Mike C drivers: v8 series, ATtiny1634 based.

Cool. Thanks.

Well, one thing led to another… now there is a F-4. It’s the same as the F-2 except with two additional 7135s that are on the back side, which in turn required a slightly smaller spring pad so they could fit. Of coarse the two new ones are on another pin, so any number of one to four 7135s can be turned on in constant current, and for any level between I will PWM the 7135 on PB2.

I’ll still keep the F2 available as it has a larger spring pad, otherwise there is no need for it if Nanjg sized springs are used. If only two 7135s are wanted, don’t populate the ones on the back (spring pad) side. I also put another 0805 cap on the back side since I removed one from the F-2. If the F-4 is made with four 7135s, then it’s a double sided board anyway, so two OTSM caps can be used for longer measurable off times if desired.

I also managed to squeeze in another switch via, so if a E-switch is soldered in place there is still an available via for acupuncture style flashing. This is for both the F-2 and F-4.

I want to test that the F-2 v5.2 works before making the v5.3 F-2 and F-4 available. As far as I can tell, no one has ordered the current F-2 so I’ve un-shared it for now because the new v5.3 (that the firmware will be working for) will have a pin change.

Work is getting the way though, I’m being sent of to Israel tomorrow, and then on to New Delhi, so this stuff will be on hold. I don’t want to take small electronics and development stuff will cells, buttons and wires with me because Israeli airport security is a lot tighter than most other countries.

You are a true jet setter Mike. :stuck_out_tongue:

It sounds really nice. I should probably get or make one after it’s known to work, for development purposes. It seems avr-gcc didn’t add support for the tiny841 until fairly recently though, which means I’ll have to upgrade some tools…

BTW, before you finalize anything, do you think there’s any chance of getting a trace from an ADC-capable pin to LED- ? This would allow the driver to sense the ambient light level while the LED is off, which means it could be configured by pointing it at a computer screen with binary code blinking in black and white. Or it could maybe sense whether to skip moon mode when the ambient light level is too high for moon.

I will do that on a next version. For testing in the meantime there is a via used for flashing that is otherwise unused. It’s on PA5 which has ADC capabilities.

TK, that really sounds wonderful. I love a true moonlight mode in almost every light and if it were there or not based on existing light… that would be awesome! The programming idea is interesting, thus having the extra space in the chip might be huge!

We seem to have new drivers showing up every day that have lots of potential right now. It is awesome! Just wish I could wrap my head around building them. I can see I will be ordering parts soon and testing this bad boy out!

Any news on this?

Updated OP with info on new v6 design.

Nice work Mike. Just wish I understood the workings a lot more than I do. :beer:

Holy super driver, batman! Great stuff, Mike!

Thanks!

I’ve just received the new v6 design for the F-2 and F-4, so I’ll build a few and make sure there aren’t any screw-ups.

The v6 component and pin design is already tested and working nicely though. I made a L4 dedicated version with 14 x 7135s which is the first of my v6 drivers to actually make it into a light: What did you mod today? - #2984 by Mike_C

I used a 22uF X7S rated OTSM cap on the L-14 which consistently gets over 10 measurable seconds at 4V and room temperature. I’ll have to lower the voltage and heat it up to test the margins.

Nice
Gotta love those many pins on the MCU :slight_smile:

This acupuncture way of flashing, is it more reliable then a clip?

Well, with the ATtiny85 without bending pins I found that the clip would pop off every now and then, and was sometimes difficult to get on right. It was much better when I bent the legs but would still sometimes be annoying to get on correctly and could still pop off easily, at least when it was installed in a light:

The acupuncture style has it’s drawbacks too, in particular if the wires used are too thin. It has happened that I’ve had to push on the wires as contact was lost, but it happens so seldom so I much prefer this method. Now I can have the MCU on the inside, keep a spring pad and flash the driver without taking it out. That’s a real advantage for a firmware developer :slight_smile:

I have a dedicated development platform which makes development and testing much easier. Without the clip I find it even easier, but having the right size “needles” is key. The paper clips I use are pretty much spot on for the via size I used:


Those are some nice looking driver boards, Mike!

I have had that problem with the ATtiny85 clip popping off, but only pretty rarely. May have more to do with the solder flow on the legs. With bent pins, never a problem - rock solid clipping. Are you using the good blue clips?

Really like this progress though. Look'n good!!

The clips are blue, and mostly worked fine, but they might not have been the good ones. Sometimes the clips contacts would get bent and need straightening out. Possibly the stiff wire I ran over the 85s as battery contact had something to do with my issues, or possibly too little clearance to the 7135s:

Anyhow, these issues are now a thing of the past. I’ve fully converted to the 841 on all my drivers.

Small update for those interested… I’ve been testing and developing firmware for these v6 boards and so far so good. I’ve also made v7 series with some pin changes but it turned out that those changes weren’t really needed. I did make some other small adjustments so I’ll be using the v7 series anyway.

I’ve got off-switch and dual-switch functionality working fine, and I’m now working on E-switch only. It’s all the in same firmware but switch configuration is set by #defines. I did have a version where I could set switch type within the light’s config menu which worked well and was useful for debugging all three switch types in a single light, but that ate up too much memory.

On a single cell E-switch light I have got the parasitic drain down too low for both my DMMs to measure very well (Metex M-3650 and a Fluke 77). The lowest reading I can get from them is 0.1uA and that’s what I am seeing. When I pull the power to the driver it goes to 0 so the DMM is seeing something. Anyone know how accurate I can expect these DMMs to be at so low currents?

In any case I can’t get any lower as I’ve configured all pins and shut everything down for lowest power consumption while waiting for the switch pin interrupt to wake it up again. The datasheet specifies 1.3uA in power down mode but that’s with the WDT enabled. In power down mode I have it disabled. Anyway, I’m fairly satisfied the drain is low enough.

I haven’t built a multi cell LDO version yet but I expect the LDO to draw more than a powered down ATtiny841 in sleep mode. I’ll be building one of these shortly to try out if the AMC A705s can replace 7135s on 2S lights, so I’ll find out soon enough.

Thanks for the update. I don’t comment much on driver developments, but I’m watching with interest just the same. :+1:

Very nice Mike!