Automated assembly of CR2032 coin cell to PCB

[VintageHenk]

I'm designing a relatively high volume (~500k), very low cost CR2032-powered board that needs a flush bottom side and will go inside a potted enclosure.

Normally you would just do a reflow for the SMT parts, use a THT tabbed CR2032, a second pick-and-place round for the battery and wave solder it to the PCB, but as the bottom needs to be (relatively) flush, I can't have the battery leads sticking out. I'm also trying to avoid the cost of using a battery holder as the battery isn't replaceable anyway. I'd preferably use something like the SMT-tabbed coin cell in the attached image but that would need to be hand soldered (?), which isn't great at these volumes. I have also found sheet metal formed side entry coin cell holders (see attached image for an example) that can meet my cost targets, but I don't think a pick and place machine can shove the cells in sideways, so that would require expensive labor and/or a jig, which again isn't great.

How would you guys and gals go about automating the attachment of the battery cell to the PCB? Any secret tricks of the trade? I'm still in the design phase, so everything is open.

[Kean]

At those volumes you should be discussing options with a battery supplier rather than looking at existing off-the-shelf offerings.
They would have seen similar requirements before, and can likely offer some custom termination and packaging to better suit mass production.

[Kean]

Further thought
As reflow isn't compatible with batteries, and wave soldering isn't going to work for your requirements - maybe conductive glue on the top-side pads before placing the tabbed cell.  The dispensing and placement can all be automated pretty easily.  I'm guessing the current demands on the battery are pretty low so the joint resistance isn't a huge problem, and the potting will add extra mechanical strength so you wouldn't need to fully rely on the conductive glue.

[David Hess]

Normally I would suggest using the flat battery holder on the top even though the battery will never be replaced, but potting a battery holder risks having the potting material push on the mechanical connections enough to disconnect the battery.

Is there any safety issue with venting when a battery is potted?

Is there an alternative to a coin cell battery which would be more suitable?

[Psi]

Avoid using any method that uses a PCB pad as one of the battery contact. Such as in your 2nd pic.

It's very hard to get it to work reliably. The connection tends to go bad and become high resistance. Combine that with the fact you're using a CR2032 that already has pretty high internal resistance and you have a recipe for lots of failed products. Or products that think the battery is low when it's still ok.

Best to find a tabbed cell that will work, either TH or SMT.

[tomgat]

I'm designing a relatively high volume (~500k), very low cost CR2032-powered board that needs a flush bottom side and will go inside a potted enclosure.

Normally you would just do a reflow for the SMT parts, use a THT tabbed CR2032, a second pick-and-place round for the battery and wave solder it to the PCB, but as the bottom needs to be (relatively) flush, I can't have the battery leads sticking out. I'm also trying to avoid the cost of using a battery holder as the battery isn't replaceable anyway. I'd preferably use something like the SMT-tabbed coin cell in the attached image but that would need to be hand soldered (?), which isn't great at these volumes. I have also found sheet metal formed side entry coin cell holders (see attached image for an example) that can meet my cost targets, but I don't think a pick and place machine can shove the cells in sideways, so that would require expensive labor and/or a jig, which again isn't great.

How would you guys and gals go about automating the attachment of the battery cell to the PCB? Any secret tricks of the trade? I'm still in the design phase, so everything is open.

I have used SMT batteries holders on most of my designs for similar reasons to what you describe.  The PNP machines do great on placing them.  I have rarely used a physical battery, however, unless the design is simple enough that I can have a high confidence that the voltage area is isolated and cannot bridge on reflow.  In addition, using a physical battery might further complicate your testing harness and test pin pad locations, so this might be something to consider as well. 

Also, as far as "tricks" to using them... Just my 2 cents so take it for what its worth... I either expand the pad or use a stepped stencil for components like this because I have noticed in the past that the big flat leads on the SMT batteries suck up allot of paste. I only started doing this after customers reporting a small percentage detaching in the field. 

[VintageHenk]

Thanks for the suggestions everyone, very useful stuff!

Further thought
As reflow isn't compatible with batteries, and wave soldering isn't going to work for your requirements - maybe conductive glue on the top-side pads before placing the tabbed cell.  The dispensing and placement can all be automated pretty easily.  I'm guessing the current demands on the battery are pretty low so the joint resistance isn't a huge problem, and the potting will add extra mechanical strength so you wouldn't need to fully rely on the conductive glue.

That's an option I hadn't considered, current is indeed low so could be an interesting solution! I've never used conductive epoxy before, so would need to do some prototyping to see how well it works.

Normally I would suggest using the flat battery holder on the top even though the battery will never be replaced, but potting a battery holder risks having the potting material push on the mechanical connections enough to disconnect the battery.

Is there any safety issue with venting when a battery is potted?

Is there an alternative to a coin cell battery which would be more suitable?

You raised a good point regarding venting - I've looked it up and found that CR2032s are sealed so shouldn't be a problem (batteries like these are also potted in RTC modules after all), but on an Energizer datasheet, I did find the recommendation to contact them before attempting potting batteries to discuss possible failure points.
I did also look into different battery types, but couldn't really find anything that would meet cost, size, longevity and availability targets.

Avoid using any method that uses a PCB pad as one of the battery contact. Such as in your 2nd pic.

It's very hard to get it to work reliably. The connection tends to go bad and become high resistance. Combine that with the fact you're using a CR2032 that already has pretty high internal resistance and you have a recipe for lots of failed products. Or products that think the battery is low when it's still ok.

Best to find a tabbed cell that will work, either TH or SMT.

That sounds like advice learned the hard way, so I'll stick to soldering/gluing tabbed cells or getting a proper dual contact holder. Thanks!

I'm designing a relatively high volume (~500k), very low cost CR2032-powered board that needs a flush bottom side and will go inside a potted enclosure.

Normally you would just do a reflow for the SMT parts, use a THT tabbed CR2032, a second pick-and-place round for the battery and wave solder it to the PCB, but as the bottom needs to be (relatively) flush, I can't have the battery leads sticking out. I'm also trying to avoid the cost of using a battery holder as the battery isn't replaceable anyway. I'd preferably use something like the SMT-tabbed coin cell in the attached image but that would need to be hand soldered (?), which isn't great at these volumes. I have also found sheet metal formed side entry coin cell holders (see attached image for an example) that can meet my cost targets, but I don't think a pick and place machine can shove the cells in sideways, so that would require expensive labor and/or a jig, which again isn't great.

How would you guys and gals go about automating the attachment of the battery cell to the PCB? Any secret tricks of the trade? I'm still in the design phase, so everything is open.

I have used SMT batteries holders on most of my designs for similar reasons to what you describe.  The PNP machines do great on placing them.  I have rarely used a physical battery, however, unless the design is simple enough that I can have a high confidence that the voltage area is isolated and cannot bridge on reflow.  In addition, using a physical battery might further complicate your testing harness and test pin pad locations, so this might be something to consider as well. 

Also, as far as "tricks" to using them... Just my 2 cents so take it for what its worth... I either expand the pad or use a stepped stencil for components like this because I have noticed in the past that the big flat leads on the SMT batteries suck up allot of paste. I only started doing this after customers reporting a small percentage detaching in the field. 

Thanks! Again sounds like knowledge learned the hard way, so I'll make sure the pads are big enough not to get detaching contacts. As for the design - it's pretty low BOM count (~10 parts) and not that critical on density, so making space for test points and preventing bridges shouldn't be a problem.


So if I tried to get wave soldering working - do you think I could get away with ordering custom tabbed cells with leads that are trimmed just a bit shorter than the PCB thickness, so they won't stick out? Or would that compromise yield too much? Maybe someone has tried that before, or with different THT parts?

[tomgat]

So if I tried to get wave soldering working - do you think I could get away with ordering custom tabbed cells with leads that are trimmed just a bit shorter than the PCB thickness, so they won't stick out? Or would that compromise yield too much? Maybe someone has tried that before, or with different THT parts?

The TH placement machine will do this for you.  When the Th component is placed, there is a rotating head on the backside of the the machine that bends and cuts the lead to hold it in place for going through the wave solder process. 

However, if you only have 10 components on the board and space is not critical, I would think that an SMT approach would be allot cheaper for the kinds of quantities you are talking about.  The problem with putting the physically battery on the board is that it not only complicates the assembly/testing, but also shipping.  No one wants to ship large quantities of batteries without all the paperwork and insurance that comes along with it.  Meaning, it would likely be cheaper and easier to use the SMT battery shell, and then pay a worker to put in the batteries on the boards once their onsite. 

[Eric_zhang]

In response to the challenges of automating the assembly of a CR2032 coin cell to a PCB, especially for a high-volume, low-cost application with a flush bottom requirement, here are some suggestions:

Custom Termination: As Kean mentioned, discussing with a battery supplier could be beneficial. They might offer custom termination and packaging solutions that are optimized for automated assembly at your production scale.

Conductive Glue: The use of conductive glue on the top-side pads before placing the tabbed cell could be an automated process. This method would require less handling and could potentially be integrated into your existing assembly line with minimal changes.

SMT Battery Holders: As tomgat pointed out, using Surface Mount Technology (SMT) battery holders can be an efficient solution. PNP (Pick and Place) machines can handle these quite well, and it might be worth exploring this option further, especially if you can ensure that the voltage area is isolated during reflow.

Avoiding PCB Pad Contacts: Psi's advice to avoid using PCB pads as battery contacts is crucial for reliability. The connection can become high resistance over time, leading to potential product failures. Sticking to tabbed cells, either through-hole (TH) or SMT, is recommended.

Wave Soldering with Custom Tabbed Cells: The idea of custom tabbed cells with leads trimmed to the PCB thickness is interesting. However, it might introduce additional complexity and cost. The rotating head on the TH placement machine can bend and cut the lead during assembly, which could be a more reliable approach.

SMT Approach: Considering the low BOM count and not being critical on density, an SMT approach could be more cost-effective. It would also avoid the complexities and regulations associated with shipping batteries.

On-Site Battery Insertion: If all else fails, using an SMT battery shell and then having a worker insert the batteries on-site after the boards have been assembled and tested could be a viable option. This would eliminate the need for dealing with batteries during shipping and could simplify the assembly process.

In conclusion, it seems that a combination of custom solutions, such as SMT battery holders or conductive glue, might offer the best path forward for automating the assembly of the CR2032 coin cell（e.g. [spam link removed]）to your PCB. It's important to consider the entire production process, from assembly to testing to shipping, to find the most efficient and reliable solution. Prototyping and testing different methods will be key to finding the best fit for your specific application.

[toybuilder]

There are robotic "hand soldering" CNC machines that will feed solder to an iron tip.  (just an example that I found from a quick search)
The challenge is to find an assembly shop that is equipped with one that you can afford.

[Bud]

Awesome tip cleaning system @1:25, me wants one 

[Kean]

Awesome tip cleaning system @1:25, me wants one 

Slightly off-topic, but yeah me too.

JBC make some - the CLMS or CLMU, which automatically sense to tip insertion.

Another is the Aoyue 128 (also sold by Digikey as the SRA AO128) but it isn't automatic.