than a single waveform representing the duration of shutter opening, I get two spikes representing the opening and closing of the blades. I also get the same for focal plane shutters.
The good news is that my lenses are all good enough - all within about 1/4 stop at speeds I most often use, running slower at speeds of 1/250 and above which I seldom use.
One unusual thing I noticed is that an old Symmar I have in a Prontor 00 has no click-stops on the shutter speed dial. I don't know whether this is common, I haven't seen it before. The lens presumably came with cam and camera, a 6x9 Linhof Technika which Bob Salomon kindly dated for me as 1948. Doing a couple of runs with this lens showed the variation in speeds you get when not taking care to set the speeds right on the line.
I think this has been a very worthwhile project - simple, inexpensive, and as far as I can tell an accurate way of keeping tabs on shutters which have a habit of losing accuracy over time.
This has been a very interesting thread... now you have me thinking about testing the shutter on my recently acquired Speed Graphic. Will have to take it to a mate who has a collection of oscilloscopes and see what happens.
Powered by vBulletin® Version 4.2.5 Copyright © 2024 vBulletin Solutions Inc. All rights reserved.