LED Enlarger Head

[eheldreth]

Lately I have been considering building an LED light source to replace the projector bulb in my beseler dicro enlarger. I only enlarge B&W so I am not worried about the filtering in the head and I have heard that blue light is better for B&W than white. Is this true or am I better off getting white LED's.

 

[Jorge]

Graded paper is sensitive to blue light only, VC paper is sensitive to blue and green. If you plan to use VC paper, then the white LEDs would be better, this way you can use filters to change grades. If you are going to use only graded paper, then get the blue LEDs, you will have shorter printing times.

 

[Jon King]

A slowly moving project of mine is to build an LED enlarger head, although what I am looking to do is a bit different than what you are describing.

The difference in color temperature between a bulb and an white LED is significant, not only in center temperature, but in the quality of the light. White LEDs are actually blue LEDs with a phosphor on top that absorbs blue and emits various colors of light. The overall emission is perceived as white.

Presuming you are using VC paper, your filtration will have to be recalibrated when switching to an LED source - the amounts of blue and green light in the two sources are really quite different.

The brightest LEDs normally available are at www.luxeon.com - the 5W versions are really quite intense!

 

[Bob F.]

Have a look here for some ideas on LED enlargers: http://www.huws.org.uk/

Cheers, Bob.

 

[Robert Hall]

Or you could try a bank of blue and a bank of green and vary the intensity of each to (try to) control contrast.


Hmmm...

 

[David A. Goldfarb]

Huw's page, which Bob referenced above, explains how to do just that.

 

[Bob F.]

Addendum...

This link does not have the advert banner: Dead Link Removed.

Also, I see that Dead Link Removed in the UK now stock some of the Lumileds, which is nice....]

Cheers, Bob.

 

[Nick Zentena]

For those of us that spilt filter isn't that head serious over kill? Just need to power the LEDs. Turn them on/off. No need to vary them to achieve contrast grades.

On the more complicated front. Somebody was talking about a red/green/blue head for colour work.

 

[Bob F.]

For those of us that spilt filter isn't that head serious over kill? Just need to power the LEDs. Turn them on/off. No need to vary them to achieve contrast grades.

On the more complicated front. Somebody was talking about a red/green/blue head for colour work.

Yup, for split-grade printing two on/off switches will do, but that's what you get when an engineer gives himself a project: he's going to stick a microprocessor in there somewhere... Main problem I see is the lack of a grade 5 with the LEDs - a pity, but I guess one can live with that for split-grade printing and it's pretty rare to need grade 5 in normal printing in any case.

Colour is a different problem as you can't adjust the colour an LED emits: they will only emit very close to one specific wavelength (which is a useful property for a safelight). Not sure how you would get over that problem for printing colour.


Bob.

 

[Nick Zentena]

I glanced at the website with the LEDs and the blue didn't seem that blue. 470 wavelength. Would something lower work better?

I don't really understand additive colour printing. My understanding is you make three exposures. You vary the colour by varying the relative exposure of each colour. So more red means a longer red exposure.

 

[argentic]

For those of us that spilt filter isn't that head serious over kill? Just need to power the LEDs. Turn them on/off. No need to vary them to achieve contrast grades.

On the more complicated front. Somebody was talking about a red/green/blue head for colour work.

I'm a complete dummy in electronics.

Hew uses very bright LED's. Is it possible to make individual LED's change light emission intensity (kind of rheostat)? With two pontiometers you would be in business.

Or do you have to switch on more or less LED's to get more or less blue/green? Then a greater amount of less light intensive LED's could simplify the contrast grades. You just switch on more or less blue and/or green LED's for different grades.

But maybe I'm talking bullsh...

 

[Mongo]

Is it possible to make individual LED's change light emission intensity (kind of rheostat)?

The Luxeon site mentions that their LEDs are "fully dimmable", thus a rheostat is one possibility.

 

[Jon King]

There is a 'Royal Blue' LED with a wavelength of 455nm. It should be a bit better match to paper sensitivity, but I don't know how much effect it will really have. My LED head is not at that point yet.

LED light output in lumens is basically proportional to the current you put through it. A potentiometer/rheostat will work, but will dissipate a good deal of heat and doesn't really regulate the current.

I have a simple (to me, at least), low cost (< $15) linear drive circuit for these LED's. My end goal is have a LED head for a Durst 138. I'm starting with 4 LEDs( 2 green, 2 royal blue) to figure out the best approach - whether to directly replace the bulb, much like the original poster, keeping the condensers, or make a mixing chamber as in a color head. Huw's approach just seems like too much work to me.

I'll happily share the schematic if anyone wants to experiment with the LEDs. If you can solder, you can build this pretty easily.

 

[Nick Zentena]

How many LEDs would be required to cover 5x7? Or even better something like 12x12? I keep kicking around the idea of building a 5x7 enlarger but if I'm building I'd like it bigger just in case. I had thought of using a flourscent light source but don't know how well that would work.

I figure a difusion head shouldn't be too hard to make.

 

[Bob F.]

As pointed out by jking, just varying the voltage does not really work: you need to adjust the current flowing through the LEDs; this needs a transistor as shown in Hew's circuit. You then either use the transistor to vary the current through the LED or use a PWM (Pulse Width Modulation) signal to switch the LED on/off at high speed using an adjustable duty cycle.

I'm sure a google search will throw up useful circuits, or check the data sheet for the ubiquitous 555 timer chip - IIRC it can be configured as a PWM generator - there is a dual version (556?) allowing both channels to be driven from one chip.

I think Hew went down the route he did to allow integration with his RH Designs timer/analyser. Plus, using the PIC microcontroller means he can add as many bells and whistles as he likes simply by changing the firmware. The PIC only costs a few quid/dollars so if you already have the development tools and are familiar with its programming language the additional cost is not great. If you need to buy & learn it all however then it does become far too much hassle....

Nick: Hew made 4x5" heads, so scale up from there for larger sizes, taking the area to be illuminated in to account (e.g. an 8x10 head would require 4 times as many LEDs for the same level of illumination). I think you are probably right about the colour printing - I don't do colour and I've never looked deeply in to it so I'm not at all sure if the specific LED colour wavelengths would be suitable for colour work. Someone needs to try it...

Jolly interesting, but I don't think I'll go down the route of building my own head myself. I could do with better and brighter safelighting in my darkroom however so I might have a look at the Red and Amber Lumileds...

Cheers, Bob.

 

[Jon King]

Nick - I have 2 green and 2 royal blue 3W LEDs. The LEDs were within one stop of the 150-250W bulb with #2 Kodak filter, that it replaced. Neither the bulb or LEDs are close to the light intensity I'm planning for my final source.

My plan is to make a big mixing box for a diffused source. Although bulky, the size covered and light intensity can be modified quite easily. That said, scaling Huw's design and using 3W instead of 1W LEDs should get you similar performance for 5x7, and the same for 5W and 8x10.


Bob F.- a 555 timer should work fine to PWM, but without a means to regulate the current, the circuit will just vary the effective voltage. Adding current control gets you much of the way to switching regulator I'm using an LM10 set up as a V-I converter - same level of complexity as a 555 circuit.

A red Lumiled works very well as a safelight and since it doesn't matter if they vary by 1/4 f-stop, a wall wart and resistor work very well to drive it. I've used one for a few years without issues.


OT question: Is there any way to change my name on APUG from 'jking' without rejoining? I'd like to change it to my full name, but haven't seen how to do it. Thanks!

Jon King

 

[Nick Zentena]

I spent some time reading. Additive colour printing uses the sharp cutting filters. I guess #29,47b and 59? It wouldn't suprise me if the combination of red,blue and green leds with some sort of filter in the head would work. The same thing could be used for high contrast with sticking a blue filter in the head.

Did Huw use 1w leds then? Makes me feel a little better-)) Those 5w ones are expensive if you need 200 of them. Plus the cost of a heat sink.

 

[Bob F.]

I was thinking of using PWM to switch the MOSFET, and hence the LED, on & off so that by varying the ON time during the cycle, you vary the effective brightness of the LED. By using a short cycle time, say 0.001 sec (1kHz clock) it will not appear to flash and even the shortest of exposure times will be fine (as long as it is over 1/100th of a second or so ).

One downside I can think of is that for a simple PWM generator using a 555, you may have problems at one or both ends of the duty cycle: fully on and/or fully off may not be possible - would need to experiment. This is of course more complicated than using a simple variable current source (a MOSFET and a few resistors) but I suspect it is more efficient as less heat should be dissipated in the MOSFET as it is either completely on or completely off (oh dear... that brings in the 'D' word..!) except for the very short time as it switches from one to t'other.

Anyway, you have something that works and that's the main thing!

Can't help re' the name change: I had a look myself a few weeks ago and could not find a way either. Perhaps Sean can make the necessary changes.


Cheers, Bob.

 

[cao]

I'm a complete dummy in electronics.

Hew uses very bright LED's. Is it possible to make individual LED's change light emission intensity (kind of rheostat)? With two pontiometers you would be in business.

But maybe I'm talking bullsh...

'fraid so. My understanding is that the dimming is strictly by varying duty cycle in a pulsed constant current source. That's how Huw's circuit works as I recall. The spec sheets for the Luxeons seem to require a constant current somewhere between 750ma and 1000ma for operation. I'm left wondering if his inability to get a harder contrast than 4.5 comes from some sort of intermittency effect. Would raising the frequency help. I have an old book by Zakia suggesting that intermittency effects go away as the pulse frequency goes above a certain point. Is low MHz sufficient? Any phys chem dudes out there?

 

[glbeas]

Seems a simple toggle switch to bypass the PWM when max output is desired would be the easiest route.

 

[Nick Zentena]

The spec sheets for the Luxeons seem to require a constant current somewhere between 750ma and 1000ma for operation. /QUOTE]


My understanding is 750 you get the stated light output and lifespan. 1000 you get more light but less life. OTOH I'm not sure the lower lifespan would really matter in an enlarger. But if you use less then 750 you'll get less light but longer life?

 

[Jon King]

Nick - The 5W LEDs are expensive, but cheaper per lumen. That's one reason I am looking at using a mixing chamber vs the approach Huw took. Even higher output LEDs are on the way - one goal of companies like Luxeon is to replace car headlights with a single LED. For my final version, I have a PWM ckt designed (a bit too much like real work - designing analog power converter ICs pays for my photographic efforts ) that will get the voltages and currents where they need to be, as well as dimming.

cao- it would be very interesting to find out what that frequency is. A problem with low cost digital solutions for pulsing circuits is that they are pretty slow. Analog controllers are actually superior in that aspect, as if most of us on APUG don't realize that already


Jon

 

[cao]

cao- it would be very interesting to find out what that frequency is. A problem with low cost digital solutions for pulsing circuits is that they are pretty slow.
Jon

There's also a limit on how fast we can switch the Luxeons as I seem to recall; the optical rise is about 10ns, and I have no idea what sort of shape that is. So let's say, we won't modulate above 2MHz. Still, you need fairly fast power FETs and the smarts to drive them. I have a number of ideas here, but capitalizing them is another thing altogether, and again for what sized market? Many folk seem happy enough with incandescent bulbs and plastic filters.

 

[Nick Zentena]

The problem for me is the cost seems too high still. 30 LEDs end up costing $450 at $15 per III model LED. Plus all the other parts of the head. The upside is it looks like the LEDs would last almost forever in enlarger use. No bulb replacement costs. I know I'm not the only one that would like a bigger enlarger for light usage. The ones out there are either huge. Expensive. Or have other issues. I think I'll have to wait for LED costs to come down some more.

 

[eheldreth]

Thanks for all the replies, I think the information on Huws home page is a great place to start. This should keep me busy enough this winter.