R1-4 = set resistors
R5-8 = 100K
Q1-4 = N FET
Q5-8 = BJT
the MCU on the clicky is an SOIC8 12f6171, on the E-switch is the QFN 16F1825 provided by tterev3 with “MELD” v2.13 FW (damn that’s a tiny chip with 16 contacts).
Im waiting on more MCU’s (along with some customization in the code) for the T4 but I have the clicky up and running now (had it up and running last night but it needed some major code revisions, its good now!).
Yep, not right at first because that’s one of the 2 lights I’m building 5 of to launch my new driver, 5 of those A6’s (with quad’s with RGB XP-E2’s and an XP-L) and 5 EAgleEyes X6’s (with XM-L Colors).
If anyone’s interested in getting on the list for one of these first 10 lights let me know ASAP, (3 A6’s spoken for, all 5 X6’s up for grab). The price is $80/ea. DJozz and anyone else, after the initial launch I will be more than happy to sell you drivers, I’ll also be releasing all boards open source so you can DIY (I’ll sell you just the programmed chip in that case).
Note that even tho you won’t have access to the FW to make changes there’s still lots of options you’ll have:
You can change max current via swapping resistors, you can change the memory reset time by altering the ratio of cap value to resistor value. Also at the time of ordering you can let me know if you be using an XML color or quad emitters. If your using an XML color turbo will be disabled (turbo is RGBW all on 100, high is 100 white either way, turbo adds an extra mode). Also note that I’ve got an update planned (tho not any time soon) where turbo wi be configurable via an in UI configuration so it can be altered by the end user.
Video demo of UI in running light coming tonight!
Well, as far as I know that CC circuit, it should work with lower values easily, and lower values should make it react faster. However I never PWMed it and don't know what side effects that has...
Ok I’ll give it a shot (during all the research I did I found everyone stating 100k worked but no explanation of why that was chosen). I wish it would be possible to make it run with jumpers so I could exclude the gate resistors all together to free up a lot of space for larger set resistors (or just another parallel set of pads).
What I really need is a scope.
Here’s a very short (<2min) video quickly going threw the UI
Nice and well thought out! I like it.
There's only one thing that I miss ( an ultimate rgb-snob feature I must admit): a colour mix mode group in which you can ramp the different colours separately, so you can make the whole colour space :love: . That could replace the third mode group.
Unfortunately that would be nearly impossible with a power cycle controlled driver. It’s way harder to do cause unlike an e-switch the MCU cant time switch presses since it dies instantly when power is cut.
There is a major update planned for this UI but not till after the first 10 lights are finished and sold, I need some funds coming in from this project, I have so much [money and time] tied up in it I have to start selling it or I wont be able to afford to continue to develop it further.
thanks for the good word tho, you probably know this but your one of my major inspirations for doing this!
Now that the FW issues are worked out and it was running I went back and tried messing with the gate resistors again, first I tried 2k, then 1k then 470 and now I’ve got them jumpered, the issue was apparently 100% the PWM level being to low in the FW, it’s running with no gate resistors now with a min value of 5.
edited: decided to go back to 220ohm for safety.
High value gate resistors do some pretty nasty things to the FET's gate signal. The drivers I've seen that come with gate resistors stock are using very low value parts, like 10-15 ohm. The higher values we used to use were a half-assed patch to hide the Nanjg circuit's voltage boosting issue, you won't have any of that with the different circuit you have here so there's nothing for the gate resistors to fix.
The PWM output/gate circuit is just a dead end, there's no real current flow once the FET is finished turning on, so there's nothing really there that could cause too much current draw on the controller's output pins. I of course don't know squat about the PIC stuff but if it's like the Atmel, trying to pull too much current from the outputs just trips some sort of internal protection and the MCU shuts down until things return to normal, even if it's accidentally shorted straight to ground.
Here’s my first quad wired up with emitters attached and tested (and also the optic modified to accept the XP-L’s dome), 2 traces must be cut to make the led’s individually addressable.
XP-l
How do you address the LEDs individually after making those cuts to the traces?
Ahah! So they share a common negative connection, the positive lead is the thing. Cool.