8 channel 40 Amp buck for NUBs

[rhd]

With separable tactile triggered MOSFET switch / power distribution daughter board.

 

[RedCowboy]

Nice, this will be a bucking type right?
I will be needing one.
I suppose I could change 8 resistors and set for 36a i.e. 4.5 each.

 

[Cyparagon]

Why are you driving them separately?

 

[RedCowboy]

Because they will be driven past factory spec and just 5% - 10 % short of fold back I would rather not count on all the diodes remaining balanced as they wear.

If powered in series at 95% of fold back and one went open the rest would be overdriven past fold back.

If powered in series any imbalance that may appear as they wear and could cause the others to be overdriven in the worst case or cause the one to be overdriven.

It seems that they would be better protected if regulated independently, also if heat became an issue in the summer months I could choose to run every other diode, it gives me more options having each diode regulated independently.

Also if a regulator it's self failed I would lose 1 diode and not all 8.

What do you think?

 

[Alaskan]

I could sure use three of those eight channel drivers, very nice.

 

[Cyparagon]

if heat became an issue in the summer months, the thermal design needs to be improved.

Fixed that for you.

If powered in series any imbalance that may appear as they wear and could cause the others to be overdriven in the worst case or cause the one to be overdriven.

You're thinking of parallel operation. Series operation guarantees EXACT matching of current, even more precisely than having precision current sense resistors in each of these 8 drivers.

https://en.wikipedia.org/wiki/Kirchhoff's_circuit_laws#Kirchhoff.27s_current_law_.28KCL.29

Disabling one of the diodes in the string is literally as simple as shorting it out (while the driver is off of course).

Call me lazy, but I prefer using one driver to 8 drivers if I can get away with it.

 

[Pman]

I could use a 40A single driver for a Phlatlight or at least 30A.

 

[Alaskan]

Maybe a high current PWM controller like this for your Phlatlight LED?

55A Constant Current HHO PWM CCPWM Digital Amp Meter 12V 24V Waterproof | eBay

 

[rhd]

You're thinking of parallel operation. Series operation guarantees EXACT matching of current, even more precisely than having precision current sense resistors in each of these 8 drivers.

https://en.wikipedia.org/wiki/Kirchhoff's_circuit_laws#Kirchhoff.27s_current_law_.28KCL.29

Disabling one of the diodes in the string is literally as simple as shorting it out (while the driver is off of course).

Call me lazy, but I prefer using one driver to 8 drivers if I can get away with it.

What he said ^

Slightly longer explanation of rationale:

1) Parallel is too dangerous for an array, it would have to be series. I want to power from a high current 7.4V Lipo, so that would require a boost. I don't have any boost designs sitting around in Eagle that can boost to ~44V @ 5A, and I'm also lazy.

2) More important, I want to eventually be able to swap out the single MOSFET tactile power distribution daughter board for one that has independent channel control. This retains the flexibility to try that out without rebuilding the driver portion.

That said, it took me a year to get to reflowing this board after design and PCB ordering, so it may be a long time until I get to #2.

 

[paul1598419]

Looks like everyone beat me to the series current issue. But, yes, a series configuration of laser diodes would insure the current would remain constant for each diode.

 

[Alaskan]

You could buy one of these for a high current series CC regulator, but it is fairly large for what you want to do with that double quad driver and probably much more expensive:

3 24V 60A High Power Infra Red Laser Diode Driver 5kHz | eBay

 

[RedCowboy]

Fixed that for you.



You're thinking of parallel operation. Series operation guarantees EXACT matching of current, even more precisely than having precision current sense resistors in each of these 8 drivers.

https://en.wikipedia.org/wiki/Kirchhoff's_circuit_laws#Kirchhoff.27s_current_law_.28KCL.29

Disabling one of the diodes in the string is literally as simple as shorting it out (while the driver is off of course).

Call me lazy, but I prefer using one driver to 8 drivers if I can get away with it.

I had thought the loads each had to be very closely matched or you would get different brightness, I know todays leds and lasers are very well matched, I was thinking about uneven wear.

You are right about the heat dissipation design, that's was a lame excuse.

I know with a current regulated driver the current each diode gets will be the same, but if resistance varies any then the voltage drop could be different, but as good as these are I doubt it will be an issue, I tend to error on the side of caution and like to over build.

I will have lots of heat sink and fins, maybe forced air but at least some fins and one 36a driver sounds a lot easier. :beer:

 

[Benm]

Looks like everyone beat me to the series current issue. But, yes, a series configuration of laser diodes would insure the current would remain constant for each diode.

It would, but there would be considerable voltage between diodes and a heatsink if they were all mounted in the same one.

Considerable is relative though as the forward voltage is perhaps 5 volts each for a total of 40 or so. I'm not totally certain of that the maximum permissible pin-to-case voltage is on there diodes, though i doubt they'd actually break down at 40-ish volts.

In general it is something to consider however, if you try and drive say 100 luxeon-style star leds in series with the pcb at ground level you'd probably get problems, and very nasty shocks if you were to poke around with your fingers while in use.

 

[paul1598419]

It would, but there would be considerable voltage between diodes and a heatsink if they were all mounted in the same one.

Considerable is relative though as the forward voltage is perhaps 5 volts each for a total of 40 or so. I'm not totally certain of that the maximum permissible pin-to-case voltage is on there diodes, though i doubt they'd actually break down at 40-ish volts.

In general it is something to consider however, if you try and drive say 100 luxeon-style star leds in series with the pcb at ground level you'd probably get problems, and very nasty shocks if you were to poke around with your fingers while in use.

True, if the case were part of your circuit. If not, no issue. I do get a chuckle out of thinking of someone driving 100 or more LEDs in series with a single driver and then poking around with their fingers in the live circuit.

 

[Cyparagon]

if you try and drive say 100 luxeon-style star leds in series with the pcb at ground level you'd probably get problems...

I've tested a chinese LED star (cathode to heat sink insulation) to break down at about 1100V. 100 in series might actually be doable. I know from experience that 50-some in series is fine.

I'm not totally certain of that the maximum permissible pin-to-case voltage is on there diodes

The M140 projectors ran 4 parallel groups of 6 diodes in series. 7 diodes in series is perfectly fine.

 

[rhd]

The M140 projectors ran 4 parallel groups of 6 diodes in series. 7 diodes in series is perfectly fine.

Further evidence of that is the fact that the datasheet itself puts all 8 in series and refers to the array as an "LD Bank" with an operating voltage of 30 to 39V.