I have an LED strobe cuicuit (here is the .brd file Dropbox - zoetrope5buck-proTrinket.brd - Simplify your life).
You can view it here: http://3dbrdviewer.cytec.bg
It is comprised of an arduino pro trinket, a capacitor and 10 buckpucks. The strobing works fine, but I find that with each additional buck puck that I add, the strobing gets a little dimmer. I have a high quality power supply. One fairly obvious thing that might help is to solder the buck pucks in, as they are not currently soldered. They are however, firmly in place, as the holes that accept them are very small.
Is there a way to modify the circuit to get more juice, more quickly out of these buckpucks? For example, beef up the signal coming from the arduino with a darlington array?
My main concern, aside from the power of the strobe, is precision, as I am using this for a 3D zoetrope and thus need the strobe to be exactly 16Hz.
Looking forward to your thoughts and happy to post any additional info needed.
That sound like a power supply problem. Do you have a multimeter to make sure your power supply voltage is "holding up"?
Your power supply has to supply the LED power, plus some allowance for your other circuitry and some inefficiency. For example, 10 3W LEDs would require a 30W+ power supply.
Your power supply may not have a power rating, but power is calculated as Volt x Amps (i.e. A 12V / 1A power supply is capable of 12W.)
revel:
I measured the current draw and it is only .11 A.
That may be the average current draw, but the pulsed current may well be dropping the supply voltage. What duty cycle are you using with the LED's? Have you measured the supply voltage? You may be able to add some more filter capacitance to the output of the supply if it is dropping under the instantaneous load of the pulse.
Hi, Your PCB design is far from adequate for the 11A you are trying to run through those copper tracks.
0.11A is not what the LEDs are pulling.
Measure the Vin voltage that supplies power to your LED Drivers, measure this voltage ON THE PCB at the power input socket and then at each of the Vin pairs on each driver.
You need to redesign your PCB, bigger tracks and a more substantial socket than the one you are using for LED Driver supply.
I am surprised this track has survived.
This is like trying to run a car starter motor through a bit of speaker wire.
Ok. so ideally I need to look at the trace length(s) and size them according the the current they are carrying.
When the LED's are running at full blast they are pulling around 5A. Software dims them for the spin up time for the Zoetrope, then pulses them for 62 milliseconds on, 500 microseconds off for the 16Hz strobe at which point in time I read the aforementioned .1A in line at the 12V supply. But I have yet to measure at the PCB or individual drivers.
Since I do not have time to remake the PCB, and if indeed the trace size is a problem, I wonder if I can use two PCB's, with 5 buckpucks a piece, and control both of them with the arduino...
That could help. The supply is capable of delivering the current. You might also make sure the 12v feed wires are securely fastened to the board by means of screw terminals or simply solder them in. A capacitor on the supply line could help hold up the voltage, but it would have to be selected carefully to not be too big as to not be able to charge (and act as a load) or be too small and not do anything useful. Of course they won't do anything helpful if all the leds are running continuous at full power.
I guess it should also be said to not use 30ga feed wire to push 5A through. You can do some voltage drop readings with what you have now and see where the problems are.
Start right at the 12v supply and see what the voltage is with them on continuous. Leave the positive probe hooked up to the + at the supply, and start going down the positive bus with the negative probe. Check where the feed wire interfaces the board, then halfway through the circuit, and finally the end. Ideally, you would see 0V all the way to the end, but that isn't going to happen.
Same goes for the ground side. An undersized connection on the ground pin can have the same effects as sagging voltage. You can check it as well by running the same probe through the ground circuit as well. Ideally, you would see the supply voltage the whole way (12v, with the positive probe still to + supply)
Hi,
Because you are using PWM, the current is still at maximum in pulses, the maximum current is however reduced by the resistance of the PCB tracks.
revel:
Ok. so ideally I need to look at the trace length(s) and size them according the the current they are carrying.
When the LED's are running at full blast they are pulling around 5A. Software dims them for the spin up time for the Zoetrope, then pulses them for 62 milliseconds on, 500 microseconds off for the 16Hz strobe at which point in time I read the aforementioned .1A in line at the 12V supply. But I have yet to measure at the PCB or individual drivers.
Since I do not have time to remake the PCB, and if indeed the trace size is a problem, I wonder if I can use two PCB's, with 5 buckpucks a piece, and control both of them with the arduino...
That would make a difference to the current on each PCB, I'm not sure if you have room to put some wires between the power inputs of each of the drivers to bypass the tracks.
Tom...
I and going to be looking at this today and will implement some of you suggestions. Another thought I had was that I could double up the PCB's, since I have an extra, running 5 buckpucks on each and thereby halving the current load on the traces.
@gps mikey @tinman13kup lets talk capacitors for a second- what size capacitor do you recommend and / or how do I calculate that? right now I have a 25v 100uF capacitor on the pcb. I am planning on replacing it with a 100v 47uF capacitor, does that sound like a good place to start?
@TomGeorge re: voltage drop, I am measuring 12v right at the pcb and right around 7.5v at each buck puck.
If you are losing 4.5V between your feed wire and the led, that is where you need to concentrate on. A capacitor will not help that problem. If the supply isn't dropping, the capacitor right there has no reason to discharge. You would need capacitors at each led, which would have 500uS to get 4 time constants in to be of any help when you are pulsing. When you have the leds on, those capacitors would rapidly discharge and do nothing.
Is there any way to make a few jump wires to go from the 12v supply to a pad in the middle of the pucks? Is the supply feed wires mechanically connected or soldered, or is it just a push in pin?
