I added a short video above.
I’d be interested to see what your tests show. Mine was really worst case scenario, with 8 7135’s giving off heat, and the board just floating in air (because that’s what this light demands). Maybe they will fare better with proper heatsinking.
Very nice, good to see a real world test of the setup. Always nice when real world lines up with theory.
If the 7135’s don’t end up working it is no loss luckily. They can still be used for 2S setups and just left off for 4S.
It really makes me keep coming back to the op-amp driver though, I just might have to rethink giving that a go.
I am curious, has anyone built one of these drivers with a zener yet?
I am working on one today and first off realized I forgot to put the R5 resistor change in the parts list when using a zener. R5 needs to be upped to about 200 ohm for 2S cells. The nice thing is this should in theory also bias the voltage divider to the correct reading without needing to recalibrate if we figure out the precise R5 resistor value.
The real issue I am having is 2 fold.
First off the 1x 7135 works fine in moon mode and in low mode (approx 10% duty) but when I try to enter 100% duty on the 7135 it will not light up if the input voltage is the full 8.4V. If I turnt he voltage down to 7.4v or less it lights up fine.
The strange part is it is not the 7135, if I manually bridge the 7135 on with a jumper wire it comes on and regulates perfectly. So it is not the 7135 at fault here.
My first feeling is that the low ~4.2V from the MCU is not enough to energize the 7135 but then why is it able to turn on moon mode?
Not sure what to make of it.
No, and that's weird. About all I can come with is maybe the zener isn't operating in its specced range, but that's a wild guess.
It does make me realize I haven't thought about R5 enough, hmm well probably it's just the zener case that makes it an issue. Shoot with a zener mod I don't see how r1/r2 is measuring anything other than the zener plus schottky voltage actually, not the battery at all really.
Does LVP even work in zener mods?
Sorry, I didn't help your question at all.
Was trying to see another way without adding more resistors. Seems like if C1 needs a couple of ohms in series a tantalum cap will do that. D1 sure needs series resistance for a zener situation , but that can be below R1 then, but it's a circuit change, and I might still be missing something.
Another thought I was coming too a bit in the other thread, the bleeder resistor seems completely redundant with the LED+-R5-R1-R2 path (which I'm talking about making just LED+-R1-R2 anyway), just that R1 R2 is usually to big, because it's wasn't designed to leak current, and especially in e-switches without tailcaps is meant to limit current. But if using a tailcap just using very low values of R1 and R2 would create the bleed.
Given how much overlap of function causes complication everywhere already, as a general principle, I'm not sure it's a terrible idea that this is separated though. As a specific situation, I'm not sure there's a good reason for it either though. I suspect with the new OTC is a good time to be revisiting all these things. It may open up some cans of worms, maybe too many, incremental changes have strong points too, but on the other hand at some point things can also become very patch kit, and some big picture review might also be useful.
The voltage divider is before the diode and the zener is after the diode, so the zener doesn’t effect the voltage on the divider.
Heck the zener doesn’t even really do anything most of the time, the 200 ohm resistor does the actual voltage drop, the zener is there just in case basically.
Also I now have another interesting issue that I can’t figure out. I have somehow “killed” 2 separate drivers playing with this.
The strange part is that I can’t find anything wrong with either of them and they just stopped working all the sudden for no reason.
Each individual part tests ok, the 7135’s work fine in manually jumped, same for the FET. I replaced the MCU and yet it still doesn’t work. Even after reverting back to the stock setup and a new MCU, still nothing except a very fast flash when I apply power then nothing.
All the voltage checks I have run seem fine as well. No clue what the problem is.
The really strange part is if I flash the battcheck firmware, that works fine.
Have you looked at the manufacturer’s application notes? Maybe contact the manufacturer about the behavior. Don’t want to play mister obvious here; but just in case.
One problem is I have no idea who makes the 7135’s I have here.
The other is I do not even know what is “broken” at this point.
Maybe try a bit lower R5 as per our other discussion? To pull the zener into regulation. I can't even begin to say I have a working theory of the problem. It's darn weird, more just throwing darts now.
I doubt that is the issue, voltage is not dropping too low and it can’t go too high due to the zener. The MCU voltage normally changes based on battery voltage so not real worried about that.
Even reverting to the stock 1S format it still doesn’t work anymore and I have no idea why. The MCU tests fine, ditto for the 7135’s and FET and all voltages appear correct as well.
Seems like you're doing the right things. This probably is an experimental game of incrementally finding some known setup that works, some that doesn't and trying to close the gap with incremental changes until it breaks or is fixed. Ok, you tried a new mcu, but did you try to put one of the killed-circuit mcu's back into operation in a working circuit? Maybe your circuit is now frying every mcu you put in it. I'm not sure what "stock setup" means. If no flashlight with any attiny in it is able to turn on in your vicinity anymore, then you should be looking for alien tractor beams, or government transmitters implanted behind your neck.
By stock setup I mean the normal 1S setup that they were working with previously.
ok, so new mcu's no longer work in completely different tried and true circuits. Yeah, you've definitely got interference from the aliens and the government trackers, and maybe sunspots too. Some days are like this with electronics.
Are you using a battery or supply? The one common element left could be the supply I suppose.
If you can not identify the vendor of the 7135’s I have a suggestion. Buy a few of the parts from a known manufacturer. The online distributors such as Digikey and Mouser will identify the manufacturer. Then look at the manufacturers web site for data sheets and application notes. This can get you to a sort of base level. It may be that the manufacturer of the parts you have are not providing the expected parameters. If nothing else, checking the web sites of the vendors that you can identify as making the part would give you some base information of where the problem may exist. I suspect that Google would easily identify most of the manufacturers of the parts.
If you have a known manufacturer, you can then contact them about the problem.
I would also think you might look at the processor specs (and app notes) to see if any of it’s parameters are incompatible.
I know from long experience what a bitch it is when things don’t work like they ought to.
A test instrument to view and record what is happening on the various “wires” may be a help. If you don’t have anything, maybe someone in the project can loan you something. Or someone with the equipment could take a sample board and test it. There are also a bunch of pretty inexpensive devices that link to a PC to do tracing.
These parts have all been used for many many many lights and particularly in the 1S circuit he's referring to, and they work. No manufacturer is going to know or care what's going on with TA's setup today. How would they know what changed and why would they spend their time figuring it out for such a small customer? It's a mystery, but one he'll have to figure out I think. If we knew a very specific function that wasn't appearing to operate according to spec, then maybe there'd be something to query.
Have you removed and tested the 7135 chips on their own?
I had a similar problem with a driver and there the 7135 chips were destroyed.
No, just tested them by jumping the connections to manually tune them on.
My next step is to try the MCU on another PCB and see if it works then.
After that I just have to start replacing parts one by one. I will start with the 7135’s most likely. If I could just figure out which part is the problem it would make things way easier.
The strangest part is when I first hooked both of them up they worked just as they should except that the 7135 x 1 channel would not work in non-pwm mode. Only in moon and low. The FET and 6x 7135 channels worked fine in non-pwm.
Then after testing it a few times both drivers just suddenly stopped working for no reason and would not come on at all. I simply disconnected and reconnected the power as I had many times before.
The biggest issue right now is the time to really troubleshoot it.
Ok, just for your info, i tested the single 7135 chips using this board.
Makes it easy to test a lot of them in short time