I have just had a rather interesting idea. Would it be possible to drive three LEDs in parallel off a single AMC7135 type driver (such as a AK 47 or Nanjg 105A) and retain the modes?
I fairly sure with a little tinkering you could piggy back a few AMC7135s on top of each other and then split each AMC7135 pile up so they will run individual LEDs in parallel at a regulated current (e.g 3x3 AMC7135 piles powering 3 LEDs at 1A each) however in doing so you would lose the modes (effectively all you have done is make 3 different boards out of one).
However aren't AMC7135s already run in parallel? So if I did what I have outlined above... could I solder an LED positive to each AMC7135 (like one on the end normally is) and draw off the current before it is passed on to the next AMC7135 (or am I misunderstanding how these things work?)
Don't run LEDs in parallel, small variances may lead to a lower forward voltage (Vf) of one of them, that one 'steals' the current supposed for the other ones, gets hotter and an even lower Vf and steals even more...
However if you cut some traces on the board you can separate the AMCs into groups so that each LED gets it's own group of AMCs, that would work.
I have been thinking of doing what you said: cutting the traces to in effect create 3 different drivers (on for each LED) however I do not understand your comment in regards to forward voltage decreasing with heat? I have just been playing round with my Trustfire X9 and as far as I can tell the Vf increases with temperature.
Resistance increases with heat so wouldn't that lead to a decrease in current draw ? (Cree's datasheets also support this I believe)
Hi Manual Man. You can retain the modes if you maintain the traces leading to the VDD pins of the 7135 ICs. Just group and cut the traces that connect to the AMC7135 output leads. Each group will power one LED, etc.
MCE, P7 and all the other multi-die LED packages work for two reasons: The dies come from the same production batch, so nearly no variances, and they are very near to each other on the same carrier and thus strongly heat-coupled, so none of them can do the thermal rundown without taking the others with it.
For separate LEDs thermal coupling is quite weak; it still can work if they are very similar.
Let's say it this way: It can work, but it's not recommended.
To run multiple emitters from multiple AMC7135 regulator boards (one board per emitter) and have all them mode-controlled by one of the boards, see this thread.
If you don't have room in your light to stack multiple boards, then you can achieve the same thing on a single board by cutting the trace routes that combine the LED- to the left leg (legs pointing down) of each AMC7135 chip. Then soldering each LED- connection individually, one LED to one AMC7135 chip (or vertical stack of chips, or equivalent, as in splitting a 6 chip board into 2-2-2 and soldering each of 3 LEDs to a separate pair of chips, etc.).
Bump... Don't know how I missed your reply Techjunkie... but THANK YOU!!!
I now have a 2/3 completed multi-LED driver sitting beside me (1 more AMC7135 to go). This is exactly what I was looking for. I will finish the board tomorrow, if I have to fiddle round with another AMC7135 (they really are small aren't they?) for another hour...
The doctor said it best. I too have run them in parallel but it’s not ideal. Cutting the output trace, desoldering that pin, or reaming out the vias that connect the output of individual chips, groups of chips, or sides if the board and linking each group to its own led will be more reliable. If an led or solder joint fails you only lose that led rather than the whole bank.
7135’s operate like people in a restroom. The battery is the keg they’re all drinking from. The traces act like the plumbing. When they flush(voltage applied to Vdd), it all goes to the same place(leds connected together). Cutting the traces amounts to having the plumbing hooked up to separate septic tanks(leds operate independently). Unlike 7135’s, bladders come in a large variety of outputs. Keeping the Vdd traces intact amounts to having one hand on all the levers.
I hope this is clear enough without being overly graphic
Yes, maybe a little buffer cap is needed, that’s something to try out. It wouldn’t have a reverse voltage protection diode nor battery voltage detection the though.
I have succeeded in driving 3 LEDS individually off one board. I removed all but one 7135 off the bottom of the board, soldered one more 7135 on to the top of the board and cut the negative traces so I have 3 pairs of 7135s. Each LED is soldered to an individual pair of 7135s so they each get 700mA and the modes are retained.
A big thank you to the guys on here who told me what to do and how to do it.
Unfortunately I had to de-solder everything and re-solder it to get it to fit into a p60 drop-in I had machined out for it, and in the process it no longer works. I may have to sacrifice a third 7135 board to get it to go again
I will stick up some pictures once I get it going again (but don’t hold your breath as I am quite bizzy at the mo). The circuit draws 4mA off a 18650 however nothing goes, so perhaps I have bumped into the same problem I did last time (PK pilot suggested the 7135s might be switched off or something). That or it is just bad soldering on my part
Nvm guys… just got the soldering iron out and it turns out the LEDs are fried, I think I must have let the board get to hot while soldering wires on. I ended up reflowing one of the XP-Es and have now de-domed it. I tried re-reflowing it but it defiantly does not work now. Going to cost me another $15 or so to get another triple XP-E red sent over from cutter.
The build cost so far has been:
p60 drop-in machining: $30
brass unit for housing driver: $1.50
triple XP-E from cutter $12 + $3 S/H
Driver boards (x2): $7
hours put in: at least 10
Looking at another possible $15-$20 to get this thing working inside a p60 drop-in.
