Fast LED Strobe for Macro

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Fast LED Strobe for Macro

Post by mawyatt »

Peter posted a very interesting thread on building a LED Light/Strobe here:

https://www.photomacrography.net/forum/ ... hp?t=40977

After reading this I looked into the current mode LED driver I was working on last year when I got sidetracked by the piezo electric stages and decided to design a PCB for it.

The design is developed to drive an LED like the Cree CXA1304 that Peter mentioned, or other type LED, with a current mode pulse rather than the usual voltage mode, so the LED current doesn't depend on the supply voltage, LED temperature, LED device variation or LED voltage drop.

We've tried to keep the components on the low cost side and have a small ~100mm square PCB that supports 4 LEDs as design goals. To keep things simple, no computer nor micro controller is utilized and everything works from a single supply. The pulse timing is created with the popular NE555 timer and controlled by a potentiometer, the 555 voltage pulse is converted into a current that drives the LED.

The 555 is triggered from the negative edge of a typical strobe/flash which is connected thru the standard 3.5mm connector. A jumper sets the RANGE to either 100us to 1.2ms or 1ms to 12ms and controlled by a pot, so total range of 100us to 12ms. Other ranges can be easily created by just changing the charging capacitors (0.1uF & 10uF) or/and potentiometer (10K). The 555 output drives a PMOS transistor which supplies a 0 to 5.00V pulse to the current mode controlling amplifiers thru a scaling pot which is the LED current range. These amps are arranged to drive a hi current NMOS device (30A) to sink LED current. Current range can be set with the pot from 0 to 2A, or can be changed with a different divider resistor with the pot and/or a different current sense resistor. Two 2 ohm 2W current sense resistors are shown which gives a scale factor of 1A per volt LED current, if you remove one of these resistors then the scale factor changes to 1/2A per volt. Of course different current sense resistor could be used. The pot output scales for 0 to 2.00 volts, by changing the divider resistor this can change. For example changing R2 from 15K to 10K causes the max pot output to be 2.5V and thus the LED current to 2.5A amax with current sense resistor shown. Changing R2 to 25K causes the peak LED current to be 1.25A.

The design support two banks of two LED current drivers, the 2 banks have independent current scaling. Each bank is controlled from a single pot but has independent current sense resistors and amplifiers, so they can be independently set within a bank by changing the sense resistors. So 4 controlled outputs which can be combined for higher currents by simply tying the outputs together.

The controller also has a continuous mode (jumper) to allow constant LED current. The PCB layout supports different pots, from single turn to precision muti-turn and has provisions for using a TO252 or TO220 NMOS device, the TO220 is recommended for use with a heatsink for higher LED currents. The LED voltage is supported to 25V (can go higher with different components) and also supports 12VDC for the AXP1304.

This is a Quad Output very flexible current mode LED controller that supports a wide range of pulse widths, LED currents, is low cost and triggered by standard strobe triggers.

We'll be ordering some custom PCBs very soon, so PM if you are interested.

Edit: Updated to Verison 2.

Best,

Image
Image
Image
Image
Last edited by mawyatt on Tue Jan 28, 2020 8:55 am, edited 2 times in total.
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

JKT
Posts: 420
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

I'm definitely interested.

The question is how many lumens can you push out of those LED:s? I've estimated that I'd need roughly 400 lm*s for basic exposure. This works roughly up to 2:1. Higher magnification needs more light and shorter times, which is a bad combination.

I couldn't find any information on APX1304.

kaleun96
Posts: 270
Joined: Sat Oct 20, 2018 3:47 pm
Location: Stockholm, Sweden

Post by kaleun96 »

Hey Mike,

I just got into PCB manufacturing myself and have a question about the small holes in the ground plane. I can't work out their purpose. They don't look like vias, I wouldn't think they're required for thermal regulation, I assume they're not helping to reduce impedance or interference with some signal, etc. But I'm quite new to all this so could be wrong about any of those :D
- Cam

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

JKT wrote:I'm definitely interested.

The question is how many lumens can you push out of those LED:s? I've estimated that I'd need roughly 400 lm*s for basic exposure. This works roughly up to 2:1. Higher magnification needs more light and shorter times, which is a bad combination.

I couldn't find any information on APX1304.
Typo, CXA1304. Here's the data sheet.

https://www.cree.com/led-components/med ... XA1304.pdf

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

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

kaleun96 wrote:Hey Mike,

I just got into PCB manufacturing myself and have a question about the small holes in the ground plane. I can't work out their purpose. They don't look like vias, I wouldn't think they're required for thermal regulation, I assume they're not helping to reduce impedance or interference with some signal, etc. But I'm quite new to all this so could be wrong about any of those :D
These are thru vias that connect the backside planes to the frontside planes, often referred to as "Ground Planes". They serve to decrease the resistive and inductive impedances, as well as thermal.

These can serve to isolate signals. In the case of the above PCB there are 4 planes, 2 on top and 2 on bottom. The top planes are connected to the bottom planes with thru vias and normal thru holes for components. The result is two planes above and below that are selectively connected together to form "common ground". If you look at the schematic, note the GND and GND-LED labels, these signify the high current LED ground reference and the normal ground reference planes. Zero ohm resistors R6 and R23 serve to selectively tie these planes together.

High performance precision analog design, including PCB, requires controlling the ground return currents so they don't corrupt the sensitive analog signals of interest. This becomes somewhat of an "Art" as well as "Science" in analog design, although the science fundamentals govern, many don't have access to supercomputers to analysis the effects, so they rely on the "Art" aspect. :roll:

Many Millions of $ and careers have been spent on these "Ground Effects". We spend lots of time analyzing and applying these at the chip level for digital to analog isolation for the first RF/Analog System on Chip back in 2000 which was a massive chip ~18mm by ~24mm, 3/4 digital (FPGA) and 1/4 sensitive RF/Analog (uW Tx/Rx). This research led to a joint patent with IBM on the isolation techniques.

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

kaleun96
Posts: 270
Joined: Sat Oct 20, 2018 3:47 pm
Location: Stockholm, Sweden

Post by kaleun96 »

mawyatt wrote:
kaleun96 wrote:Hey Mike,

I just got into PCB manufacturing myself and have a question about the small holes in the ground plane. I can't work out their purpose. They don't look like vias, I wouldn't think they're required for thermal regulation, I assume they're not helping to reduce impedance or interference with some signal, etc. But I'm quite new to all this so could be wrong about any of those :D
These are thru vias that connect the backside planes to the frontside planes, often referred to as "Ground Planes". They serve to decrease the resistive and inductive impedances, as well as thermal.

These can serve to isolate signals. In the case of the above PCB there are 4 planes, 2 on top and 2 on bottom. The top planes are connected to the bottom planes with thru vias and normal thru holes for components. The result is two planes above and below that are selectively connected together to form "common ground". If you look at the schematic, note the GND and GND-LED labels, these signify the high current LED ground reference and the normal ground reference planes. Zero ohm resistors R6 and R23 serve to selectively tie these planes together.

High performance precision analog design, including PCB, requires controlling the ground return currents so they don't corrupt the sensitive analog signals of interest. This becomes somewhat of an "Art" as well as "Science" in analog design, although the science fundamentals govern, many don't have access to supercomputers to analysis the effects, so they rely on the "Art" aspect. :roll:

Many Millions of $ and careers have been spent on these "Ground Effects". We spend lots of time analyzing and applying these at the chip level for digital to analog isolation for the first RF/Analog System on Chip back in 2000 which was a massive chip ~18mm by ~24mm, 3/4 digital (FPGA) and 1/4 sensitive RF/Analog (uW Tx/Rx). This research led to a joint patent with IBM on the isolation techniques.

Best,
Thanks for all the info!

Have you turned off the layer showing the plating on the top and bottom layer for the vias, or is the plating super small, or is it some special kind of via that looks like a regular through-hole but without plating shown on the exterior?

I'm glad I haven't had to deal with RF or analog signals where you have to be careful about this kind of thing and place your vias in special patterns around the signal traces. I can imagine it ends up being a bit of an art and as you say, much spent on the research into these effects.

I'll share my (already printed) PCB in another thread some day soon and will look forward to your feedback. It's a fun process once you get the hang of it, though I'm sure I'm flouting some obvious "do's and don't's" of PCB design.


I'm interested in this LED project of yours too but will keep an eye on it from the sidelines for now, got too much on my plate already :D
- Cam

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

kaleun96 wrote:
mawyatt wrote:
kaleun96 wrote:Hey Mike,

I just got into PCB manufacturing myself and have a question about the small holes in the ground plane. I can't work out their purpose. They don't look like vias, I wouldn't think they're required for thermal regulation, I assume they're not helping to reduce impedance or interference with some signal, etc. But I'm quite new to all this so could be wrong about any of those :D
These are thru vias that connect the backside planes to the frontside planes, often referred to as "Ground Planes". They serve to decrease the resistive and inductive impedances, as well as thermal.

These can serve to isolate signals. In the case of the above PCB there are 4 planes, 2 on top and 2 on bottom. The top planes are connected to the bottom planes with thru vias and normal thru holes for components. The result is two planes above and below that are selectively connected together to form "common ground". If you look at the schematic, note the GND and GND-LED labels, these signify the high current LED ground reference and the normal ground reference planes. Zero ohm resistors R6 and R23 serve to selectively tie these planes together.

High performance precision analog design, including PCB, requires controlling the ground return currents so they don't corrupt the sensitive analog signals of interest. This becomes somewhat of an "Art" as well as "Science" in analog design, although the science fundamentals govern, many don't have access to supercomputers to analysis the effects, so they rely on the "Art" aspect. :roll:

Many Millions of $ and careers have been spent on these "Ground Effects". We spend lots of time analyzing and applying these at the chip level for digital to analog isolation for the first RF/Analog System on Chip back in 2000 which was a massive chip ~18mm by ~24mm, 3/4 digital (FPGA) and 1/4 sensitive RF/Analog (uW Tx/Rx). This research led to a joint patent with IBM on the isolation techniques.

Best,
Thanks for all the info!

Have you turned off the layer showing the plating on the top and bottom layer for the vias, or is the plating super small, or is it some special kind of via that looks like a regular through-hole but without plating shown on the exterior?

I'm glad I haven't had to deal with RF or analog signals where you have to be careful about this kind of thing and place your vias in special patterns around the signal traces. I can imagine it ends up being a bit of an art and as you say, much spent on the research into these effects.

I'll share my (already printed) PCB in another thread some day soon and will look forward to your feedback. It's a fun process once you get the hang of it, though I'm sure I'm flouting some obvious "do's and don't's" of PCB design.


I'm interested in this LED project of yours too but will keep an eye on it from the sidelines for now, got too much on my plate already :D
These are just standard thru vias that have a drill and diameter parameter. The holes are drilled then plated creating a thin wall Cu tube between the top and bottom. There are optimum sizes and placement depending on what you want to achieve, but this does get somewhat complex, especially when you work with the EM fields and Maxwell's Equations! :roll:

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

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Here's some simulations of the LED Current using the Strobe Controller trigger at 10ms time.

Edit: I've run the sims with a proper op amp model and updated the text and results and deleted the previous to help avoid any confusion.

Best,


Here's the controller set to 100us Pulse Width with 1A (1000ma) Constant LED Current, the simulations shows ~89us and ~999ma. Note the ringing caused by the NMOS input capacitance creating a delay in the feedback loop. Since I'm not using the exact NMOS device model (which has more capacitance) I'm not too worried about this and the design has a compensation capacitor to handle just this effect.

Image

Here's the controller set to 1.2ms Pulse Width with 1A (1000ma) Constant LED Current, the simulations shows ~11.87ms and ~999ma.

Image

Here's the controller set to 12ms Pulse Width with 1A (1000ma) Constant LED
Current, the simulations shows ~11.92ms and ~999ma.

Image

Here's 3 more 1000ma @ 1.2ms graphs triggered at 1ms with the LED Supply Voltage changed from 12, 16 to 24VDC. Note the lack of influence on the pulse timing and LED current, indicative of true current mode operation!
Image
Image
Image
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

JKT
Posts: 420
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

mawyatt wrote: Typo, CXA1304. Here's the data sheet.

https://www.cree.com/led-components/med ... XA1304.pdf
That clears things up - Thanks!

Unfortunately the result is what I feared ... at least if have gotten the 10% I understand of the datasheet right. :roll:

The 18V version gives about 350 lm and extending the curves to 1 A would give roughly 1250 lm and 24.7 V. With 4 LEDs that would total to 5000 lm, which would give my exposure in about 80 ms ... or with 80 units. I'll pass. :wink:

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

JKT wrote:
mawyatt wrote: Typo, CXA1304. Here's the data sheet.

https://www.cree.com/led-components/med ... XA1304.pdf
That clears things up - Thanks!

Unfortunately the result is what I feared ... at least if have gotten the 10% I understand of the datasheet right. :roll:

The 18V version gives about 350 lm and extending the curves to 1 A would give roughly 1250 lm and 24.7 V. With 4 LEDs that would total to 5000 lm, which would give my exposure in about 80 ms ... or with 80 units. I'll pass. :wink:
I have little experience with high power LEDs, but if you need more illumination then maybe use a higher power LED like the XHP70.2 (or other type) which is 4300lm @ 2.4A, or parallel up devices for more power. The basic controller design should support much higher currents as well, the power NMOS devices are rated at 50A 60W in the TO220 case (use a heatsink).

Here's a quick simulation with 4A and 8A LED current @ 1.2ms pulse-width, only the current sense resistor was changed to 0.5 and 0.25 ohms respectively.

Image
Image

You could also extend the design with more "sections" to support more LEDs, the design is easily adaptable to more sections and higher currents as you can see from the schematic (why it was provided for DIY).

Anyway, I think you have a few options to try and get what you are looking for.

BTW keep us informed if you find any additional info on these high power LEDs, this is a thread we all can learn from :D

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

JKT
Posts: 420
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

mawyatt wrote:I have little experience with high power LEDs, but if you need more illumination then maybe use a higher power LED like the XHP70.2 (or other type) which is 4300lm @ 2.4A, or parallel up devices for more power. The basic controller design should support much higher currents as well, the NMOS devices are rated at 30A 60W in the TO220 case (use a heatsink).
It seemed to me ( https://www.cree.com/led-components/med ... XHP702.pdf ) that the output is around 3000 lm for CRI90, but still quite an improvement. In any case that would make 12000 lm for the set of four and 33 ms or 33 units or 132 LEDs.

If I loosened the time requirement, the numbers would drop. The question is how long could the duration be without burning the controller or LEDs. And the other limit is the max magnification, where I want to freeze vibration.

I'm afraid I'll have to keep the cheap Yongnuos. The value of 400 lms is my estimation of three Yongnuo YN560s at 1/64 power with a diffusor, which directs most of the light relatively well on (large) target. With 200 ISO that works up to something like 2:1 with MP-E. You'd likely be able to do with less if you concentrate the light better. My system will always light up ~d150 ring.

In any case, I would suggest that you start by estimating the required amount of light. Then you'll know what might or might not work.
mawyatt wrote:as you can see from the schematic
Me? From electric schematic? It looks kind of symmetric... :lol: Hooking up LED stripes as modeling light was challenge enough for me...

Saul
Posts: 1781
Joined: Mon Jan 31, 2011 11:59 am
Location: Naperville, IL USA
Contact:

Post by Saul »

Frans (Fotoopa) made very nice system, which produced great results:
https://www.flickr.com/photos/fotoopa_h ... 4329376535

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Agree, knowing how much light is required will help determine the required LEDs, one reason the controller design is somewhat LED independent. One could combine LEDs to achieve various lower levels.

Pulsing the LEDs keeps the average power and temperature down, an LED current of 10A at 10V for 1ms is only 0.1 watt, whereas continuous is 100 watts.

The controller electrical design can support over 100A peak currents for each section with different NMOS output devices, of course this would require a much heftier PCB design and different connectors. Here's a simulation at 100A peak with a 20 milli-ohm sense resistor (2V/100A = 0.02 ohms) and NXP PSMN0R725YLD Output NMOS Device. :shock:
Image

Of course one would want to include various PCB parasitic effects and carefully evaluate the NMOS device model at these currents before committing to a PCB fabrication. However, this shows what the Current Mode Controller electrical design is capable of achieving.

Seems to me the LED is the limiting factor and finding acceptable LEDs at a reasonable cost is the task ahead. I suspect the LEDs will cost more than the controller :roll:

Best,
Last edited by mawyatt on Wed Jan 29, 2020 11:01 am, edited 1 time in total.
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Saul wrote:Frans (Fotoopa) made very nice system, which produced great results:
https://www.flickr.com/photos/fotoopa_h ... 4329376535
Saul,

That is a nice LED pulsed system, thanks for posting.

It's using a Cree CMA2550 LED which looks like a good LED to consider. It's a higher voltage rather than current LED, but the current mode controller can handle this with some different components.

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

Smokedaddy
Posts: 1951
Joined: Sat Oct 07, 2006 10:16 am
Location: Bigfork, Montana
Contact:

Post by Smokedaddy »

Saul wrote:Frans (Fotoopa) made very nice system, which produced great results:
https://www.flickr.com/photos/fotoopa_h ... 4329376535
Frans must have a lot of time on his hands. Incredible link, that's for sure. Thanks for sharing.

-JW:

PS: I wonder if he has a website.

Post Reply Previous topicNext topic