Converting my Omega D-2 enlarger to LED illumination

[Bernice Loui]

Heat rises, better cooling would be achieved by reversing the flow of the fan from blowing on the LED array to pulling the hot air up and out of the top.

Another way would be to figure out what light wave length(s) variable contrast paper is most sensitive to, then pick LEDs that best match the VC paper sensitivity range, apply a light output controller to each color LED (wavelength of light), put the LED's light output into an integrating sphere with the light output of this sphere into the enlarger (similar to a single light bulb). This results in a variable contrast paper light source that is more efficient with a very high degree of VC paper contrast control.

Not a simple project any more,
Bernice

[Tin Can]

That does it Bernice!

Of course you are correct, but sometimes we settle

I have an insane made by GM skunkworks Cold Lamp controller

I am cutting it out of my life literally, it is heavy, odd wiring, and unnecessary for ME

Pics after the snowstorm

PS I really appreciate your expertise!

Heat rises, better cooling would be achieved by reversing the flow of the fan from blowing on the LED array to pulling the hot air up and out of the top.

Another way would be to figure out what light wave length(s) variable contrast paper is most sensitive to, then pick LEDs that best match the VC paper sensitivity range, apply a light output controller to each color LED (wavelength of light), put the LED's light output into an integrating sphere with the light output of this sphere into the enlarger (similar to a single light bulb). This results in a variable contrast paper light source that is more efficient with a very high degree of VC paper contrast control.

Not a simple project any more,
Bernice

[Oslolens]

Few observations..

"white" LEDS (in this case 3,000K) have three individual LEDs of Red-Green-Blue inside and they are coupled to specialized phosphors to increase their light output efficiency. The trade off being the three peak light output color outputs instead of a uniform light spectrum output as would be from an incandescent, sunlight or similar light source. Humanoid eyes perceive three peaked light output from LEDs as "white" photographic materials might not record light in the same way.. This is an important consideration due to the way variable contrast B&W papers work and how contrast grade control is done with VC black and white papers..
Attachment 233792

https://www.sciencedirect.com/scienc...3107051830029X

White LEDs typically have about a 3.5 volt forward voltage to make them go.. and this voltage changes as their temperature goes up. If the LEDs are driven with a constant voltage source, the current into the LED goes up causing an increase in light output and more heat generated in the LED. This is why LEDs are driven with a constant current source to reduce this baked in problem.

"Why Drive White LEDs with Constant Current?"
Attachment 233793

https://www.analog.com/en/technical-...t-current.html

~It is also why the power unit takes time before the LEDs can be powered to illuminate them as the power source uses negative feedback and more to control the current output into the LED.

Connecting the LED array in parallel is not good. This forces the power dissipation in the power unit to be a LOT more than it needs to be and aggravates the forward voltage issues that is baked into how LEDs work. Simply wire the LED array in series to stop this problem. Check the max output voltage on the LED power unit to determine how many LEDs can be wired in series and meet the LED power unit's specifications. This might be two strings of six LEDs each instead of twelve LED in series..

12 LEDs at 3.5 volts = about 42 volts needed at the driving current of 1(?) amp..

6 LEDs at 3.5 volts = about 21 volts per 6 LED string at the driving current of 2(?) amps total into the two LED strings.

Keep in mind, current to control LED light output intensity/power, NOT voltage as this varies with each LED and operating temperature.

There will be a time delay before the power unit can make the LEDs go. There are ways to make the power unit do "near instant on" this is a design detail that required knowing how all the individual parts of the system works on their own then, designing the system to allow each individual part to work at their ideal best.


Bernice

I would worry less about the size and cost of a power unit when connecting in parallel and worry more about the effect of losing one LED (not short circuit) so the rest get let say 12% more current and soon fail, also when connecting in series, if one LED short circuit, the constant current is increased for the rest and soon fail.

Changing all LED once one of them breaks sounds like a costly design.

Any thoughts before I go totally pro-constant-90%-voltage?

Sent fra min SM-S901B via Tapatalk

[stevea]

Oslolens: With a constant-voltage supply, the biggest issue is how part-to-part Vf variation will impact current sharing. In the I/V curve Bernice provided you can see a two-fold variation in current for the same supplier's parts for the same Vf (although this is probably based on data sheet limits, not a random sample of three parts). Since light output is (roughly) proportional to forward current, this means at a minimum you'll get a lot of variation in illumination across the LED array. In the worst case, if you don't design for the maximum current, you'll burn out the low Vf parts. The current through the other LEDs won't increase because their current is (essentially) only dependent on the supply voltage (the total overall current will decrease as LEDs burn out). Note you can mitigate this with series resistors, but you then end up burning a lot of power as useless heat.

As Bernice also points out, thermal runaway can be a problem, but this can be addressed with proper thermal design and, if needed, throttling with the aid of a temperature sensor.

-steve

[Havoc]

Few observations..

"white" LEDS (in this case 3,000K) have three individual LEDs of Red-Green-Blue inside and they are coupled to specialized phosphors to increase their light output efficiency.

Might be, might not be.

"RGB" white leds have the 3 individual leds inside. And they are almost always separately connected and controlled and are bare (no phosphors).

The common "white" led these days is in fact an UV or blue led with on top of it phosphor to convert the UV/blue to visible light. They are a solid state version of the TL. Much cheaper and easier to make as there is only a single type of led involved.

Little consequence for the rest of your explanation as whatever type of "white" led you use because you end up with a series of peaks from various emitters, be it individual leds or phosphors.

[Larry Gebhardt]

Looking at your arrangement you have removed the variable condenser and replaced it with a diffuse light source. It would likely be better to move that diffuse light source up into the position of the old bulb and add the variable condenser back in. But if the projected light is even ignore my suggestion.

My first LED enlarger setup was very similar in that I switched a MeanWell power supply (could be the same model you're using) on with the timer. For reasonably long (over 1/4s) I didn't notice any issues with startup times. I had no heat issues with 12 3W LEDs on a 5x7 sheet of 1/4" aluminum with minute long exposures, but without active cooling you will want to be careful to not leave the focus on too long. The head I converted did have some passive airflow which probably helped with cooling. I think you should run some tests to see how long you can leave the LEDs running before it gets too hot.

[Bernice Loui]

Connecting up all 12 LEDs in parallel, not gonna work. LEDs do not work like incandescent "hot filament" bulbs. LEDs don't fail like incandescent bulbs.
Ponder why millions of LED bulbs in current service and vast production are wired in series connected arrays.. Another fact and reality, connecting LEDs in parallel will likely produce un-even light output and varying light output with temperature and individual LED device variations due to the solid state physics production..

Take the time to read this publication by Analog Devices (yes, these folks actually DO know something about LEDs and how to drive them properly)..

https://www.analog.com/en/technical-...t-current.html

Constant voltage driving LEDs is not a good idea, these are not incandescent filament light bulbs or why LEDs have been driven constant current since they were invented.


Bernice

I would worry less about the size and cost of a power unit when connecting in parallel and worry more about the effect of losing one LED (not short circuit) so the rest get let say 12% more current and soon fail, also when connecting in series, if one LED short circuit, the constant current is increased for the rest and soon fail.

Changing all LED once one of them breaks sounds like a costly design.

Any thoughts before I go totally pro-constant-90%-voltage?

Sent fra min SM-S901B via Tapatalk

[Bernice Loui]

Yes, and why LED choice needs to match the needs of the foto materials to be used..

Bernice

Might be, might not be.

"RGB" white leds have the 3 individual leds inside. And they are almost always separately connected and controlled and are bare (no phosphors).

The common "white" led these days is in fact an UV or blue led with on top of it phosphor to convert the UV/blue to visible light. They are a solid state version of the TL. Much cheaper and easier to make as there is only a single type of led involved.

Little consequence for the rest of your explanation as whatever type of "white" led you use because you end up with a series of peaks from various emitters, be it individual leds or phosphors.

[Larry Gebhardt]

I would worry less about the size and cost of a power unit when connecting in parallel and worry more about the effect of losing one LED (not short circuit) so the rest get let say 12% more current and soon fail, also when connecting in series, if one LED short circuit, the constant current is increased for the rest and soon fail.

Changing all LED once one of them breaks sounds like a costly design.

Any thoughts before I go totally pro-constant-90%-voltage?

Sent fra min SM-S901B via Tapatalk

The LEDs and driver ASA1000 has could be wired all in series or in a mix of series and parallel strings. The Meanwell driver he's using puts out a constant current of 1.05A and it will increase the voltage up to 60V to push that much current through. At 1.05A the forward voltage of the LEDs are just under 3V, so it is possible to put 20 of those LEDs in series and still be within the design parameters of the power supply and LEDs (based on the voltage vs current graph in the spec sheet). If you run two strings of 6 in parallel each LED would be driven with ~0.5025A and produce less light and heat and last much longer, but 12 in series will produce more light and still be well within the 1.5A limit specified by the LEDs and under the 60V limit of the driver.

The spec sheet for the driver is here, https://www.meanwell-web.com/content...PC-60-spec.pdf and from it we can see the driver has a "SETUP, RISE TIME = 500ms, 120ms / 230VAC 500ms, 120ms / 115VAC at full load". In my experience with that same driver it didn't seem that long.

For reference the LED spec sheet is at https://www.ledsupply.com/content/pd...umentation.pdf Based on the Relative Spectral Power Distribution of the 3000K LED I'd expect softer results. The neutral white has much more blue light, and it looks more balanced compared to the green.

Edit: Note my mention of 60V max is wrong. It's 48 per the spec sheet, so it definitely can't run the 20 I mentioned above.

[domaz]

Looking at your arrangement you have removed the variable condenser and replaced it with a diffuse light source. It would likely be better to move that diffuse light source up into the position of the old bulb and add the variable condenser back in. But if the projected light is even ignore my suggestion.

You don't need diffuse LED light source anymore. A single or 50 or 100W LED COB unit would be easier than using strips IMO. Only one unit to connect and one controller. And it functions closer to a point source like the original.