Seven lenses checked out at camera repair shop

[jloen]

I had accumulated quite a few modern large format lenses over the last 6 months or so, from local sales and the web auction place. These ranged from 65mm to 210 mm, made by Fuji, Rodenstock, Nikon, and Caltar. They seemed to be in nice shape but I was uncertain about the precision of the shutter timing, so I took them to a camera repair shop for testing.

The story was the same on each lens!! All seven! : Fast times were overexposed. Biggest problems were with 125 to 500, which were slow by 100% or more. Shop owner said that the mainsprings weaken with age and need to be replaced (if they are still available; many are not). Obviously this is a chronic problem even with lenses made since the 1970's. Somehow I had expected them to be a bit better than this....

Slow times (1 sec to 1/60) were good to exact.

The shop kindly charted out the times for each lens, and offered to repair lenses that they could get parts for. Rather than opting for major surgery I simply made tiny P-type labels and stuck them over the old numbers on the lenses.
The tests cost me $160 for 7 lenses (which cost me about $1500 total).
Now I know what I have here.

They also tested my Pentax and Soligor spot meters, which each were one stop off, but were easily recalibrated for $10 or so each.

So if you have any question about your gear, you can get accurate answers fairly cheaply. If you are like me, your gear is not as good as you think it is.

[Sideshow Bob]

I don't think I have ever exposed a sheet of film at a shutter speed faster than 1/60 of a second so I wouldn't worry about it (YMMV).

Gale

[eddie]

jloen,

what kind of shutters were they? i have a shutter tester and most all my copal shutters run almost right on. some of the older shutters are slow just like you suggested. the 1/10 seems to be right on on most every old shutter i have tested....most



I don't think I have ever exposed a sheet of film at a shutter speed faster than 1/60 of a second so I wouldn't worry about it (YMMV).

Gale

i shoot many fast lenses wide open...even the not so fast ones i shoot wide open. i wish that some of the bigger ones did 1/500....most are only 1/50 and they are several stops slower than i would have liked. but i can get some good speed out of the little copal 0....now if i could just get some f3 lenses in a fast shutter....:)

eddie

[Struan Gray]

At working apertures fast times should be overexposed - it's an unavoidable consequence of the way leaf shutters work. 100% at f22 is about right.

The marked shutter speed is an effective speed at full aperture. Because the shutter blades take some time to open and close, at the fastest speed the aperture is only fully open in the middle of the exposure, the rest of the time it is partially obscured by the shutter blades. The marked speed is the equivalent speed that would give you the relevant exposure if the shutter blades could move infinitely fast. At smaller apertures the shutter blades clear the opening much more quickly, and so overexposure results.

If you use leaf shutter lenses at high speeds often you either need to make a table for your usual apertures, or do the math in your head at exposure time.

[GPS]

Struan, although I agree with you about the small apertures overexposure, the OP did not state that the repair shop checked the shutters using these small apertures. I suppose that being professionals, they know how to check lenses... Unless they know how to incline amateurs to give them shutters for a repair...;-)

[BarryS]

"Obviously this is a chronic problem even with lenses made since the 1970's. Somehow I had expected them to be a bit better than this...."

Not obvious at all--It may be a chronic problem is with your shop. :) I have a shutter tester and a good number of lenses of varying ages and consistently find better accuracy than you describe (in Copal shutters). The very top speeds tend to run a little slow, but more like 1/500 is actually 1/400. The diagnosis of seven mainspring replacements sounds misguided--I don't believe it.

[Sevo]

The notion that "mainsprings weaken with age" is widespread, but entirely wrong. Springs do not weaken unless overloaded, which they'll never be in such a contained, sparingly used low-power device as a shutter, unless some previous owner once forced the shutter when stuck.

Sevo

[ic-racer]

I used to futz with my escapements, but now that I have over 25 shutters, I just make a note of any speeds that are fast (B&W Negative film user only).

[jloen]

Thanks for the interesting comments. I should have said these were Copal (five 0, and one 1) shutters, and one Seiko (on the Fuji 65 mm).
The 1/500 generally rated 1/200-1/300. The work was done at a camera repair shop near Seattle, recommended to me by a professional photographer.
Personally I rarely shoot faster speeds than 1/60 and more often 1/2 to 1 sec in the murky northwest woods.

[Sevo]

The 1/500 generally rated 1/200-1/300.

That is what I'd expect even for a perfect shutter, if you meter time rather than light flow. Shutters are supposed to be calibrated so that they perform a exposition equivalent to their nominal value. At their shortest setting the total runtime and fully-open time diverge quite significantly from the nominal value, to either side respectively (for 1/500, expect e.g. 1/250 total and 1/1000 fully open).

Shutter time testers cannot properly test for that peculiarity unless you do multiple measurements with a laser spot across the shutter radius and calculate the real speed from that and the shutter's mechanical specs - metering continuous light shots with a flashmeter and TTL probe is the only easy way short of test exposures which will get you a direct test.

Sevo

[Struan Gray]

I suppose that being professionals, they know how to check lenses...

Mostly I'd agree. But. That comment about mainsprings raises my bullshit detector.

[GPS]

Mostly I'd agree. But. That comment about mainsprings raises my bullshit detector.

I agree. But what do you want, business is slow and cash flow weak these days, heck, why not to repair the whole batch of them at $ 99 a piece...?;)

[Struan Gray]

To be fair, they may just not have much LF experience. For example, some of the F-series Hasselblad lenses do have a known problem with the tensioning of the stop-down mechanism. But an LF shutter mainspring getting weak through mere age is pure foboff waffle at best.

[jloen]

I agree. But what do you want, business is slow and cash flow weak these days, heck, why not to repair the whole batch of them at $ 99 a piece...?;)

Not quite. Price to repair each was $138 plus tax...

Personally I think this shop is honest and reliable. Older, experienced staff, no shortage of business either. (I run a repair shop myself in a different field, so I'm definitely watching and thinking about integrity when I get something repaired FWIW).

Thanks for the interesting comments. When you pay $400 for a Nikon lens you expect the 1/500 to come in on target, not at 1/200. I'm still a bit shell shocked! :eek:

[rguinter]

At working apertures fast times should be overexposed - it's an unavoidable consequence of the way leaf shutters work. 100% at f22 is about right.

The marked shutter speed is an effective speed at full aperture. Because the shutter blades take some time to open and close, at the fastest speed the aperture is only fully open in the middle of the exposure, the rest of the time it is partially obscured by the shutter blades. The marked speed is the equivalent speed that would give you the relevant exposure if the shutter blades could move infinitely fast. At smaller apertures the shutter blades clear the opening much more quickly, and so overexposure results.

If you use leaf shutter lenses at high speeds often you either need to make a table for your usual apertures, or do the math in your head at exposure time.

Struan: A nice analysis. At slower exposure settings, obviously the speed of the shutter blades opening and closing has less of an effect on the total time the shutter is open than at fast speeds. So I'm wondering, at what speed setting would you recommend making appropriate mental adjustments for modern shutters to prevent overexposure? Thanks for you comments. Bob G.

[HMG]

This seems fairly common to me. I have no empirical evidence but it seems that most of the time when someone shows the results of shutter speed tests after CLA, the fast speeds are slower than marked. I don't know if this is just a "bad" sample (i.e. the ones I remember), the test process, or actual speeds.

For example:
http://www.largeformatphotography.info/forum/showthread.php?t=60629&highlight=cla+flutots
http://www.largeformatphotography.info/forum/showthread.php?t=49750&highlight=cla+flutots

[Doremus Scudder]

Unless you need to use the faster speeds for some reason, just tape a sticker with the equivalent correct speeds for each lens on the lens board and use those when calculating exposure.

It is typical for older shutters to have slower faster speeds. Modern ones, however should not be 100% off. That said, it is important to test the shutter speeds at an average working aperture. If your tech tested wide open, then the speeds may be close to correct at f/22. Maybe a quick re-test at your favorite aperture is in order.

Best,

Doremus Scudder

[GPS]

Alternatively, buy yourself a Calumet shutter tester (I think it costs less than what you paid for your test) re-test your shutters at will in all of their aperture settings and you will know where you are - with the repair center and your lenses.

[Eric Woodbury]

I have the test results on the 14 LF lenses I use. Mostly Copal #0 and #1, but a couple of #3. They are all modern lenses, they all look new, and half or more I bought new. One of my #3 is right on the money, and the other is 125% slow at the 1/125. It is a 1970s lens and was just CLA'd. It is a speed I never use, so it doesn't matter and it is consistent, so I could. They are all within +/- 10% below a 1/30.

Remember that a 1/3 stop is a 26% change in exposure. Maybe with color 26% is a lot, but not usually.

I had a 121mm SA for years. Speeds were all weird on that, but I just put a sticker on it and used it as is. They never changed. It was fine.

If I were you, I wouldn't have any of the shutters worked on and I would get my own shutter tester. $160/7 is a little rich. Only have them worked on when there are different every time or when they get cold.

[Struan Gray]

So I'm wondering, at what speed setting would you recommend making appropriate mental adjustments for modern shutters to prevent overexposure?

Bob, if I needed any real accuracy I would measure the intensity of a focussed lamp as a function of time for all shutter speeds and apertures of interest. The photodiode and sound card shutter testers often recommended are not really good enough, but it's not too hard to put together a better system with a fast photodetector and a real usb datalogger or a storage oscilloscope. Such a setup would allow you to also test how repeatable the shutter is.

The whopping sample of two shutters I tested had intensity-time curves at full aperture which looked very much like a triangle. In this case, there is a simple geometrical proof that if the exposure is correct for the marked top speed at full aperture, the same speed at smaller apertures will have an effective exposure of:

-1 stop: 1 + 1/2 (+0.6 stops)
-2 stop: 1 + 3/4 (+0.8 stops)
-3 stop: 1 + 7/8 (+0.9 stops)

The pattern is clear (it continues 1+15/16, 1+31/32, etc), and the series converges to a factor of two, i.e. one stop overexposure. At half the highest speed the additional exposure is halved for each stop. At a quarter the highest speed it is divided by four, and so on.

There are caveats and small corrections for lenses where the aperture blades are not fully clear of the shutter opening at the widest marked aperture (true for some of my wide angles and process lenses), and for the exposure of moving objects or out-of-focus background highlights. But the corrections are small, and if you really care about them it's best to measure your actual lens and shutter.

[rdenney]

Bob, if I needed any real accuracy I would measure the intensity of a focussed lamp as a function of time for all shutter speeds and apertures of interest. The photodiode and sound card shutter testers often recommended are not really good enough, but it's not too hard to put together a better system with a fast photodetector and a real usb datalogger or a storage oscilloscope. Such a setup would allow you to also test how repeatable the shutter is.

Struan, this got me thinking, as I've been trying to understand the photo-transistor tester that I have.

In mine, current is pulled through the transistor at 1.5 volts using a 3.9K resistor. I do not know what diode I have, and as I peruse a few data sheets in a google search, it would seem that rise and fall time depend on current flow, which results from the brightness of the light and the resistor. Assuming the light is bright enough to bring the diode to saturation (which seems to be the case with what I have, even measuring light on the ground glass), the rise time is pretty crisp to the saturation point. The rise time in my test of the Speed Graphic focal plane shutter is maybe 300 uS, and the fall time to negative saturation is perhaps 150 uS. These are long enough to introduce error at 1/1000 shutter speed, but I would think enough to render the measurement unuseful. There is, on several data sheets, also a delay factor before the diode begins to flow current, but I have no way of measuring that without having enough test equipment so that I wouldn't need the diode tester in the first place.

I'm assuming there is some process latency in the sound-card and recording software, but I would think it necessary to be quite consistent in order to render music reasonably with frequencies up well over 10KHz. At 44KHz sampling rate, I get at least ten samples during the rise time only in my measurement, so that seems more than adequate. My old Tektronics scope isn't a storage scope so I can't use that to bypass the sound card and processor, but I'm not sure it's bandwidth is all that much better in any case. (Hey, I paid $50 for it at a hamfest.)

Thus, I'm back to that diode response delay before the rise begins. The datasheets I've seen show no delay on fall, so if the delay was significant it would shorten the measured opening rather than lengthening it.

I'm just wondering what troubles you about the diode testers--perhaps it's just uncertainty.

Rick "diodes are cheaper than film" Denney

[Struan Gray]

Rick, the photodiode bandwidth isn't really a problem any more. It is possible to buy a diode whose rise-time is too long, but it's unlikely if you just stick a pin in the photodiode page of most component suppliers.

There are two problems with photodiodes, but the largest is with the computer sound card. Sound cards are AC coupled and they reject DC. Worse, there are no standards for the low frequency and DC rejection circuits, and they are rarely documented, so you cannot be sure which bits of your signal are missing. A step voltage fed into the sound card will produce a spike followed by a decay back to zero, and both the height of the spike and the rate of decay to zero are set by parameters you don't have.

Photodiodes are highly linear when run in reverse-bias mode, but in simple circuits there will be nonlinearities and drift with temperature, drive current and voltage. Perhaps a worse problem is that cheap fast photodiodes acheive their speed by having small photodetector areas, which means they saturate quickly and are noisy.

Put it all together, and you get a detector which is perfectly good enough for measuring the length of the light pulse at the focal plane of a LF lens, but which is going to be a royal pain in the backside if you want to measure the time-dependent light intensity with any accuracy. The spanner I threw into this thread was exactly the point that just measuring the start and stop times can lead to easily-misinterpreted results.

Perhaps the biggest problem with the simple photodiode circuit is that you can now buy much better devices for very little money. There are now very nice combined photodiode and amplifier components which address the drift and linearity problems with a raw photodiode. Some also include a voltage-to-frequency converter, so that you can avoid the DC rejection shenannigans of your sound card too.

One supplier whose devices seem to be in all my local components catalogues is TAOS (www.taosinc.com). With a battery (or a USB power cord) and one of their TSL235 sensors you can put a square wave into the sound card whose frequency is directly proportional to the light intensity. Use one of the programmable chips in the TSL230 series and you can match the full dynamic range of the device to the input frequency range of your sound card. The last (best) step would be to interface one of these chips to a microcontroller with a counter/timer. The widely-documented (and slightly cult-like) Arduino is probably the best bang-per-buck, but companies like National Instruments (www.ni.com) also do reasonably priced USB-based dataloggers.

When I looked at this I used a large area, NIST-traceable photodetector with a known and calibrated bandwidth. (This one (http://newfocus.com/products/?navid=3&theView=modelDetail&productLineId=3&productGroupId=207&modelGroupId=1053&modelRangeId=1053&modelId=1401) from New Focus, if anybody cares). Overkill for photographers, but I just happen to have one lying about :-)

[rdenney]

Perhaps the biggest problem with the simple photodiode circuit is that you can now buy much better devices for very little money. There are now very nice combined photodiode and amplifier components which address the drift and linearity problems with a raw photodiode. Some also include a voltage-to-frequency converter, so that you can avoid the DC rejection shenannigans of your sound card too.

The DC rejection issue explains the shapes I'm seeing, indeed, though the steps still seem to be clearly visible. I hadn't thought of that. Feeding it with high-frequency AC might solve that problem, but probably would lead to aliasing with the rise time of the diode.

Maybe on some rainy afternoon, when you sitting around bored, you could sketch up a schematic of something that would be cheap and functional. I can solder and build things, and I can figure some things out, but I'm not an electronics designer at any level. I found one on the Internet that used a programmable logic controller, but that seems a bit ambitious for those unlikely to have the development environment.

Rick "who never uses the fastest speeds anyway" Denney

[Photomagica]

The notion that "mainsprings weaken with age" is widespread, but entirely wrong. Springs do not weaken unless overloaded, which they'll never be in such a contained, sparingly used low-power device as a shutter, unless some previous owner once forced the shutter when stuck.

Sevo

Sevo,

Mainsprings do weaken over time.

Hasselblad specifies the replacement of the mainspring in their V series lenses at regular service intervals because it does weaken after repeated cycling and the high speeds drift slower. This is true of other shutters as well. I have restored the high speed performance of Copal shutters by mainspring replacement. Some older Compur shutters have a series of notches that engage one the end of the mainspring. The idea is to move to a higher tension notch as the spring looses strength.

Mainsprings are under a heavy load for their size. Thus from time to time I do see one that has broken.

(Hasselblad shutters (Compur/Prontor) are stored cocked. There is no evidence this causes faster or added deterioration of the mainspring. Copal mainsprings, typically left uncocked, show a very similar rate of deterioration.)

I did not want your remark here to stand unchallenged because it inappropriately casts doubt on the advice of responsible, experienced repair technicians who have seen the results of mainspring replacement, and the advice Hasselblad.
Photomagica

[Sevo]

Mainsprings do weaken over time.

Hasselblad specifies the replacement of the mainspring in their V series lenses at regular service intervals because it does weaken after repeated cycling and the high speeds drift slower. This is true of other shutters as well.

Hasselblad used Compur and Prontor shutters. I own service manuals for most the former, and some of the latter. Neither demand any routine mainspring replacement. Nor do they need them - I haven't ever needed to replace a main spring on one of my shutters, or had such a replacement on the bill of shutters that I had to send out.

[Struan Gray]

Maybe on some rainy afternoon, when you sitting around bored, you could sketch up a schematic of something that would be cheap and functional.

Just what I need - another project :-)

I think the biggest problem is software. There are any number of freeware programs which will record signals from the sound card and measure a time interval. Software to measure, plot and analyse real voltages is more specialised, which means less available and more platform specific. Freeware solutions like the Arduino require you to at least be able to compile simple programs.

My kids are getting to the age when electronic projects start to make sense, and the only part of computers which ever fascinated me was interfacing to the outside world, so perhaps, one day soon, I will get round to it. No promises though - I'm still trying to find time to scan my backlog of photos.


Struan

[Paul Metcalf]

Because the shutter blades take some time to open and close, at the fastest speed the aperture is only fully open in the middle of the exposure, the rest of the time it is partially obscured by the shutter blades. This has long made me wonder (and wonder is only as far as it's got!) what the effect of this is on DOF, etc. If the aperture is only fully open (at its particular setting) for a very brief period, the film then is only fully exposed at that aperture for a very brief period. It's exposed at a continuum of ever increasing apertures from zero to the particular full-open setting, then back again to zero. What's that doing to DOF and OOF areas, etc. etc? I'd be curious to see what differences in the final image there would be with the same optics taken at different set apertures with a typical leaf shutter, instantaneous open/closed shutter (if there is such a thing) and maybe a roll-film shutter like on a speed graphic. But the curiosity is not all that great...:p

[rdenney]

Mainsprings do weaken over time.

Let's think about this a bit.

Steel is highly elastic, in that it returns the energy stored in it linearly, with nearly no loss through absorption. That's what makes it good for springs. The linearity between stress (load) and strain (how much the spring winds up) is maintained up to the point where the steel deforms permanently. This point is the "yield strength" of steel--the point at which the material finally yields to the load being applied. When it deforms permanently, it has become "plastic".

Some steels, of course, have a higher yield strength than others, depending on alloy, heat treatment, work hardening, and so on. Coat-hangers have very low strength, but music wire of the same thickness can be flexed into a pretzel and still pop back straight.

Steel has another characteristic, known as fatigue. Fatigue results in the formation of cracks (sometimes microscopic at first) that then begin to travel through the material as it is exercised. Steel has to be stressed to close to its yield strength many times before fatigue sets in, unless there were already high internal residual stresses left over from manufacturing processes. Stress relieving usually addresses those, and springs are routinely stress relieved during manufacture. Steel loaded into its plastic state fatigues quickly, as anyone who has bent a coat hanger back and forth can attest.

Here's the point: Unless steel has been stressed to the point of fatigue, and assuming it hasn't been stressed to the point of yielding, the stiffness of the steel will remain unchanged over time, even very long periods of time. Steel does not "go soft."

An old hobby of mine is rebuilding antique clocks. The steel mainsprings in my 150-year-old 8-day clocks still run those clocks for 8 days without any trouble.

Usually, when springs break or lose their stiffness, it's because of one of several causes: 1.) the spring rubs on something, causing a notch to be worn, that acts as the root for crack formation, 2.) it was poorly manufactured and is suffering from residual internal stress and cracks are spreading (in which case it will break soon and cannot be repaired), and 3.) the spring was forced into the plastic state by the application of one overwhelming load, which reduces the preload force that it applies in service. I suspect the latter is the usual cause of spring weakening in camera shutters, and the thin, long springs are more subject to this than short, stout springs.

Mostly what causes shutters to run slow is friction, not weakness in the spring. When the dirt, gummy or improper lubrication, and wear is addressed, things seem to speed right back up.

I don't know much about electronics design, perhaps, but I do know about steel.

Rick "a (sometimes) civil engineer" Denney

[engl]

This has long made me wonder (and wonder is only as far as it's got!) what the effect of this is on DOF, etc. If the aperture is only fully open (at its particular setting) for a very brief period, the film then is only fully exposed at that aperture for a very brief period. It's exposed at a continuum of ever increasing apertures from zero to the particular full-open setting, then back again to zero. What's that doing to DOF and OOF areas, etc. etc? I'd be curious to see what differences in the final image there would be with the same optics taken at different set apertures with a typical leaf shutter, instantaneous open/closed shutter (if there is such a thing) and maybe a roll-film shutter like on a speed graphic. But the curiosity is not all that great...:p

I made a post showing what the illumination from a leaf shutter looks like in the following thread:
http://www.largeformatphotography.info/forum/showthread.php?t=62197

There is also a reply by Struan Gray showing what it can look like in practice, affecting the look of OOF areas.

A DSLR seems to me like the perfect way to test a shutter, at least at higher shutter speeds. Dedicated testers or home built devices test at a single point, and seem inaccurate due to electrical component properties (rise/fall time etc.). Imagine using a single point tester away from the center, the shutter time reported will vary greatly between different points at the same distance from the center. Look at the image in the linked post, imagine testing the shutter either at one of the bright lines or a point in the much darker area adjacent to the line.

If I had more spare time than I do Id consider writing a piece of software for analyzing shutters based on RAW or even JPEG images. It would be possible to show the effective shutter speed at all apertures. I dont have the time though, and I doubt anyone would care to use it given that it only applies at rarely used high shutter speeds, and the small exposure variation would only be visible on slide film.

[Photomagica]

Let's think about this a bit.

. . .
Here's the point: Unless steel has been stressed to the point of fatigue, and assuming it hasn't been stressed to the point of yielding, the stiffness of the steel will remain unchanged over time, even very long periods of time. Steel does not "go soft."

An old hobby of mine is rebuilding antique clocks. The steel mainsprings in my 150-year-old 8-day clocks still run those clocks for 8 days without any trouble.

Usually, when springs break or lose their stiffness, it's because of one of several causes: 1.) the spring rubs on something, causing a notch to be worn, that acts as the root for crack formation, 2.) it was poorly manufactured and is suffering from residual internal stress and cracks are spreading (in which case it will break soon and cannot be repaired), and 3.) the spring was forced into the plastic state by the application of one overwhelming load, which reduces the preload force that it applies in service. I suspect the latter is the usual cause of spring weakening in camera shutters, and the thin, long springs are more subject to this than short, stout springs.

Mostly what causes shutters to run slow is friction, not weakness in the spring. When the dirt, gummy or improper lubrication, and wear is addressed, things seem to speed right back up.

I don't know much about electronics design, perhaps, but I do know about steel.

Rick "a (sometimes) civil engineer" Denney

Rick,
Thank you for the thoughtful and useful discussion. While in many situations, like your clock, springs do not loose strength over a long period of time, in other situations they can, even without the application of overwhelming force - force outside the basic design limits of the spring. In automotive suspension spring design it is simply called "sag". There is a good article about it from the automotive perspective here:

https://tspace.library.utoronto.ca/bitstream/1807/12732/1/MQ45601.pdf

The same effect applies to the tiny mainspring in a camera, which is called upon to be wound with considerable force and then to release its energy with suddenness many times over. This causes the camera equivalent of "sag". Thus weakened, the spring stores less energy and cannot achieve the highest shutter speeds.

As you note, in most cases, the cause of slow speeds is just a dirty shutter. However when the dirt and thickened lubricants have been taken care of, the shutters I see are frequently slow at the very highest speed(s). In this situation replacing the mainspring returns the high speeds to normal.

Of course for most of us the top two speeds don't matter. However if we are using flash with sunlight or bright studio lights or stopping action, these speeds do matter.
Photomagica

[rdenney]

In automotive suspension spring design it is simply called "sag".

Sorry, I didn't read the article. But I do know what causes sag in old car springs: Impact stresses beyond the yield strength of portions of the metal where residual stresses are still high. Residual stresses occur when the formation of the part left parts of the material fighting other parts due to differential cooling, or the way it was shaped, or the lack of stress relieving after welding, and so on. This is not a process that should be taking place in a shutter, in which the spring is not, or should not be, subject to impact forces at maximum stress.

Rick "wondering if an actual materials engineer has studied the mechanical processes in a shutter in detail" Denney

[Photomagica]

Rick,
Hmmm - the author of the automotive spring investigation points out:

"During service, springs are subjected to fatigue loading which is normally lower than
the yield strength of the spring material. However, the design and mount is such that yielding anywhere in the spring is prevented. Moreover, the springs are given a preload before being mounted on the vehicle. This process called scragging or presetting, in addition to strain hardening, gives the necessary plastic accommodation. Nevertheless, with service time, springs progressively deform or sag."

So this deformation takes place without exceeding yield strength. This is what is going on in camera mainsprings and why, in practice, we need to replace mainsprings to restore high speeds.
Cheers,
Photomagica

[rdenney]

Rick,
Hmmm - the author of the automotive spring investigation points out:

"During service, springs are subjected to fatigue loading which is normally lower than
the yield strength of the spring material. However, the design and mount is such that yielding anywhere in the spring is prevented. Moreover, the springs are given a preload before being mounted on the vehicle. This process called scragging or presetting, in addition to strain hardening, gives the necessary plastic accommodation. Nevertheless, with service time, springs progressively deform or sag."

I disagree with this description, actually. I just wrote a long reason why, but it's not really relevant to the forum or this thread, and I erased it. Let's just say they missed the effects of internal residual stresses at the grain level.

But you see what you see. The issue is to what that effect can be attributed.

Rick "creep in steel would be bad" Denney

[tom thomas]

I've been following this thread about shutter speeds/tolerances on our old lenses (and newer ones for repair) and thought I'd add what I found in the Graflex Shutter repair manual for speed tolerances. Graflex themselves stated that these are typical Operation Ranges of Shutter Speed Settings when repairing the shutters. I didn't include all speeds.

Speed Setting Min Millisec's(speed) Max Millsec's (speed)
1/400 1.75 (1/570) 3.3 (1/303)
1/50 16 (62.5) 24 (42)
1 sec 800 (1.25) 1200 (0.8)

The chart in the repair manual just lists the millisec's, I calculated the speed by dividing 1000/speed in millsec. Hopefully I'm close.

As you can see,the tolerances were generous so the photog had to know his camera and lens.

I've forgotted the site where I found the manual but it's the same one I was referred to seeking repair helf for the top mount rangefinder on my Graflex 4X5 Pacemaker. Great manuals there too.

Tom Thomas

[tom thomas]

Geez, that's hard to read when tabs don't work.
Translate as 1/400 is from 1/303 to 1/570th of sec
1/50 as 1/42 to 1/62 of sec
1 sec as 0.8 sec to 1.25 sec.

[Arne Croell]

I've been following this thread about shutter speeds/tolerances on our old lenses (and newer ones for repair) and thought I'd add what I found in the Graflex Shutter repair manual for speed tolerances. Graflex themselves stated that these are typical Operation Ranges of Shutter Speed Settings when repairing the shutters. I didn't include all speeds.

Speed Setting Min Millisec's(speed) Max Millsec's (speed)
1/400 1.75 (1/570) 3.3 (1/303)
1/50 16 (62.5) 24 (42)
1 sec 800 (1.25) 1200 (0.8)

The chart in the repair manual just lists the millisec's, I calculated the speed by dividing 1000/speed in millsec. Hopefully I'm close.

As you can see,the tolerances were generous so the photog had to know his camera and lens.

I've forgotted the site where I found the manual but it's the same one I was referred to seeking repair helf for the top mount rangefinder on my Graflex 4X5 Pacemaker. Great manuals there too.

Tom Thomas

Actually, except for the highest speed, that's only 1/3rd of a stop (factor of 1.26 in timing) or less of a deviation from the set value. If I remember correctly, 1/3rd of a stop is the ISO tolerance for shutters, aperture settings, and film speeds. The maximum in that list is 1/2 stop, on the fast side of 1/400th (somewhat unlikely to happen), and on the slow side of the 1/400th it is a factor of 1.33, somewhere between 1/3rd and 1/2 stop.

[neil poulsen]

I never rely on nominal speeds written on the shutter. Years ago, I purchased a Calumet shutter tester, and I use it to determine actual speeds. From these speeds, I create a table giving me f-stop corrections.

What impressed me in making these tests was the consistency of modern shutters. A quite large proportion of speeds were within 1/10th stop from high to low and several were within 1/20th stop from high to low. To determine consistency, I take about six readings per speed.

At the time of purchase, my tester was about $90. In the last year or so, I've seen these testers reach upwards of $200 on EBay.

[Lynn Jones]

I've checked hundreds of new between the lens shutters and the only brand that I've encountered which routinely were 1/500th when set at 500 were the Seikosha Rapid units. To begin with, the ansi spec for 500 is +/- 30% of indicated. Most of the recognizible shutters don't make that.

Spring technology in the western world is truly excellent going back to the 1930's, your repair tech is full of beans, they either last or break.

Well used shutters are usually poor in speed due to dirt or lack of proper lubrication. Exercising shutters is the best thing you can do, once a month trip each of the slow shutter speed (100 and slower) 10 times. It will last longer than you will.

Lynn

[Ernest Purdum]

When you contemplate all the possible causes of exposure error, it's a wonder that we are ever able to make correct exposures (more or less).