17mm RGBW CC drivers for both E and clicky switches. Video of running light in post 41!

Flashlight Modding and DIY Parts
1

First RGBW_clicky light is up and running! Video here

I'm making this thread to document the progress of the 17 and 20mm drivers that are for tterev3's MELD UI (his FW is proprietary but he's happy to sell programmed MCU's) .If you havent yet checked out what MELD is capable of I highly suggest you check out his blog for a demo of its amazing functionality.

*edit* there is now another 17mm driver with a different MCU and a FW for power cycle mode changes (no e-switch needed, just the tail clicky!!) in post 9, I will be building 5 identical RGBW Supfire A6's using these drivers with quad emitters and 5 X6's with XML colors, if you think you may be interested in purchasing one please let me know ASAP (2-3 A6's spoken for, one X6 spoken for).

These started out in my thread about my SL-3 RGB SST-90 built (which runs DrJones RGB FW and doesnt have a white channel, instead mixing RG and B for white). I've really gotten on a RGB kick recently, its as fun as lasers without the danger. If you dont have a RGB light you NEED one, they're really fun to play with and do actually give off white light even without a white emitter so its actually useful as a flashlight, not just a toy.

These are a pair of 17 and 20mm CC drivers, since we cant tune the PWM level in the FW I needed to find a way to control current per channel, the current can be set via the RSET resistors. When runnig a 2502 FET the max current would be 4.2A, with the AO3400 it would increase to 5.7A. *NOTE* max current on the 17mm version is limited by the power the 1206's can dissipate and will be around 2.5A per channel, the larger 20mm driver is being upgraded to the double size 2512's and will be able to do the full power of the FET's

Some people have tried to say the red emitter wont work cause of its lower vf but I have a RGB light running now at >6A (as well as a few other red only lights) and thats simply not the case, it runs fine and is even output with the other channels, it may technically overdrive the red channel but it doesnt negatively effect operation. This would also be a great driver for an XML-Color. Note the FW REQUIRES a e-switch, its not able to be modified to tail click only, there is just to much function for it to be controlled by power cycle. (see reply 9 for a power cycle controlled option)

These are going to be very voltage intensive circuits, as such I decided to leave polarity protection off cause of its v drop (also since the PIC MCU handles voltage monitoring internally even a slight drop over a protection diode would cause it to read a fair bit lower), I briefly considered a PFET for polarity protection but that would give the same issues. Again there is NO POLARITY PROTECTION on either of these drivers

[20mm version not happening]

and shrunk down (was NOT easy) to 17mm:

Here the SW+ has a trace to a masked over via, this is for if you are installing this piggyback and need to make connections from the bottom (if the SW+ connects from the top use the pad, if it connects from the bottom just scrape the mask from that via and connect there)

Both drivers can be piggybacked onto your host contact board or can be installed directly in the 17 or 20mm pills but again an e-switch is required, there is no way around this.

edited for spellin'

2

The 20mm version sure is a lot prettier…

Also note getting rid of the protection diode does away with the whole CAP location debate!

3

Wow…looking very very good man! Getting good at Eagle…keep up the awesome work man

4

I can’t wait for your results when you build these up.

5

17mm _v2.0.3

v2.0.4

v3
I was able to find a new way to ground the BJT which I’m happy about, I really didnt like how I had that done before. Other changes include the signal traces, they’re ran on the top, only going to the bottom for the gate resistors then coming right back up. Also moved the white components back even with blue and tried to take care of some errors.

v4
LED- via’s, bottom ground copper changed, clearance changes, components & GND via’s moved to be more symmetrical, also did some work getting rid of unrouted issues it was giving me and managed to get rid of all the yellow X’s. I think this one should be ready to order!

6

Wonder if the Quad XML board (illumn.com 20mm-copper-direct-thermal-quad-xp) with the Quad optic (http://www.ledsupply.com/quad-carclo-led-optics) in a SupFire A6 (or equal size light) would work for this 17mm RGBW driver.

Ooops - no e-switch, so no dice... Dang!

Do you have an EDC host in mind? The Quad board and optic for XM-L2's would be perfect. Maybe it can be a "choose-your-tint" variant rather than multi-color?

7

The A6 is my dream, probably take $100 worth of machine work to a $15 light but if I can get it modded to an e-switch I wouldnt have to buy another EDC light for myself ever. I plan to run RGB XP-E2’s and a white XP-L (an XML color would be a cool light too)

8

Amazing CK.

I have followed your work as best as I can since you wrote that you'd get started with eagle and fw programming and I am just amazed at your skill and results. I never commented anything before because it might interfere the thread in some cases or because I hardly got the basics of what you were doing. But this is soo cool. Well done man. Well done!

9

^ +1

10

Thanks guys.

Today I adapted the MELD UI driver to take a 12F617 and deleted the e-switch inputs, this here is the driver for power cycling mode control. I hope to have FW for it ready by the time the parts get here, these are going in Supfire A6’s, I’ll be building 5 of these to sell (well keeping one), I expect to sell them for around $75, if your interested speak up now as I imagine these will be spoken for quick once word gets out.

I’m not asking for any money now, I’m building them out of pocket but if you think you’ll want one better let me know ASAP.

these drivers are ideal for 4 XP-E2’s or an XML-Color, they can deliver about 2.0A per channel from a single cell (so ~8A tail draw on full power), the output is adjustable via sense resistors and you CAN go higher but I wont ship them doing any more than that, its a good power level for either emitter option and more power than that and you risk putting to much power threw the 1206’s. The light can be operated in white only without having to ever see the colors if you just need a normal flashlight.

Note that a 0805 cap AND a 0805 resistor stack on the C1 pad for off-time memory checking (I may add a dedicated pad for each before ordering, I’ve got time, the hosts are coming from China)

11

Are these PWMless drivers, using an analog signal to control the mosfet, a la led4power’s driver?

12

No the PIC is a different type of MCU…still PWM

13

The MELD UI driver is here, built and running on the bench from an external PWM input. I expect the QFN chip to get here Monday from tterev3!

First host is also coming but from FT so probably a few weeks still. It’s a SolarStorm T4, it’ll run RGB XP-e2 colors with an XP-G2, XM-L2 or XP-L. No UV on it.

Next MELD host will have an even more advanced driver (including a smart-latching e-switch to cut power to the rest of the driver for a 2-button interface), don’t want to give it away but let’s just say this- the host comes as a WRB light (R&B using 5mm emitters), I’ll be turning it into RGBWUV with all high power emitters.

14

I am in awe! Please oh please do a video review of one of these once you have them running. They are sooo cool!

15

As with all driver threads, Ive been subscribed since day one. Thanks for all of your efforts! :bigsmile:

16

That looks excellent! Would this driver in the 17mm flavor fit into a P-60 with a quad TIR?

17

Dan, those boards look brilliant! Working with tterev3 and using his outstanding firmware is a awesome idea. Is it at all possible to use his firmware and your driver with one of the XM-L RGBW LEDs? If so, I want one when you are ready to start offering them so consider this my heads up. :D

I NEED one of these in an EE X6!!

18

Thanks guys!

Yes these work with the XML-color! Only thing is I highly suggest limiting max current to around 1.75A, with an XML-color, reason being is even tho each die can handle >2A running (Djozz’s testing) running at >8A and having the heat try to get out threw a single thermal pad is gonna cause you heat issues.

19

Will the quad optic fit a P60? This driver will and you can use any of the RGBW emitters or a quad (assuming it fits)

The built driver above is the one that requires an e-switch so this one wont work in either a P60 or a EE X6 but the power cycle one will work in either! Everett tells me the power cycle FW I’m buying from him is almost ready too, if yo uhavent see it yet this is how it’ll operate (note main mode sequence swapped, in the video its H-L, its been changed L-H)

One other thing I’m going to see if its possible to implement is what everett calls “dim to red”, in his MELD FW its a configuration option the end-user can do from the config menu (no coding, done by clicks). Now I’m not sure if its even feasible with a power cycle controlled light, I’ll have to ask him about it, but when enabled the moon mode in the main sequence is red, when disabled its white. He does it like this on his MELD and has configurable options in several of his e-switch FW’s but I have no idea if its possible with power cycling only so absolutely no promises at all but thats one change that if possible may be coming, otherwise the FW should be very close to what you see there.

Note anyone that wants one now you know Matt has a RGBW power cycle driver that runs a tterev3 FW he commissioned. a little more simple than this one but he has everything in hand and could make and ship you a P60 drop in right now, this is still at the minimum a month out.

20

Great work, nice collaboration. It’s really something to see what you all can do.