[UPDATE:v1.7.1,Q8&1chanOTSM]bistro-HD, bistro your way. OTSM, eswitch(devel), Vcc reads, safe_presses, turbo timeout...

I’m having trouble figuring out how to get Bazaar set up, so here’s a patch for bistro-HD.c instead.

Everything compiles that I’ve tested, but I can’t guarantee it doesn’t break other configs :P.

Google Drive Link

Hey folks, I feel a little bad abandoning this thread. It look like others have given it a little life, maybe even the originator of it all TK herself. I'll see if I can catch up soon.

I got one problem with 2S OTSM drivers with the 17mm TA board LDO I got it working
but with 22mm other LDO its not working at all

Cool, I'm a little surprised someone tried to get 2S working at all, I mean it should have worked though. I wrote up some details in the manual didn't I? It's actually been awhile since I thought about it, but I think it needs pretty tight tolerances on the resistors. Maybe tolerances and variation are the issue. I wouldn't think there is any difference in the 17mm and 22m. Do you know of any difference in the circuit for the two? Is it just the number of 7135's?

This is a good point for patching HD. You mean you tested compiling all the provided build configurations, just not with every conceivable config permutation? That's probably pretty good. It's good with changes to HD to run the full build process to see if everything still builds. If it does, it's some good assurance that the patch didn't break any of the config options permutations. This was a big part of the work in HD actually was cleaning up a bunch of older config options that had become incompatible and making them all compatible again.

Exactly right for OTSM. Somewhere I wrote the necessary versions, or you can find it in the attiny manual under BODS. In theory compatible versions exist for all three versions as I recall. In the real world, only the 45V's seem to be selling with the needed version level. You could call the 85v iffy, if by iffy you mean there are no known, to us, accounts of anyone ever receiving one of a late enough version to work. Presumably someday the early version stock will run out and that will change, but it could be years if ever. The 45V is the way to go for OTSM. (see next post)

You know I've confused myself on this now... been too long. I'm pretty sure my 25V did work. It's just the regular 25 that didn't. Yes, half the point of the HD was to get it to fit in a 25. I'm pretty sure this is right. You do have to check the revision letter on the back though to be sure. If it's a late enough revision, it will certainly work.

Edit, see post 69, yes I found the attiny25V's I ordered working, and I think even attiny25 non V's actually (it's coming back to me, an attiny25 non-V was my original test chip in fact where I first got BODS working, just dug the board out of the drawer to check) , although non-V isn't good for OTSM anyway.

So it's just the 85's that don't work. But do see post 69 for revision letters and be sure to check the revision code on your chip before wasting time soldering it.

As for stock, 25V's were listed on aliexpress last time I checked. Be sure to search for SSU to get the small footprint. It's not always very cheap to order just one there though.

Everything builds (at least all the default included configs) but I can’t test all the different builds.

Glad to see you’re back Flintrock. I’ve successfully built a couple 26mm TA LDO 3s OTSM and several 1s 17mm OTSM drivers. They all used the 25V and 45V. I have some 85v 10su (have not tried them yet) that came from arrow and mouser.

Oh yeah, I’ve been meaning to ask about the 47uf tantalum capacitors. The description for the p/n you gave says 47uf , 10 volts, 20%. If it’s rated for only 10volts what will be the long term effect of running it at 3+ cell voltage? Should we find one that’s rated higher for 3+ cell drivers?

The patch is simple enough, but I can’t build the whole suite of hex files unless I rewrite the Makefile to make it unix-compatible. That will have to be a task for when I’m a bit more awake.

Well, if you built it right, the OTSM cap should be under the regulated LDO output voltage. Yes, actually placing 12V on those caps would be risky at best, but that shouldn't be happening, even in 3S.

Hmm.. I've written more Makefiles in linux than windows, and didn't particularly notice the syntax difference, but usually I start with something and make changes, so easily could have missed it. It might be that all that is needed is to define the paths correctly though (like the batch file does), and/or a couple of command aliases. I'm not sure. I like linux, but for this atmel stuff, I don't know how anyone can do as much developing as you do without the gdb interface/simulator in atmel studio. It's just too great for understanding the machine state and finding bugs, and seems completely indispensable for embedded code like this where the only other testing option is to flash it and test things one blink and re-flash at a time. Is there a simulator for linux?

However, while I didn't provide a linux-tested Makefile, I did provide one that works without atmel studio from the command-line, which might at least make a linux person a little more comfortable. Still if you're not using windows then you're not.

Oh, it looked like I just need to remove the hardcoded paths and .exe extensions, and rewrite the shell portions for Bourne shell instead of DOS. It shouldn’t be hard, but I haven’t slept much lately because I’m trying to be diurnal and that doesn’t come naturally for me.

OTOH, programming without a debugger? That’s pretty natural for me. I haven’t even looked to see if there’s a simulator, because I’ve never really needed one. I can generally figure out what’s happening by simulating in my head.
The main parts I need hardware for are to blink out measurements from the sensors, which the emulator can’t do anyway.

Yeah, ok, there might be some slashes that need to be rotated 90 degrees. Sure even simple things require making time to do it. It would be nice if there's a way to generalize it to work. Dos might even accept the linux slashes these days and I'm not sure the command syntaxes are different other than the .exe's. There may be some other details though.

It may be fine. I didn't base that selection just on the label specs though. I looked for something with good DC bias curves and good temperature derating curves. Neither are shown for that. I didn't take any chances on the prototype because I wanted to make it work, not leave myself wondering why it didn't. Of the two I'd be more worried about DC bias. Even if it's worse it will still work, it may just struggle more for example with very low batteries and/or higher timing resolution settings. Or, it may work just as well. 10V is pretty high, most DC bias curves don't fall too much before 30% of rated voltage, so it's likely to be ok. Half price is certainly nice.

Trying to catch up on a few questions:

I'm not going to say it's impossible to get a zener working, but... I think usually you use a balast resistor to set a bias current on the zener right? If you skipped that and if the zener had low enough reverse leakage, maybe it could work, maybe (or I'm wrong and it's not even maybe). Anyway, I'm not taking credit for that mod. At your own risk.

Someone asked about blf emu issues... I can't say. I don't have any attiny13's and actually I never tested it on one. I think I did test an attiny25 version of it and had it working, possibly with a modified board layout to match my test board. Not sure what's going on there. hmm...

The 17mm and 22mm 2S just differed in the used LDO,
all other parts were exactly the same,
same firmware used

Well, the precompiled one was compiled by me on windows, but it was never about windows, it was about the version of the compiler used in atmel studio 5.0. I'm pretty sure (not 100%) with exactly the same compiler version and same arguments windows and linux will give bit for bit identical binaries. The precompiled hex's were compiled with atmel studio 7.0 not 5.0. I forgot the compiler version number. On 7.0 the few times I checked, I got within a couple of bits of what I was getting on linux, with similar compiler versions.

Well that can well be the difference. The LDO needs a very low reverse leakage current. This is not being picky. The diode I first used was draining a bunch of the cap's capacity, and it wasn't a bad diode. I called out a different diode from the usual TA one for a reason. And I don't know, maybe some of the standard TA ones also work, but maybe they all don't. I think I have a note in the manual or maybe somewhere around post 69 about how low the reverse leakage current should be, probably ~ 1uA preferably when warm if I remember right, and many are much much much higher. You really have to look at the graphs for the temperature dependence too as this increases quickly with temperature on some parts.

A problem with this OTSM stuff is that the chinese budget builds are likely to run into problems if they aren't paying attention to specs.

I am building the drivers with the TA parts list, I ordered only 50 LDOs of one kind
So if I would add a diode you have listed for the 1S driver after the LDO this would fix the problem?

R6 and 7 are just 0 Ohm resistors