[retired] [WIP] 20mm single sided & 17mm double-sided ?-amp linear driver - surprisingly good!

With the PWM frequency at 1 Khz, what do you see on the output? Is it still pretty smooth? I still find it intersting that the 9.4 kHz runs so well in the M6, but it's got to have something to do with the component network that feeds the 7136 its PWM signal.

I also have not found the parasitic drain to be objectionable.

In STAR Momentary, do the ALT_modes need to match the regular modes? Or do I just comment them out?

The ALT_modes are for the ALTernate, or secondary PWM out. If you're only using a single PWM output, then leave them commented out.

Right, so does this driver use the ALT, or just the primary? Sorry, my question wasn’t clear.

You can see that the PWM output is on Pin 6, so the normal output is what is being used.

  • Here is what I recall from my last testing with the scope (I’ve never placed one of these in an actual flashlight). The waveform is stretched out a lot of course. Output is still pretty smooth, especially at low levels where (as I’ve mentioned) PWM appears to bring arbitrary output levels rather than ones which are really based on the set current + PWM duty cycles. At the higher PWM levels it’s easier to see what’s going on (eg the waveform isn’t actually stretched out, but rather less chopped off than at lower freq. I did not see anything which seemed like it would definitely be a problem.
  • I am confident that it is not “something to do with the component network that feeds the 7136”.
    A. Your normal M6 mod bypasses that entire network: the “P” pad is attached directly to the 7136’s.
    B. The driver in this thread works fine at 9.6Khz with the DTU30N02 pulled from the M6 driver.
    Therefore the issue is entirely with FET selection…

Ah that’s cool to hear I was a bit worried about the 1khz only part of this driver. This fet looks like a good option for my particular usage since when it comes to fet size, the bigger the better. :stuck_out_tongue:
Is there a known source for this mosfet that isn’t from an m6 driver? I can’t seem to find it available at the usual places, ie farnell, mouser or digikey.

No known source. I see the :stuck_out_tongue: smiley, but just to be clear: bigger is not better. Bigger is because it’s an older, less advanced, higher thermal resistance package.

Argh too bad…and yes of course bigger is only sorta better for my particular scenerio where it’s easier for me to mount onto a heatsink and airwire to the rest of the driver.
Those tiny little LFPAK fets are quite remarkable and will work also, as long as I don’t drop them and lose them while installing that is… :wink:

Don’t forget that there’s no reason not to use led4power’s expander PCB w/ this driver: OSH Park ~ Shared Projects by led4power

Of course that big copper pour must be electrically isolated from your heatsink!

So am I the first non-Wight person to fully assemble it? It was a piece of cake to assemble, but I must not have the FW right, because it only lights up on turbo. Here is my source file.

sorry for the blurry cell-phone pic

mine is all in the mail should be here any time soon

You didn’t share that file properly, I get the “You need permission” message.

Just keep dropping the freq until it works.

EDIT: Build looks good BTW.

Thanks for posting your code. I think there is what I would consider a minor regression redundancy of stuff in this version. I’ll take a closer look when I’m more on point.

Please comment out “#define FAST_PWM_START” and then change this line:

“TCCR0A = 0x23; // phase corrected PWM is 0x21 for PB1, fast-PWM is 0x23”

to read like this:

“TCCR0A = 0x21; // phase corrected PWM is 0x21 for PB1, fast-PWM is 0x23”

I fixed my link… I think… and fixed an obvious error in the FW, but now it appears as if moon is missing. The FW has Moon>L>M>T, but it I am only seeing three modes. It is lighting up on Medium and Turbo, but nothing on low.
So you’re saying to take it from :8 to :64?
………….
Oops, missed your last post

I tried the “TCCR0A = 0x21” and I upped moon from 3 to 5 pwm, and there appear to be no changes since my last post. 3 modes, one of them doesn’t light.
Rev 3 code

Update: I also just tried setting “TCCR0B = 0x03” and moon PWM to 8. Again, no changes.

Up the PWM levels a bit. The response curve isn't the same as 7135s or normal FETs.

Ah, sorry. I didn’t think about / notice the low moon mode you were using. That won’t work! :wink:

… Crap, that took forever. I couldn’t find where I posted on the subject! (Post #45.)

I’d start with around 10 to 15 and work from there.

Ok I tried it with this code.
Modes L>M>H>T with PWM levels at 20, 50, 100, 255
Still acts like there are three modes, but only two of them light up. very strange.

Well, just keep going upwards until you get something.

EDIT: I see now that you are saying that you’ve added a mode. Hmm.