kWeld - "Next level" DIY battery spot welder

[Hummina Shadeeba]

Anyone local to San Francisco who can fix these? I have two I’ve broken by contaminants getting in and shorting it. If u fix one u can keep the other.


I’m also desperate to borrow a welder to weld six cells to finish a battery.

 

[Zambam]

Anyone local to San Francisco who can fix these? I have two I’ve broken by contaminants getting in and shorting it. If u fix one u can keep the other.


I’m also desperate to borrow a welder to weld six cells to finish a battery.

What kind of contaminants and under what circumstance did the short occur?

edit: If contaminant was water (salt water the worst), probably not worth trying to repair. If you shorted out the power FET's or diodes, those can be identified with a DVM and replaced (if you can source the same quality parts).

 

[Zambam]

circuit board repair? Kweld to be specific.

So I made a bone head move and got distracted by something and I touched a meter probe to my kweld. I know I know, why is it not in a case, well I printed one but it was mirrored for some reason and I didn’t have any mor…

forum.esk8.news

 

[Hummina Shadeeba]

What kind of contaminants and under what circumstance did the short occur?

edit: If contaminant was water (salt water the worst), probably not worth trying to repair. If you shorted out the power FET's or diodes, those can be identified with a DVM and replaced (if you can source the same quality parts).

I’m not sure exactly what it was that did it but something slipped passed the kapton I had it wrapped in

 

[pegaz666]

Hi,
Can someone please explain to me why they’ve selected these particular Mosfets for kWeld?

So far I’ve seen in there:
1. Fairchild FDB0105N407L (discontinued now)
2. IOR AUIRFS8409-7P

There are 6 of them in parallel. People mainly seems to look at the max amp ratings in the datasheet BUT by looking at max SOA graphs, they BOTH should fail at given working conditions:
1. Drain to Source voltage = 12V (this is kWeld’s recommendation or 3s Lipo)
2. Welding time: usually the weld takes ~30-50ms (i know this device operates on joules rather then ms)

So: for a 30ms weld @ 12 volts we are looking at 2-3 Amps only if we want to operate within SOA even if there are 6 fets in parallel we’ve got ~20A max. Fairchild is a bit better as it indicates 6-7A but still we are light years from 2000A? They say in the datasheet 522A or 460A @ 10V BUT that’s for 10uS or 100uS ONLY not for times like 30ms+!!?? How does it even work without blowing the whole thing during the first shot? Could someone please explain this to me?


https://www.infineon.com/dgdl/auirfs8409-7p.pdf?fileId=5546d462533600a4015355b745e314f0

 

[TDA78]

Hi,
Can someone please explain to me why they’ve selected these particular Mosfets for kWeld?

So far I’ve seen in there:
1. Fairchild FDB0105N407L (discontinued now)
2. IOR AUIRFS8409-7P

There are 6 of them in parallel. People mainly seems to look at the max amp ratings in the datasheet BUT by looking at max SOA graphs, they BOTH should fail at given working conditions:
1. Drain to Source voltage = 12V (this is kWeld’s recommendation or 3s Lipo)
2. Welding time: usually the weld takes ~30-50ms (i know this device operates on joules rather then ms)

So: for a 30ms weld @ 12 volts we are looking at 2-3 Amps only if we want to operate within SOA even if there are 6 fets in parallel we’ve got ~20A max. Fairchild is a bit better as it indicates 6-7A but still we are light years from 2000A? They say in the datasheet 522A or 460A @ 10V BUT that’s for 10uS or 100uS ONLY not for times like 30ms+!!?? How does it even work without blowing the whole thing during the first shot? Could someone please explain this to me?


https://www.infineon.com/dgdl/auirfs8409-7p.pdf?fileId=5546d462533600a4015355b745e314f0

The Drain to Source voltage is not 12 V when the MOSFETS are fully on. They would measure about 0.183 V @ 2000 A (2000 A / 6 FETs * 0.00055 Ohms). You should be looking at the "limited by Rds(on)" part of the graph.

 

[pegaz666]

The Drain to Source voltage is not 12 V when the MOSFETS are fully on. They would measure about 0.183 V @ 2000 A (2000 A / 6 FETs * 0.00055 Ohms). You should be looking at the "limited by Rds(on)" part of the graph.

Thanks,
If that’s the case then why they say you can’t use more then 12V ? I wanted to use a 4s LiPo instead. Why input voltage doesn’t play any role in your equation? Thanks

PS. I don’t know about all these mosfets specs but someone was explaining to me these SOA graphs when i was trying to choose the new mosfet for my artificial load (atorch dl24p) - they were saying that the mosfet works there in a linear mode so DC part of the graph I should be looking at, so based on the above example of a FET at 10V input I could only put 2A max load on it if i wanted to drive it in constant mode, hence my thinking about reading this graph regarding spot welders but obviously spot welders are working only few ms as opposed to electonic loads which work constantly.

 

[TDA78]

Thanks,
If that’s the case then why they say you can’t use more then 12V ? I wanted to use a 4s LiPo instead. Why input voltage doesn’t play any role in your equation? Thanks

PS. I don’t know about all these mosfets specs but someone was explaining to me these SOA graphs when i was trying to choose the new mosfet for my artificial load (atorch dl24p) - they were saying that the mosfet works there in a linear mode so DC part of the graph I should be looking at, so based on the above example of a FET at 10V input I could only put 2A max load on it if i wanted to drive it in constant mode, hence my thinking about reading this graph regarding spot welders but obviously spot welders are working only few ms as opposed to electonic loads which work constantly.

The kWeld will work with up to 30V but 12V is ideal. Here's a quote from the designer on this subject:

"Desired current is 1500A. The resistance of the kWeld system is roughly 3.2 milliOhms (stock cables). The weld spot itself contributes another 1 milliOhm typical. This gives 4.2 milliOhms total. To push 1500A through that, you need U = R*I = 6.3V from an ideal battery. For every milliOhm of battery’s internal resistance, you need another 1.5V accordingly. Or in other words, for every volt above 6.3V that the battery has, it needs to have 1 milliOhm of internal resistance to stay in the sweet spot for the current. A 3S battery therefore needs to have 12-6.3 = 5.7 mOhms. Then calculate the power loss in the battery: P = 0.5 * R * I^2 = 6412 W. That explains why they heat up quickly, even though the pulses are short. Now consider a 6S battery. That must have an internal resistance of 24-6.3 = 17.7 mOhms. Then P = 19912W. You see that the power loss in that battery is huge. This is why I recommend 2S to 3S batteries for the welder."

The operating manual has a deep dive into the theory of operation.
https://www.keenlab.de/wp-content/uploads/2018/07/kWeld-operation-manual-r3.0.pdf

Spot welders must switch past the linear operating region of a MOSFET as fast as possible to survive.