Manufacturing a product with 18650 batteries

[Mad ID]

Hi everyone,
I'm designing a product which uses one 18650 cell without battery protection circuit.

The cell has small tabs for contacts and the battery is soldered directly to the PCB via the tabs. Tabs provide both electrical contact and serve as mechanical holders.

The battery is firmly in place and electrical resistance is minimal.  Battery protection IC is located on the PCB.

If would like to hear what you think of this solution, it has some benefits and shortcomings

Benefits:
- holder cost is zero
- very reliable electrical contact

Shortcomings:
- battery needs to be soldered
- no easy way to remove power from the electronic module

This last shortcoming worries me. After soldering the battery I need to program the MCU with a software which will go into low power otherwise all batteries would be drained.

Did you have a similar project and how did you solve the volume production / handling of batteries? I'm thinking of adding some kind od ship-mode or battery disconnect circuit. Is there a possibility of some kind of software bug or glitch that only power toggle (reset is available) can fix? Can think of none, that's why I ask

Thanks.

[T3sl4co1l]

Yes... Unless your low power state is really low power, a disconnect will be very good, especially for during storage!

A possible scenario for needing a disconnect, beyond a reset pin alone:
ESD on a logic pin, or miswiring causing power backflow.  In either case, the ESD protection diodes end up forward-biased, and under extreme conditions, you get CMOS latchup.  This can only be stopped by interrupting the power rail.

A current limiting switch might be wise?  These can auto-restart (fault current is drawn until the device turns off, then it waits a fixed time before turning on again), and come with enable/disable inputs and stuff.  They're pretty cheap, being used in huge quantities for USB port power and stuff like that.  (And, lots of battery managers with similar features.)

You can make a very simple, "no leakage" switch, by wiring up a complementary MOSFET pair like an SCR.  That can get you a voluntary on/off function, if not a current limited / protected switch.

Huh... I never thought about this before: thru-hole batteries aren't really a good idea to wave solder.  Or, if you make sure the pins are farther apart than the width of the solder wave, and the board is always soldered lengthwise (so the pins are soldered separately, not in the same wave at the same time -- shorting!), I guess that would be okay.  Otherwise, it's something that's always got to be hand soldered (which isn't a bad idea anyway, since, preheat? ew).

I've seen lots of stuff that uses packs/holders on plug cables, and nothing (so far) that used a hard wired 18650. (Packs on soldered leads, yes.  Close enough?)  Doesn't seem to be a problem for anyone else.  But, it's up to you, of course.

Tim

[station240]

You could put a jumper on the PCB so the battery can be isolated, for transport or testing.
If possible I would also use different shaped tabs, so the step where the 18650 is mounted to the board cannot result in wrong orientation.

Good point on wave soldering, you definitely do not want solder in the vent cap (positive end).

[Mad ID]

So all in all, no obvious drawbacks to derail the project

The batteries will be hand soldered. The low power state is really low power, and on top of that the PCM will disconnect the battery if over-discharged.

I will probably do a modification to the PCM circuit based around BQ29700DSER to allow for some kind of ship-mode just to be sure.

Thank you both.

[vealmike]

So all in all, no obvious drawbacks to derail the project

The batteries will be hand soldered. The low power state is really low power, and on top of that the PCM will disconnect the battery if over-discharged.

I will probably do a modification to the PCM circuit based around BQ29700DSER to allow for some kind of ship-mode just to be sure.

Thank you both.

Nothing wrong with your thinking here.
Please be aware of your legal restrictions when it comes to shipping.

Without UN38.3 approval for your battery system, you must ship your LI-ions as dangerous goods. You'll need explosion proof packaging and an expensive courier.

You may not ship by air freight if the battery has more than 30% charge.
There are limits on the packing density when shipping Li-ions (don't think this one applies if the battery is permanently fitted in equipment). These are IATA regulations.

You may not courier a faulty battery (or system containing a faulty battery) unless you can prove that it is safe. If you can't it's dangerous goods again. "safe" doesn't mean "I designed it to fail safe" it means I know that whatever the fault is, the safety circuit is still operating properly. It's a big ask if you have a dumb battery and a big seller for intelligent safety circuits that can be interrogated over I2C.

Yes, I know, yada yada, who cares? This guy's really anal. Well, breaching these rules is considered pretty serious. Fines are well into six figures, if you're selling this thing, tread carefully.

[Mad ID]

Yes, I know, yada yada, who cares? This guy's really anal. Well, breaching these rules is considered pretty serious. Fines are well into six figures, if you're selling this thing, tread carefully.

What is the difference of a device which has soldered battery and the one which has a battery connected via a small cable, both being transported in the product enclosure?

I'm aware of shipping restrictions, but don't see any difference here.

[vealmike]

There is no difference from a shipping perspective.

There are regulations for shipping batteries and regs for shipping product containing batteries. It's dumb, 'cus we all know that it makes sod all difference to safety, but rules is rules.

[Mad ID]

There is no difference from a shipping perspective.

There are regulations for shipping batteries and regs for shipping product containing batteries. It's dumb, 'cus we all know that it makes sod all difference to safety, but rules is rules.

I contacted the supplier and they sent me UN38.3 test report together with Certifications for safe transport by air and sea. This should help. I've send the files to our CE certification house for review.

[vealmike]

Can you get 38.3 on a raw cell? I'd have expected that you needed the safety circuitry in there.

Perhaps our supplier are telling us we need 38.3 as we have a pack. If that's the case, is a 38.3 cert for a raw cell still valid if the cell is shipped connected to a circuit?

[Mad ID]

Can you get 38.3 on a raw cell? I'd have expected that you needed the safety circuitry in there.

Perhaps our supplier are telling us we need 38.3 as we have a pack. If that's the case, is a 38.3 cert for a raw cell still valid if the cell is shipped connected to a circuit?

I would gladly answer that but I don't know. I got 38.3 for a bare cell. Report looks legit with pictures and everything. If you find out anything let me know and vice versa

[vealmike]

Cheers! They don't exactly make it easy do they?
Not only do us electron herders have to become chemists to understand the cells, we're expected to understand international air freight regulations as well.
Doing three jobs, paid for one.

[Mad ID]

Cheers! They don't exactly make it easy do they?
Not only do us electron herders have to become chemists to understand the cells, we're expected to understand international air freight regulations as well.
Doing three jobs, paid for one.

I look at it as part of the job
Will try to get as much info from my EMC lab as possible, they also do LVD testing of my product.