Building an LED light/flash

mjkzz · Post by **mjkzz** » Fri Jan 24, 2020 1:38 am

Too busy lately and finally got time to do this.

For a lot of people here who probably just want to skip to the part where a buck-boost power module is used as dimmable driver. The module is ZK-4KX and can be purchased from eBay or Amazon (not sure as I do not buy on Amazon). I have been using it for other projects and it is a heck of device, I have shorted it at 3A settings, dropped it, beat it up badly, yet it is still going. The best part is that when you dim the LED, even at lowest level, NO FLICKER :-) It is small and you can change voltage, set limiting current. . . sort of like a miniature lab power supply.

The only drawback is that it can only handle 50W of power, given that most of these stuff are made in China, I would limit max power to 60% of rated, so it is about 30W. Still not bad.

I have seen a lot LED projects here and it seems the driver is the problematic ones. So, to contribute more, here it is.

viktor j nilsson · Post by **viktor j nilsson** » Fri Jan 24, 2020 2:51 am

Thanks for the tip, Peter!

I agree that the driver is often the problem. I use a Buckpuck 1000mA driver to drive a Cree XM-L2, and I'm not entirely happy. The lowest setting is too bright and flickers, and the 1A output is limiting.

Not sure how much these problems (flicker and high intensity at lowest setting) are related to the Led itself or the driver.

Have you tried it with a Cree XM-L2 or similar 10w single-die led? How does it behave? For the price, I am willing to buy one or two of these to try it out. As you say, they seem like a miníature lab power supply!

viktor j nilsson · Post by **viktor j nilsson** » Fri Jan 24, 2020 3:49 am

I picked up two of these units to test. Thanks for the tip!

mjkzz · Post by **mjkzz** » Fri Jan 24, 2020 3:59 am

ah, you did :-)

I highly doubt XM-L2 would be any different from other LEDs, that said, LEDs are current device, as long as you do not push the current over too much and for prolonged period of time, it should work.

Older LED drivers are designed having normal lighting in mind and dimmable drivers use PWM. On the other hand, ZK-4KX adjusts output voltage with current limiting, no PWM.

viktor j nilsson · Post by **viktor j nilsson** » Fri Jan 24, 2020 4:20 am

The Buckpuck 1000mA driver is a constant current power supply that is dimmable via a trim pot.

https://www.ledsupply.com/led-drivers/b ... ed-drivers

But I still don't feel that it works as well as I would like. I'll get back with my impression of these ones in due time (which is likely in 8 weeks shipping + the time it takes for me to find time for projects).

viktor j nilsson · Post by **viktor j nilsson** » Fri Jan 24, 2020 4:40 am

Peter, how do you like using the LED as a flash? Do you get consistent exposure and color rendition between shots?

jurkovicovic · Post by **jurkovicovic** » Fri Jan 24, 2020 1:43 pm

Hi,
what kind of LED flash controller did you use?

mjkzz · Post by **mjkzz** » Fri Jan 24, 2020 4:45 pm

viktor j nilsson wrote:The Buckpuck 1000mA driver is a constant current power supply that is dimmable via a trim pot.

https://www.ledsupply.com/led-drivers/b ... ed-drivers

But I still don't feel that it works as well as I would like. I'll get back with my impression of these ones in due time (which is likely in 8 weeks shipping + the time it takes for me to find time for projects).

I do not know internals of that driver, but most old drivers uses PWM and the POT is used to adjust pulse width.

Peter, how do you like using the LED as a flash? Do you get consistent exposure and color rendition between shots?

If you notice that in my video, sometimes the power output "fluctuates on its own", but I think that is because the latency of the power supply -- I was turning the knob too fast.

However, I did observe a little bit of thermal reaction when turning LED on -- as LED gets turned on, it heats up, and I do observe that the current increases, slowly and very little, but it does. This is true with my 100W water cooled LED setup.

But I do not think the fluctuation can cause more than 1/10 of stop over a few seconds of time, and if you flash it in short duration, like 1/10s or even 1/2s, the observed thermal effect can be ignored.

So I think pulsing LEDs is better.

mjkzz · Post by **mjkzz** » Fri Jan 24, 2020 4:45 pm

jurkovicovic wrote:Hi,
what kind of LED flash controller did you use?

Oh, sorry, I should have pointed it out, it is my own creation.

mawyatt · Post by **mawyatt** » Mon Jan 27, 2020 6:59 pm

mjkzz wrote:
viktor j nilsson wrote:The Buckpuck 1000mA driver is a constant current power supply that is dimmable via a trim pot.

https://www.ledsupply.com/led-drivers/b ... ed-drivers

But I still don't feel that it works as well as I would like. I'll get back with my impression of these ones in due time (which is likely in 8 weeks shipping + the time it takes for me to find time for projects).
I do not know internals of that driver, but most old drivers uses PWM and the POT is used to adjust pulse width.

Peter, how do you like using the LED as a flash? Do you get consistent exposure and color rendition between shots?
If you notice that in my video, sometimes the power output "fluctuates on its own", but I think that is because the latency of the power supply -- I was turning the knob too fast.

However, I did observe a little bit of thermal reaction when turning LED on -- as LED gets turned on, it heats up, and I do observe that the current increases, slowly and very little, but it does. This is true with my 100W water cooled LED setup.

But I do not think the fluctuation can cause more than 1/10 of stop over a few seconds of time, and if you flash it in short duration, like 1/10s or even 1/2s, the observed thermal effect can be ignored.

So I think pulsing LEDs is better.

Peter,

Agree the pulsing keeps the temperature down. If you are using true current mode then the LED current should remain constant and not influenced by device temperature, nor supply voltage variations.

Also curious about the "sometimes the power output "fluctuates on its own", but I think that is because the latency of the power supply -- I was turning the knob too fast". Could this be caused by the power supply "hunting" trying the reach the final value?

These Cree LEDs look nice, where did you get the reflectors and heatsinks for these Cree LEDs?

All this LED stuff got me to revisit an older design that got shelved last year due to the piezo stage controller developments. This control technique is based purely on linear current mode control, so no switch mode power supply artifacts.

https://www.photomacrography.net/forum/ ... hp?t=40999

mjkzz · Post by **mjkzz** » Tue Jan 28, 2020 6:32 am

Mike,

OK, I really appreciate your signature:

Research is like a treasure hunt, you don't know where to look or what you'll find!

That is research. In terms of engineering, I think it is the same.

What I found is that as temperature goes up, not necessarily overall temperature, but I suspect diode junction temperature, LEDs tend to let more current pass, sort of like self destruction. But that is just my observation using a high powered lab power supply. This is why a constant current power supply is needed, to prevent LEDs go self-destruct by limiting current, for elongated operation of the LED

I am a bit busy to dig too much into this, but so far, what I have seem to do well, in terms of consistency of power output without using CC circuit. On the other hand, I do have a networked LED driver in mind and actually designed one last Chinese New Year, but did not have time to do so.

On the other hand, since you like to do research, I would highly recommend you to do more experiments than just theoretical design. Sometimes, what you think in theory might be very different from engineering, ie, actual hands-on build.

And if you are generous, please share it with us here.

Cheers, good work on your design.

mawyatt · Post by **mawyatt** » Tue Jan 28, 2020 8:53 am

mjkzz wrote:Mike,

OK, I really appreciate your signature:

Research is like a treasure hunt, you don't know where to look or what you'll find!
That is research. In terms of engineering, I think it is the same.

What I found is that as temperature goes up, not necessarily overall temperature, but I suspect diode junction temperature, LEDs tend to let more current pass, sort of like self destruction. But that is just my observation using a high powered lab power supply. This is why a constant current power supply is needed, to prevent LEDs go self-destruct by limiting current, for elongated operation of the LED

I am a bit busy to dig too much into this, but so far, what I have seem to do well, in terms of consistency of power output without using CC circuit. On the other hand, I do have a networked LED driver in mind and actually designed one last Chinese New Year, but did not have time to do so.

On the other hand, since you like to do research, I would highly recommend you to do more experiments than just theoretical design. Sometimes, what you think in theory might be very different from engineering, ie, actual hands-on build.

And if you are generous, please share it with us here.

Cheers, good work on your design.

Thanks, yes I've spent lots of time the lab, really enjoy the instruments especially the Tektronix Scopes and HP equipment. I've just refurbished an old Tektronix 2465 scope and a couple HP/Agilent HP34401A 6 1/2 digit DVM for Mike's Labs now that I'm retired.

Engineering is the application of science, and good engineering takes advantage of every possible parameter within the constraints of cost, time, size weight and environment....plus maybe a few others.

Anyway, where did you find those nice reflectors?

Best,

Adalbert · Post by **Adalbert** » Tue Jan 28, 2020 9:10 am

Hello everyone,

I have been using something common for some years:

https://www.ebay.com/itm/CC-CV-Adjustable-5A-Step-down-Charge-LED-Panel-Voltmeter-Ammeter-Display-Module/312123127133

But I can be used as a power supply for the Arduino too:
https://www.photomacrography.net/forum/ ... highlight=

BR, ADi

Admin edit [RJL] for URL formatting

mawyatt · Post by **mawyatt** » Tue Jan 28, 2020 12:40 pm

Adalbert wrote:Hello everyone,

I have been using something common for some years:

https://www.ebay.com/itm/CC-CV-Adjustable-5A-Step-down-Charge-LED-Panel-Voltmeter-Ammeter-Display-Module/312123127133

But I can be used as a power supply for the Arduino too:
https://www.photomacrography.net/forum/ ... highlight=

BR, ADi

Admin edit [RJL] for URL formatting

ADi,

That power supply is a switch-mode type with a 300KHz switching frequency. Great for efficiency and static use, but much too slow in response for use with a fast strobe type LED control I think.

Looks like a great little supply though!!

Best,

mjkzz · Post by **mjkzz** » Tue Jan 28, 2020 6:35 pm

Anyway, where did you find those nice reflectors?

Ah, sorry, I missed this question. I think I bought it in a building in HuaQiangBei (google it or even search it on Youtube, it is a famous electronic district) a few years back, along with a bunch of other thing, even that LED which I think it is out of production, I can not find it anymore.

And sadly, these buildings are following what happened to Frys or Radio Shack -- more and more final products, increasingly less component vendors.

My intention was to show case the power supply, how easy to use it for LEDs. I have seen so many posts here about LEDs, and in particular, the power supply. So I figure, why not share it here so others can benefit. People here are very resourceful

Here is a test shot, single shot, ISO 200 1/60s f/8 (effective f/16) 1:1 magnification. I think it is pretty good. For a 5X lens at f/2.8 (effective f/16.4), ISO 200, you can set shutter at about same speed, that is not bad.

The doll is made of ceramic, matte surface but still very shiny. Shot hand held

Setup