Why does the cell log have a max 4.9v limit?

[Kin]

Hi everyone,

At work some guys wanted me to set something up to easily measure 9 voltages at once. At first I was excited by the prospects of throwing together an easy celllog setup, splitting the 9 voltages over two different cellogs (5 on one, 4 on the other).

Today it came to light that the max voltages are actually as high as 5.7v, which was not what I was originally under the impression.

So I have been looking through the cell log hack threads and trying to understand why the cellog limit is 4.9v. Last time I went through the threads I saw an easier to read schematic of the op-amps, and I could sort of work through that, but I can't find it right now. The best thing I found was Doc's pictures on viewtopic.php?f=14&t=12815&start=15, but I'm not familiar enough with electronics to understand the pictures of the tiny components.

I also remember there was a component that could not deal with 32v or something around that level, but technically I think my 5.7 split would put max 28.5v on any single Cell log (since I'm splitting the 9 voltages, 5 on one cellog, 4 on the other). Yet, not having that old schematic, I don't know if the split would help at all.

TL;DR Version
: With that in mind, I'm asking if anyone knows why the cell logs have a 4.9v limit, so I can get an idea whether or not 5s of 5.7v units (28.5v total) would kill the unit.

In the worst case, I can use voltage dividers, but right now I do not have the necessary very low tolerance resistors I would need on hand (precision is pretty important for this use), I'm not 100% sure how the load that the Cell log draws would affect a voltage divider, and, well, it would be kind of messy on each of 9 wires :-/. But! I will do this if I can't get around the 4.9v limit.

[neptronix]

...hm....

what kind of cells are you using that are above 4.9v per cell?!?!

[Kin]

Not cells, actually! I removed the part where I mentioned the use, because I figured it wouldn't be interesting to anyone. It's 9 peltier modules in series. They're matched by measured impedance, but then tested again through several heat cycles, this time to see that the voltages through all the peltier modules in series are similar (which should mean they're still well matched). Right now someone painstakingly uses a single voltmeter to read, record, read, record, and tries to do it very quickly so they can be within a decent timescale (voltage moves over time).


Another forum member mentioned using voltage dividers to isolate some 12v lead acid batteries under load, but I'm hoping I might not need to use voltage dividers.

It's been surprising to me how much Ebike tech and RC tech proves to be useful outside of these two hobbies, simply because decent quality equipment is made in such large scale for those fields. I hope I'm not bothering anyone to ask this question that I'm not directly using for Ebiking. I've tried to do all the research on the cell log hack threads, but maybe I just missed the right piece or the right thread, because I never found an answer.

[heathyoung]

Have a look at my post in the cell-log hacking thread in regards to voltage dividers + opamp (circuit provided) - used it on 8S 12V batteries just fine.

You can calibrate each channel individually to get past the resistor tolerances (or intolerances as the case may be).

[Kin]

thanks heath young. I'll look into that, and good call on the calibration option, I forgot about that. I'll probably go that way rather than blow a cell log without knowing if it has any chance of tolerating a higher voltage. Still would be a lot neater ofa package to avoid the resistors. I also cnsidered signal diodes but didn't find a consistent drop.

[dogman]

Sounds to me like a bank of inexpensive voltmeters would be a better thing for this application.

By the time I say 4.9v on a lipo cell, I'd be getting out of there. So perhaps the cellog designers never considered above 5v to be needed.

[bigmoose]

Here is a drawing of what is believed to be the power supply layout of the Cell Log 8. Note that the lower LM324 is powered off of 6 cells. It has a maximum Vcc of 32 volts. 32/6=5.3 volts/cell maximum. This drawing shows a LM358 up top also with a max Vcc of 32 volts. It is powered from cell 8, but ground is supposed to be the 3 volt MCU rail. so (32-3)/8=3.6 volts/cell maximum. You can add the diode drop if you want to use that 0.6 volts.

It appears they are overstressing the upper LM358 (about 35.6 Vcc) at a rated 4.9 volts/cell. I did not reverse engineer the schematic, and cannot verify that it is correct. It is off of ES by a member.

CellLog 8M front end schematic.jpg (60.03 KiB) Viewed 357 times

[Kin]

Ah! That is the schematic that I found the first time I was looking, but which I couldn't find last night. Thanks.


It seems to follow somewhat my memory of it. I'm not very experienced with electronics, but I'm thinking that I might be ok with my intended use, because although individual units will be 5.7v MAX, no cell log will see more than 5 of them (5*5.7 = 28.5v). Which is below the 32v limit for LM324. Even if I misunderstand the implication of 3v MCU, 32v-3v /5 = 5.8v [though quite close to my max limit]. If I'm not mistaken, I will not have to worry about LM358 because there will be no cell 7, 8, (or for that matter, neither cell 6).

Hopefully, there's no firmware limit to the 4.9v limit.

Now I'm surprised that the jumper mod (sharing power over all 8 logs) doesn't damage the unit, since if 8 cells are at 4.2v, that's 33.6v, and above the limit of LM324. However, I suppose that should not matter for me, and I suppose it suggests I don't fully understand what's a'happening.


Thanks again for finding the schematic, Bigmoose (while I acknowledge the risk that it might not be accurate). I wish we had someone around here that wasn't just a kid with electronics, like me, but it's a small place and doesn't have anyone (which is part of the fun of getting to work here, and the only reason I'd be paid while knowing so little).

Dogman, as for using a bank of voltmeters I have considered that. But, i have not found inexpensive meters that are accurate past the first digit. Maybe I'm not looking in the right place, and it probably doesnt help that I've just looked in a few nearby stores and not online, where things are actually affordable. But I would like to get three sig figures (like, 4.50v), and the cheap meters ($10), seem to suck. While we could probably even spend $180 on the 9 meters, in which case we could get 9 halfway decent meters that could do three sig figs, there would be another big problem that the cheap meters don't have auto-off, and that a set of 9 meters would be much more bulky on the workbench.

[dogman]

Yep, It would take costly meters to get more accuracy. But the cellog 8 does seem to be the better of the cheap lipo meters, more accurate than any of my chargers anyway.

Part of why I don't sweat .05 off on my lipos. May not even be real.

[Kin]

True, and I did just make a quick search on the fleabay, but I couldn't find anything cheap that would also have a hold function. Things are accurate enough, though. Celllogs I agree are nice; I consider them to be sufficiently accurate for my use.

[John in CR]

: With that in mind, I'm asking if anyone knows why the cell logs have a 4.9v limit

It's because they're dreaming of the day when we have higher voltage cells. Otherwise the limit should be lower.

[Kin]

: With that in mind, I'm asking if anyone knows why the cell logs have a 4.9v limit

It's because they're dreaming of the day when we have higher voltage cells. Otherwise the limit should be lower.

To be very fair, I was going for the physical reason, not the meta reason :p. It's a joy to boast of technical capabilities beyond what is needed, so I've been told. If you're sitting on a 4.9v LiPo, anyway, I hope you're not in a wooden house.

[Kin]

Aannnnnnddd, the verdict is, (finally just decided to potentially sacrifice one),

There's a firmware limit. The cellog did not die when I put the higher voltages in (5.5v, but only 5 lines, so it was below the upper limit), BUT, the cellog booted up and went "Cell Volt Err"

I think if I'm going to go do some voltage dividers, I may as well use a multiplexer and run this whole business on an Arduino. I can couple it to a thermal printer so it can do the few calculations that the numbers are used for, and poll the voltages when the temperature hits the temperature needed.