Zeiss GFL LED

Here you can discuss DIY adaptations to the microscope.
Post Reply
Message
Author
User avatar
imkap
Posts: 750
Joined: Thu Nov 25, 2021 9:44 pm

Zeiss GFL LED

#1 Post by imkap » Mon Feb 14, 2022 10:25 pm

Hi everyone,
I just made a DIY LED adaptation for my Zeiss GFL inspired by this:

http://www.geocities.ws/dushang2000/Mic ... 20GFL.html

it turned out great... It is very bright and the colour seems ok too. I thought it will be too blue but not really. The image is better now by a long shot...


I bought LEDs on eBay cheaply, they are advertised as 12V 10W, but actually have less power. I have tried one without a resistor and measured 0,72A, so it's power was 8-9W and it got burned. I have added a 2.8Ohm resistor in series (+) and now it at 9.75V and 0.4A , which is around 4W but still very bright. It has to be dimmed to half for bright field, not possible to look at. I will order a few more LEDs of these days then experiment with smaller resistors when they arrive, it should work at 10-11V I think with no problem and be even more powerful maybe 7W. Not sure about the temperature... They fit just right into the hole when you remove the corners of the aluminium base with a Dremel tool.

Maybe someone has an another LED chip recommendation, just for fun I looked at CREE and Osram Oslon, they are mostly made up to 4V voltage. I'd probably have to change the dimming module if take these but not sure how much would I benefit from it?

Used a LM2596 as a driver for dimming. Added a fuse and a connector for the resistors, so I can change without hassle... It's great i got an original Zeiss enclosure so it looks great too.
I have one more hole in the enclosure so I'll add an LED so it is obvious when it is on. Better not to leave it overnight...


By the way I did plug in the original when I first bought it, as it was quite powerful with 2.5A on 6V. It didn't work, but now I realized that the main switch wasn't working, so that might have been the problem... Still I'm not sure it would be safe to use it for prolonged periods, so it is better off waiting in the drawer...

Here are some pictures:

Image
Image Image

dtsh
Posts: 977
Joined: Wed May 01, 2019 6:06 pm
Location: Wisconsin

Re: Zeiss GFL LED

#2 Post by dtsh » Mon Feb 14, 2022 11:12 pm

Nicely done, I can appreciate the reuse of the original transformer housing.
I have a 12w LED in my AO10 and have to run it at it's lowest setting when visually observing, sometimes engaging the ND filter to reduce it even further to make viewing comfortable.
For simple viewing, I think 3w LED would probably be sufficient for brighfield, but the extra wattage is nice to have for phase, darkfield, and imaging.

User avatar
imkap
Posts: 750
Joined: Thu Nov 25, 2021 9:44 pm

Re: Zeiss GFL LED

#3 Post by imkap » Tue Feb 15, 2022 8:30 am

dtsh wrote:
Mon Feb 14, 2022 11:12 pm
Nicely done, I can appreciate the reuse of the original transformer housing.
I have a 12w LED in my AO10 and have to run it at it's lowest setting when visually observing, sometimes engaging the ND filter to reduce it even further to make viewing comfortable.
For simple viewing, I think 3w LED would probably be sufficient for brighfield, but the extra wattage is nice to have for phase, darkfield, and imaging.
Thanks, it is great to have options. Now I understand why you need a dimmer with a, microscope :mrgreen:

Hobbyst46
Posts: 4277
Joined: Mon Aug 21, 2017 9:02 pm

Re: Zeiss GFL LED

#4 Post by Hobbyst46 » Tue Feb 15, 2022 9:09 am

Congratulation again on a good job!.
The original incandescent lamp is definitely inferior. A 9W/10W LED is indispensable for phase contrast and DF at high mags. When the LED is located within the base at the same exact location of the original bulb (to provide Kohler illumination), the distance from the LED to the field aperture justifies a higher wattage than 3W, I believe. On my microscope it is a commercial LED retrofit, 12V, 10W with heat sink and high-freq PWM dimmer (not an optimal dimmer but I use ND filters anyway).

May I ask, which Cree LEDs did you buy ?

One thing, although the all-metal stand conducts heat, a dedicated heat sink for the LED is important.

User avatar
imkap
Posts: 750
Joined: Thu Nov 25, 2021 9:44 pm

Re: Zeiss GFL LED

#5 Post by imkap » Tue Feb 15, 2022 10:39 am

Hobbyst46 wrote:
Tue Feb 15, 2022 9:09 am
May I ask, which Cree LEDs did you buy ?
Hi and thanks. I didn't buy a Cree LED, just a generic "10w" 12v, available on eBay, aliexpress or elsewhere.
Like this:
https://www.ebay.co.uk/itm/353408231084 ... 2573623422

I was wondering about buying a branded LED of higher quality, but the ones I found are using lower voltage (3-4V), so I would need a different module for dimming and probably a more powerful power supply. They don't seem to be very available on the retail market. If anyone has a source for a 'better' LED of 6-12V in EU, please send...
Hobbyst46 wrote:
Tue Feb 15, 2022 9:09 am
One thing, although the all-metal stand conducts heat, a dedicated heat sink for the LED is important.
This one conducts heat to the back and the led is glued with a thermal glue, I think it should be ok I have tested it for 20mins on max and it wasn't too hot. We'll see in time.
I could shorten this aluminium thing on the front where the led is and glue a small heatsink if needed. Maybe add a small fan inside, I'll see how it works in time, for now it is good.

User avatar
patta
Posts: 402
Joined: Sun May 10, 2020 6:01 am
Location: Stavanger Norway
Contact:

Re: Zeiss GFL LED

#6 Post by patta » Tue Feb 15, 2022 11:25 am

Very nice, I'm dreaming to get something similar; but never ventured in electric engineering.

The components of the above systems are like a total of 10$; but won't be easy or cheap to get one by one.. and the right ones. Do we really need to study electronics too?
Some questions:
- why you choose a "COB" large led (large yellow emitting surface)? Shouldn't the source be as small as possible, like an halogen lamp? I'm looking at CREE XM, XP, XHP models, that have a very small emitter area, like 2x2mm;
- Part list would be: Transformer 220-??, Driver, fuse holder, Dimmer resistor, switch, extra resistor, LED, heatsink; some more?
- isn't there some "kit" with LED, heatsink, driver, transformer... already matched and assembled ? I could pay 20$... Then hack it to fit the old stand.

Hobbyst46
Posts: 4277
Joined: Mon Aug 21, 2017 9:02 pm

Re: Zeiss GFL LED

#7 Post by Hobbyst46 » Tue Feb 15, 2022 12:57 pm

patta wrote:
Tue Feb 15, 2022 11:25 am
Do we really need to study electronics too?
Actually the sub-units exist on the market and there are Cree chips of 90+ CRI. The challenge is more in the mechanics in this case than in electronics. A basic LED circuit is very simple, since the voltage drop across the LED is constant. Ohm's Law covers it all...

User avatar
imkap
Posts: 750
Joined: Thu Nov 25, 2021 9:44 pm

Re: Zeiss GFL LED

#8 Post by imkap » Tue Feb 15, 2022 7:09 pm

Hobbyst46 wrote:
Tue Feb 15, 2022 12:57 pm
patta wrote:
Tue Feb 15, 2022 11:25 am
Do we really need to study electronics too?
Actually the sub-units exist on the market and there are Cree chips of 90+ CRI. The challenge is more in the mechanics in this case than in electronics. A basic LED circuit is very simple, since the voltage drop across the LED is constant. Ohm's Law covers it all...
Well, for what I made you don't exactly need a degree. Just put the + on the + and - on -... :)
I'm far from an expert in electronics, but kind of learned how to solder, that is the main reason I'm so excited with making this simple circuit :D
I use internet and call my father or friends who are experts if I don't understand something...
patta wrote:
Tue Feb 15, 2022 11:25 am
- why you choose a "COB" large led (large yellow emitting surface)? Shouldn't the source be as small as possible, like an halogen lamp? I'm looking at CREE XM, XP, XHP models, that have a very small emitter area, like 2x2mm;
I chose it as it was easily available and very cheap and 12V 10W, didn't find a Cree or a clone with more than 3V. I had a 12V power supply and found a dimming module/step down converter for 12V. The whole thing was cheap. I can easily swap this LED for any other, as I didn't make any screw holes but glued it on with thermal glue. So when I find a perfect LED I'll make holes, screw it in and use thermal paste not glue.
Now as you told me I see that Cree makes some XHP model working on 12V, I'm quite interested but not sure where to buy them in Europe. Also after trying this one out, I now understand how much light does a 4W LED give (this one is currently working on 4W), so maybe can go with 6-8W Cree. Not to get too hot
patta wrote:
Tue Feb 15, 2022 11:25 am
- Part list would be: Transformer 220-??, Driver, fuse holder, Dimmer resistor, switch, extra resistor, LED, heatsink; some more?
I used a regular power adapter with variable voltage from 3-12V and 2.25A

Driver is actually a step down converter 3-4$ on eBay in which I swapped the trim pot for voltage regulation with a regular potentiometer of 10kOhm. So can dim the LED that way...

Fuse might or might not be needed, depending on your enclosure, you can put it inside or outside on the enclosure, for easier swapping.

Extra resistor is to limit the current as this LED doesn't really work at 12V, it gets burned quickly. It is actually closer to 9-11V. So this resistor limits the voltage to 9.75V. There are calculators for this on internet.

I'll add an extra LED so the enclosure has an ON/OFF indicator light. When the iris is closed, and/or microscope covered sometimes I don't see the light... Better don't leave this stuff unattended :mrgreen:
patta wrote:
Tue Feb 15, 2022 11:25 am
- isn't there some "kit" with LED, heatsink, driver, transformer... already matched and assembled ? I could pay 20$... Then hack it to fit the old stand.
Probably there is a kit, didn't look. But the holder is best if you make it to fit just right. Although you might be able to hack it.

dtsh
Posts: 977
Joined: Wed May 01, 2019 6:06 pm
Location: Wisconsin

Re: Zeiss GFL LED

#9 Post by dtsh » Tue Feb 15, 2022 8:24 pm

imkap wrote:
Tue Feb 15, 2022 7:09 pm
Also after trying this one out, I now understand how much light does a 4W LED give (this one is currently working on 4W), so maybe can go with 6-8W Cree. Not to get too hot
When I was bulding mine I overheated the first due to a lack of good contact between the "star" substrate and the heatsink which resulted in th LED partly burning up after about 2 hours of continuous use. On subsequent builds I have been using the common 22mm x 22mm x 10mm aluminum heatsinks instead of being attached to an aluminum block. My testing process has been to run them for a short period of time and record the temperature. If it was an acceptable temperature, I increased the time and repeated until I finally had mine running on high for 4 hours with stable temperatures so considered it good.

User avatar
imkap
Posts: 750
Joined: Thu Nov 25, 2021 9:44 pm

Re: Zeiss GFL LED

#10 Post by imkap » Tue Feb 15, 2022 8:46 pm

dtsh wrote:
Tue Feb 15, 2022 8:24 pm
When I was bulding mine I overheated the first due to a lack of good contact between the "star" substrate and the heatsink which resulted in th LED partly burning up after about 2 hours of continuous use. On subsequent builds I have been using the common 22mm x 22mm x 10mm aluminum heatsinks instead of being attached to an aluminum block. My testing process has been to run them for a short period of time and record the temperature. If it was an acceptable temperature, I increased the time and repeated until I finally had mine running on high for 4 hours with stable temperatures so considered it good.
Great, thanks for the info, I thought about shortening the block and adding a heatsink, I think for more power it might be necessary.

User avatar
imkap
Posts: 750
Joined: Thu Nov 25, 2021 9:44 pm

Re: Zeiss GFL LED

#11 Post by imkap » Sat Jun 04, 2022 8:25 pm

Hi,
I have modified the LED dimming driver, soldered a trim pot in parallel with a regular pot. So now I can adjust the maximum voltage with the trim pot (10K), and dim the LED with a regular one. No need to put the resistors in series anymore. I think it is a good idea to use a pot with exponential curve, but I don't have one. With linear dimming works in the upper range only and turned off below 6...

I found a laptop power supply, which is rated at 3.34A and 19V. So with a new driver mod I can use it for any LED I want (12V or 18V). These laptop PSUs are quite powerful and easily available. Now I'm using an 4000K 18V Cree led and it is way better than the unbranded one, heats less for the same amount of light. I keep it maxed at around 10W and it doesn't overheat on my small aluminium holder.

Unfortunately the blue CA is still present as it was on the unbranded LED. I didn't expect 4000K to be so yellow when inside the microscope, so I must use a blue filter with it. This one is CRI 70, which is pretty low. I have some other models of the same LED with a CRI 90-95 but 4000K, specs say they are a bit less bright, not sure if CRI will affect the CA.

I'm thinking maybe if I buy a 2700K similar LED with CRI of 90-95 and use a blue filter the blue CA might disappear, or at least lessen ...

Below is the graph which indicates a peak in blue wavelenght for 4000K version, I'm thinking this might be the culprit for the CA. The 3000K LED has a lower peak in blue. I found a 2700K with even lower peak in the blue range. I'm thinking this might reduce the blue CA, or maybe not :D I'd use the 2700K with a blue filter.

We have talked about my CA woes here, so this is a mixed up post from both topics: viewtopic.php?f=25&t=15405
I'd like to get a similar result with the LED as I had with a halogen...

Suggestions are appreceated as always, thanks :)
image_2022-06-04_221855683.png
image_2022-06-04_221855683.png (82.99 KiB) Viewed 2694 times

Post Reply