Shadow SL3 + RGB SST-90's + custom 4ch driver (new CC driver in post 101)!

The green and blue will be allright, the red @ 2v is as issue tho I’m running a RGB light with XP-E2’s now and it works allright, the red is by far the brightest. Thats what the addition of the limit resistors are for on the 17, I’m gonna start from scratch and do another 22 tomorrow.

I’m figuring I’ll use one or two 1206 pad per channel and I can stack them on top of each other if more are needed.

So R G B leds will be near white light? Anyways, WOW this needs to have video!!! 3 FETs on a board lol.

I’m sure it’ll be a heck of a light. Good luck with the build!

He has a working RGB right now…Cereal you should post it up to see what they are gonna get…just BIGGAH! :smiley:

These are XP-E2’s at ~1.2A each (that’ll soon be raised to 2.4A) the strobe it 19hz however each color only pulses at 1/3 that rate cause it does RGBRGBRGB (the iphone camera has no chance, video just doesn’t do it justice)

Thanks for the video!

Poor dog, getting blasted with RGB light.

Did you mean 19Hz instead of 19kHz, because that would be something like PWM, not strobe.

Wooo doggies! ;D
If only i can see it in person! A guy on the “other forum” posted about strobe.
Study of Tactical Strobe
Most important part:

Having said that, the right strobe needs some clear specs.
We found, that with a slow or changing strobe, the opponent was able to (slowly) move their hands without being detected. The strobe is still very disorienting from the slower 5hz to 12hz and even higher, but the effect seems to get less when going much faster than 24hz.

We worked the strobe up to around 18hz to 20hz and that problem (noticing movements)was tackled. At 18hz to 20hz, the strobe was still extremely disorientating and the opponent could not move a finger without detection.
So, strobes at 8hz, 10hz, 12hz etc ARE very disorienting, but could have unwanted negative side effects for the LEO.
Thats why we chose for the 18hz to 20hz, less unwanted side effects.

Do you think you can do a “no delay strobe” as in, there is no time delay between the color switching? maybe that could be pretty cool.

Poor doggies, they were never designed to suffer anything like that

You are lightning fast CerealK! I can't have a night sleep and the first version of a new rgb fet-board is done!

What is the driver now in that rgb-xpe2 light in the video?

About the values of the current limiting resistors on the little rgb(w) board: perhaps it is best to tune it while measuring output current. It will be dependant on the used battery and other resistances in the used host. I'd say, better be a bit on the low side with the resistors , then it still gives good output with lesser hosts/batteries. It looks like 2A is perfect for all xpe2's (and xml-colour dies) if on copper/Sinkpad, and it seems that up to 3A is still handled fine.

Yes 19*Hz*, sorry bout the multiple typo’s, my phone insists on changing Hz to kHz every time I type it, it must assume nothing runs at JUST Hz anymore… I wish I have a video camera that could show it, its intense to say the least.

As for the driver, let’s just say its proprietary lol…

That thread you linked is the exact reason 19hz was choose. BTW a new driver is in development for the current light (with the assistance of my buddy tterev3) that will do even more (that’s why the drivers in this thread use the 8k tiny85 so I can program it AND have lots of room to play with) but the new one for the tk45 will be based on a PIC MCU to fit even more programming for fancy stuff.
I love you guys but please don’t expect much info on it, I’ll show it off sure but its one light I intent to keep the spec’s / production info on hush hush, atleast for awhile.

What do you guys think about using a Zener diode to limit the voltage to the red emitter, could a ~2.2v Zener be found that could handle ~2.44-2.6A being passed threw it?

Probably not practical… it would have to dissipate a LOT of power. If you are dropping 2 volts at 2.5A that would be 5 watts.

Electronics noob here to trash your thread!

Why would a Zener be better than a cheapo black barrel silicon diode, or a silicon diode in one of those heatsinkable packages that you see FETs in? Do you just need something that will drop a few volts, may be inefficient, but may be easiest.

Can we discuss this more please guys, what other options are there to physically limit the red channel (so the PWM settings in the PCM can be equal)?

Remember I need it to fit on a 17mm board!

You know what they say, always go full retard.....I think.

:smiley:

hehe

Don’t know enough about FET’s to be be technically accurate enough for the discussion

I wonder if you could calibrate different gauge wires, as in on the red one, use a 30 ga wire instead of a 22 ga and see if it restricts current flow enough not to overdrive the red led

In a fet driver with a led that can not have the full current the question is just: where do you want to burn the power, and perhaps a resistor (near the edge of the board) is a better place than a fet or a wire?

I’ve asked comfy to weigh in on it, I’m thinking about the possibility of purposely picking out a bad fet, that is one with higher RDS (on), to lower the current by adding resistance in there instead of needing another resistor. Same idea as with the different wire size, it shouldnt take a dedicated resistor to add resistance into the circuit if its determined thats all it would take to adjust for the lower vf.

I have the one light running now with everything equal and while red is brighter cause its being overdriven it is running otherwise on evenly powered channels, this one will be alot more power tho

If it needs one where should the limit resistor go? In line with the FET’s source (GND) or in the red LED’s + path (or both)?

I’m now using a 14 pin QFN PIC chip and will be running 4 channels (R G B and W) with the option to add one more for UV so space is very very precious on this 17mm single board driver, I still intent to have a batt+ contact on the bottom for direct installs.

Don’t put it on the FET ground… the FET may not turn on.