Battery packs, 3 questions mostly current-capability

Hi all,

There’s some generality here, but specifically this is for P639-F025-ND, batt pack with 10 C-size NiMH cells.

  1. How much current can the battery pack deliver and with what consequences. I’m not asking about the Panasonic cells, I’m asking about further limitations resulting from them being in the pack – most specifically related to the 9 interconnections between the cells, and to the 2 output tabs, and to anything else I’m overlooking. To be more specific, in one case: I expect to draw about 18 amps for 8 seconds; 0 amps for let’s say min 22 seconds; repeat that cycle up to 5 times; then 0 amps for many minutes or hours other than some low-current recharging. I want a) to be confident that none of the interconnections will be degraded, and b) to know how much resistance … like, milliOhms … is introduced by the pack (I’m not asking about the cells’ resistances, just the added interconnections). Or, if that resistance is dominated by the materials and not the welds(?), then if I can know the interconnection materials & length/width/thickness, I’m fine calculating it myself.

  2. What are the diameters of the holes in the + and - output terminals of the battery. I’ll be choosing wire gauge and deciding whether/what multiple wires into each terminal. We have battery packs but I’m not at the site where we’re keeping them, so I can’t measure.

  3. On this and other products’ DigiKey webpage, I see “Note Battery packs made with NIMH cells include a thermostat”. Could that actually be true, or is it a mistake for this pack and others? If it’s true, then I’m interested to know whether/how there’s electrical access, and/or other details.


-- brucep

Fair questions… I’ll ask around and see if there’s any actual documentation available, and float a couple of educated guesses in the meantime.

  1. I’d expect the interconnects to be less a factor than the cells themselves considering production variances, temperature dependencies, and performance shifts over time; if the use case is aggressive enough for it to be an issue, there’s a fair chance that there are bigger, uglier issues hiding in the bushes. Has any thought been given to the use of supercaps as an alternative approach?

  2. Scaling from the photo suggests a hole diameter in the neighborhood of 0.1" to 0.125".

  3. I suspect the term “thermostat” in this context to be a sub-optimal translation from documents originally written in Japanese, which has been faithfully reproduced… Some sort of bi-metal switching element in series seems likely, but a one-shot thermal fuse might also be a possibility. It would seem pertinent either way and a thing deserving of some documentation.

Got some more information; the interconnect strips used in this pack are 0.2" x 5mil nickel 1" long, which by my math yields about 2.7mohm apiece. The terminations are the same base material roughly 3/4" long with a 0.1" hole. Characterization of weld junction electrical characteristics is not available at this time.

The protection device is indeed a bimetallic switch type soldered inline with a nominal switching threshold of 60°C. Unfortunately, it’s only rated to switch 10A @ 12VDC, and has a derating curve which can be approximated as 1.5*I2 °C/A. Those are presumably steady-state ratings, but I’d doubt that the thermal characteristics of the switch would be sufficient to carry an 18A pulse load for 8s in close succession without significant risk of false opening at room temperature.

Due to the issue of the thermal protection device being under-rated for the intended application, I’d suggest that the stock battery pack may be ill-suited to the application as described. I expect that ordering a similar pack as a custom part assembled without this protection device would be a possibility.

1 Like

Hi rick_1976,

Thank you for your very good work! This is the first time I’ve used the Digi-Key TechForum, and I’m impressed with your initial conjectures, then followup with technical info & more reasoning, all with quick response time.

Unfortunately as you noted, some of the answers are problematic for us, so I’m looking into some possible directions. Yes some of the solutions involve a custom battery pack. I’m not particularly a battery/pack expert but in hopes of minimizing NRE, I would plan to design a pack and propose it to a company that does such work, and we/they would modify as it make sense, or they may choose to start from the ground up.

  1. I’m open to pointers to what company(s) might consider doing just a half dozen packs and no expectation of more … maybe the same company that did the existing pack.
  2. If I could, I’d like to have part number and/or the fuller specs of the cutout that’s in the device now, I might be able to search for a beefier one in a related family. Plus I might learn a little about cutouts … I may not be understanding the derating approximation you gave, but I’m inclined to think: at merely 3 amps, 1.5*(I^2) is 13.5 … so it’s 13.5 degC per amp … but I=3 so that’s 3*13.5 = 40.5 degC rise which would exceed the 60degC trip point with only the 3 amps? It seems 118-CB85A1B-ND is almost a nice choice, but I’m casting a wider net to find more current-margin.
    FWIW, I’m not strongly space-constrained which might help the cells thermally, although I currently have essentially no forced-air in that area of the overall device (which by the way is a medical research, Investigational Device, for National Institute of Mental Health).
  3. As for strap-connection resistance … the cell spec I’m using is at Nice that it shows a discharge curve at 20 amps, constant and much longer in duration than my 18 amps. It also says Rinternal about 5mOhm under conditions not so much like mine, but combining that with some analysis of curves, I’ve long been analyzing/simulating with 70mOhms for the 10-pack. For the Ni straps, my math came out the same as yours about 28mOhm for the pack, 35mOhm at 60degC. That’s a significant part of the total pathway resistance unless there’s more R than I see at this point, so I hope to explore increasing strap thickness or width. Plus, a) straps’ cross-sectional area is approx like 20awg wire, too small in my view for these currents, and b) being Ni, the straps’ resistivity is >4x that of Cu.
  4. We’re also using AA battery packs, I believe SY115-F025-ND or similar, at about 2.6 amps with the same timing as I described in original post. If you have info about its construction, that would be welcome too.

Actually you can skip #2 above after the first sentence, and skip all of #3. Just comment/criticize about them if you get a chance and want to.
Thanks again! for your other two postings even if you can’t help further at this point.

   -- Bruce P.

Nor am I… Competence is often a sufficient substitute for genuine expertise however, and it’s usually cheaper.

Battery packs like the one in question are assembled on-site at DK. If you’d send a note to specifying desired alterations to the standard stock number, I expect we could accommodate.

Wish granted.

Methinks you’re ending up with an I3 term there… The switch contacts have some finite resistance, hence dissipate some amount of power according to I2R, resulting in self-heating of the contact assembly. Thermal transfer to the surroundings is generally some 1st-order process, with a coefficient of (x)°C/W. Smoosh the two together, and you end up with self-heating being a function of current flow squared. Regression to the plot in the linked datasheet came up with a coefficient of about 1.5 for that squared term.

The cells themselves have a lot more thermal mass than the contact assembly of the thermal switch. I expect they’ll warm up some, but probably not too badly. The interconnect straps on the other hand, they’d be burning close to a watt apiece and seem likely to become rather toasty, with attendant increase in resistance. I don’t know if the folks in the assembly group have any thicker interconnect stock on hand, but it’s conceivable that one could use 2 or 3x and address the problem that way.

As for the SY115-F025-ND, the interconnect stock is .125" x 5mil x 0.78", connect tabs 0.156"W x 5mil x 0.6" with 0.062" hole, no thermal switch. This seems a much more viable use case than the other.

I’d again float the idea of a supercap-based approach, as delivering power seems more the issue here than storing energy. Devices such as the TPLH-3R0/400SS35X66 might merit attention.

One other wild-card option which might be viable are LiFePO4 batteries, such as the following:


These are protected battery packs in lead-acid style form-factors. They are a bit larger than the P639-F025-ND, are somewhat more expensive, and mass is slightly, to 50% greater than that pack, but they do have some compelling characteristics – specifically, high current discharge capable, fast charge capability, no memory effects, low self-discharge, and long life.