Battery packs and literature
-
I've got a situation at work. One customer needs our products individually solar-powered. So one of our electronics engineers got the task of integrating a battery and solar charger into our product. We were a bit time-constrained, so he decided the easiest way was to use off-the-shelf 18650 cells, with integrated protection ICs, in a 4s2p pack. With the individual cells in spring-loaded holders, which are connected to a backing PCB with the protector/balancer IC for the whole thing.
Now to me this screams
. Let us number the ways:- I'd understand soldering matched cells into pairs(*), but random off-the shelf cells, without testing them thoroughly?
- *=If there were no protection ICs between the 2 in a parallel pair.
- Spring-loaded contacts add a further resistance difference between the cells in a pair. And oxidation of the contacts, won't that be a problem over time?
But... is it wrong? I'm a software engineer, so I have some healthy awareness of the holes in me electronics knowledge, even if I have been able to cobble together a PCB every now and then.
So, my first question here is: Am I worrying for nothing, or is this an actual ?Also, even if I were correct in that it is a problem that will blow in our collective faces sooner or later, I'm going to have to make my case via literature. I'm arguing against a full-time electronics designer, for benefit of a superior who doesn't have a technical degree. And I'm doing it from a software engineering position. So I'm going to need a rather plainly-worder "don't do this" in black and white.
Does anyone happen to know handy literature for this?
(And if I was wrong, I assume the same literature can be used to educate me. So do please tell me if you have something that proves me wrong, too.)ETA:
Here's a picture to illustrate the current design:
-
@acrow Isn't it a bigger problem that spring loaded holders scream "user replaceable", but batteries with built-in protection ICs scream "the user is going to replace these with more readily-available unprotected cells"?
So, even if everything works out, there's a likelihood that you will have product failures because customers simply don't pay enough attention.
-
@acrow said in Battery packs and literature:
I'd understand soldering matched cells into pairs(*), but random off-the shelf cells, without testing them thoroughly?
If they're a matched pair, everything will be fine.
If one of the cells in parallel has more charge than the other one (typically because it is newer), it will attempt to charge the older cell of the pair. What happens at that point depends on the battery chemistry, but for most cells the effects vary between Not Good and Exciting!
You're right to be worried.
Spring-loaded contacts add a further resistance difference between the cells in a pair. And oxidation of the contacts, won't that be a problem over time?
That should be OK. Better to have good physical contact, and the springs are quite effective at that.
-
@Jaime said in Battery packs and literature:
@acrow Isn't it a bigger problem that spring loaded holders scream "user replaceable", but batteries with built-in protection ICs scream "the user is going to replace these with more readily-available unprotected cells"?
So, even if everything works out, there's a likelihood that you will have product failures because customers simply don't pay enough attention.
Since we sell strictly business-to-business, this is the one thing that I don't have to worry about: Either they replace them with the exact same thing, or they take responsibility of whatever happens afterward. On pain of lawyers.
-
-
@dkf said in Battery packs and literature:
If one of the cells in parallel has more charge than the other one (typically because it is newer), it will attempt to charge the older cell of the pair. What happens at that point depends on the battery chemistry, but for most cells the effects vary between Not Good and Exciting!
I read somewhere that a 20% difference in the cells' internal resistance will cut cell life 40%. Since the voltage in the cells is the same to within tens of millivolts, how much current flows out of each when the pack is discharged depends on their relative internal resistances. And when the drain stops, the cells will start to balance each other again which makes for, well, another stressful current.
But I can't remember where exactly I read it. And I'd need an authoritative paper to stand on anyway, to tackle the second problem. And I found nothing on how much the built-in protection circuitry affects the observed cell resistance either.
(And don't take this as me contradicting you. Just begging for help in finding resources.)Spring-loaded contacts add a further resistance difference between the cells in a pair. And oxidation of the contacts, won't that be a problem over time?
That should be OK. Better to have good physical contact, and the springs are quite effective at that.
Are you sure? My worry here is two-fold:
- A cell not making good contact for any reason will be masked by its parallel fellow; the pack seems to work initially. But the pack's capacity will be effectively halved. (The balancer is only capable of a few mA, so it's not able to mask a cell-pair falling to half capacity.)
- I've seen contact oxidation happen in flashlights a lot, which has probably lowered my confidence in spring-loaded battery setups. But if the oxidation should happen to one cell out of the pair, effectively cutting it off the circuit at a time when it's near full. And then, when it's partner gets drained and the voltage difference grows, this isolated cell finally punches through the oxidation layer. We'd have a full and an empty cell directly connected then. But how likely is something like this happening outside of cheap Chinese flashlights?
-
I'm totally clueless about all that (in our local jargon:
), but do you really have to worry about oxidation?What I mean is, while I've also seen it happening in e.g. flashlights, it takes quite a bit of time to happen, and/or a somewhat humid environment. So yeah, a flashlight that you have a for a few years, that's stored e.g. in the basement or some sort of outside shed, it's going to be an issue. But for example, I've never seen it happen in a TV remote that stays inside the house and is changed every couple of years (
shitty DVD player that never last more than 2 years + 1 day).You're saying "solar panels" so I assume it's going to be an outside device, but if it were an inside device designed to work for maximum a year or two, I'm not sure that would really be something to worry about?
-
@remi 10 years. Long warranty is a selling point, because the installation locations are often out of the way, so maintenance visits are costly.
And they're mostly installed around a coast. But I've seen stuff installed inland (but still outdoors) gather 2cm of water inside, even if they're IP67 tested to be "waterproof". Moisture finds the smallest flaws to seep through over time.
-
@acrow OK, corrosion is very likely going to be an issue during the life of the device, so forget I said anything.
-
@remi No, its a thing that I forgot to mention. Glad we got that sorted out. I tend to forget that not everyone works with equipment installed outside for years on end.
Well, it's not like e.g. PCBs corrode easily. If you have a splash-proof casing with a 1mm hole on the bottom, and at least some heating, you can leave a device outside almost indefinitely without notable corrosion damage. Unless you're right on the seashore. And even then lacquering the PCB may be enough to keep it working through the natural life of the constituent components.
Pity spring-loaded contacts and lacquering don't mix.
-
@acrow said in Battery packs and literature:
10 years
Do lithium cells last 10 years nowadays? With that charge/drain pattern, what is their expected capacity after that amount of time?
-
@PleegWat Good question. And here's the answer I got from the electronics engineer now that I just went to ask him: The cells should lose at most 20% of their capacity after 5 years. The drain and charge current are small-ish and fairly constant, so the wear on the cells will be light. ...He says.
And I guess after 5 years the customer's technician will have to replace the cells.
-
@acrow said in Battery packs and literature:
I tend to forget that not everyone works with equipment installed outside for years on end.
I know that you're likely not going to answer (and it's definitely starting to derail your help thread, sorry about that!), but now I'm curious to know what kind of stuff you work on...
-
@remi said in Battery packs and literature:
@acrow said in Battery packs and literature:
I tend to forget that not everyone works with equipment installed outside for years on end.
I know that you're likely not going to answer (and it's definitely starting to derail your help thread, sorry about that!), but now I'm curious to know what kind of stuff you work on...
I'd tell you, but I'd be doxxing myself doing it. It's a very niche field. Let's just say it involves packages of sensors and electronics, and is usually installed on towers. Sometimes with real-time monitoring, sometimes without.
I'll tell all as soon as I change companies, whenever that will happen. There's enough
going on in this company for one or two front page articles. But until then, sorry, I have to keep my job.
-
@acrow said in Battery packs and literature:
Let's just say it involves packages of sensors and electronics, and is usually installed on towers. Sometimes with real-time monitoring, sometimes without.
Well if the weather over this mystery ever unclouds I'm sure we'll be deluged with details.
-
@acrow said in Battery packs and literature:
It's a very niche field. Let's just say it involves packages of sensors and electronics
Interesting. Very similar to my job.
There's enough
going on in this company for one or two front page articles.Poor boy. I published here on the forum much more than that.
-
@acrow said in Battery packs and literature:
spring-loaded holders
This, to me, screams "regularly desired replacement" i.e. Phones before manufacturers decided two years of life was enough for anyone. If that's the case, and someone will be out to fuck with the deployed devices every few years (and replace only the batteries instead of the whole unit, for raisins) then that's not too bad of a choice I guess. I doubt there would be sufficient corrosion in that timeframe to significantly hamper performance more than it would without the springs.
On the other hand, I see mentions of demidecades and ponder if the Internet of Shit wouldn't make replacement effectively mandatory regardless...
