JP spot-welder, FET-switched, timed adj. pulse

okashira

10 kW
Joined
Nov 17, 2014
Messages
963
Location
Houston, TX
[this post is labeled as being from "okashira", but I have split this topic into its own thread with approval from Riba, and taken over this post in order for the first post to reflect this threads' mission and content. The okashira content was merged into the next post, -spinningmagnets]

ES member riba2233 (from Croatia) was selling a low-voltage/high current spot welder, and also selling less-expensive kits for customers who want to save money by doing the assembly themselves. Due to health and family concerns, he is no longer selling them.

"JP spot welder"
https://endless-sphere.com/forums/viewtopic.php?f=31&t=68865

riba2233's "sales thread" has devolved into a scramble to find ways to copy this type of DIY spot-welder, so I am splitting this off into its own thread in the battery section.

Since nickel strips are in physical contact with the 18650 cell-ends, high voltage is not necessary. 12V automobile starter batteries are common, and are already capable of high temporary currents. What makes this style of spot-welder interesting is the feature-set it provides compared to common industrial desk-top spot-welders, which typically need a high-amp power supply to load a large capacitor bank. Common wall-sockets in homes are sometimes overloaded by the demands, which customers may not find out until after they have purchased a spot-welder.

Since the voltage does not need to be high, some builders are enjoying satisfactory results using lithium batteries for the current using 2S and 3S packs, with recycled 18650's or LiPo (8V, 12V).

When dealing with high-amp DC, the on/off switch can experience arc-ing, and in a worst-case scenario, the contacts can be welded together themselves. With no way to turn the DC device off, it can lead to a spectacular melt-down. This style of spot-welder uses MOSFET electronic on/off switches, which are common inside high-powered ebike controllers.

The length of time that the current is "ON" is vital to control in order to achieve consistent results. It should be long enough to make a solid connection, but not so long that excessive heat is built-up and damages the internal cell chemicals. The length of time that current is flowing is called the "pulse", and it is measured in one-thousandths of a second, called milli-seconds. Thicker metal bus-strips need a longer pulse, thinner strips need a shorter pulse. This device uses a programmable microprocessor to adjust the pulse-width.

Content in this post from below this line is from ES member okashira:

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

The welder works fine as-is. You can weld up to 0.2mm nickel easily with the kit supplied by Riba using a lead acid starter battery. You can even weld 0.3mm nickel, but be mindful that the mosfets stay as cool as possible and don't use too powerful of a battery.

The diode is an upgrade for those of us that want 2000 amps and longer welding leads and want to weld lots and lots of 0.3mm nickel. Just use a lead acid starter battery. Don't worry about amps, S, P, whatever, just use a battery like Riba recommends and it will be fine. You can probably use the one from your car.

This is the BEST 18650 spot welder you can buy for under $2,000. Period. Your next option is a Sunstone Engineering welder. This one can work just as well.
 
It really took me a while to figure this out, but I finally figured out how to fix the problem when using the welder at higher currents with longer welding leads.
I have blown over 100 mosfets so I got tired of it...

You can attach a high current diode like shown in black:
file.php

The black wires should be as short as possible to minimize inductance. Now the inductance of the welding leads can be high and not cause anymore mosfet failures.

Since I have a bunch of this awesome aerospace grade 4awg silicone wire, I might offer a high end welding packing upgrade for riba's welder - longer super flexible leads/electrodes + diode

One side of diode gets attached to JP welder "welding lead" side.
Other side to battery positive terminal.
Diode points in direction of battery.
 
okashira said:
It really took me a while to figure this out, but I finally figured out how to fix the problem when using the welder at higher currents with longer welding leads.
I have blown over 100 mosfets so I got tired of it...

You can attach a high current diode like shown in black:
file.php

The black wires should be as short as possible to minimize inductance. Now the inductance of the welding leads can be high and not cause anymore mosfet failures.

Since I have a bunch of this awesome aerospace grade 4awg silicone wire, I might offer a high end welding packing upgrade for riba's welder - longer super flexible leads/electrodes + diode


Can we get some more details on this? I may have to increase my current because I want to weld .3 nickel. If placing a simple diode will help me increase my current without blowing my fets then I will do this.

Just need more information about this and if it really works? How much current can you now safely use with this? Like can I use 1400 cranking amps instead of the 1100 limit set by riba?

Pictures of a completed setup and part number for a suitable diode will also be nice.

Thanks
 
Offroader said:
Can we get some more details on this? I may have to increase my current because I want to weld .3 nickel. If placing a simple diode will help me increase my current without blowing my fets then I will do this.

Just need more information about this and if it really works? How much current can you now safely use with this? Like can I use 1400 cranking amps instead of the 1100 limit set by riba?

Pictures of a completed setup and part number for a suitable diode will also be nice.

Thanks

Same here. Would like to be able to see the original pic posted but it is no longer on the ES server.
 
Fixed the pic.
I'm using about 2000+ amps but only get a few hundred welds before FET's blow
This diode should work fine: http://www.mouser.com/Search/ProductDetail.aspx?R=DSA300I45NAvirtualkey65120000virtualkey747-DSA300I45NA
Mine will come in tomorrow and i'll hook it up and post pics/results.
 
Offroader said:
okashira said:
It really took me a while to figure this out, but I finally figured out how to fix the problem when using the welder at higher currents with longer welding leads.
I have blown over 100 mosfets so I got tired of it...

You can attach a high current diode like shown in black:
file.php

The black wires should be as short as possible to minimize inductance. Now the inductance of the welding leads can be high and not cause anymore mosfet failures.

Since I have a bunch of this awesome aerospace grade 4awg silicone wire, I might offer a high end welding packing upgrade for riba's welder - longer super flexible leads/electrodes + diode


Can we get some more details on this? I may have to increase my current because I want to weld .3 nickel. If placing a simple diode will help me increase my current without blowing my fets then I will do this.

Just need more information about this and if it really works? How much current can you now safely use with this? Like can I use 1400 cranking amps instead of the 1100 limit set by riba?

Pictures of a completed setup and part number for a suitable diode will also be nice.

Thanks

The FET's are only rated for an avalanche current of 150A peak. That puts welder current as designed to 900A peak, because you can expect full current of the pulse to flow thorough the mosfets when they're turned off. (inductance sucks) even if it's for100uS.

The Diode circuit gives the current somewhere to go once the FET's are opened... no more avalanche current. Of course there is inductance in the diode diode, but it's much smaller then the 4-6 feet of welding leads.

I was so focused on SOA curves when I was helping pick MOSFET's, I finally realized it's the avalanche current after turn off that kills them. That's why when we tried higher voltage MOSFET's they would blow even faster. Because the avalanche power = Vdsrated*I. You want as low voltage MOSFET as possible
 
okashira said:
One side of diode gets attached to JP welder "welding lead" side.
Other side to battery positive terminal.
Diode points in direction of battery.

I think that's a good idea. I've thought about that too.

You could use a couple of FETs for the diode. Just tie gate to source and use the body diode.
The body diodes in the FETs used in the welder can handle pretty high currents for a short pulse.
 
I got the diode hooked up and It seems to be working.
Been running at max power (20ms) welding 0.3mm nickel like a champ in rapid succession. I have a special firmware with a 600ms delay so I'm welding FAST.
Even welded some nickel right to the Model S Aluminum bus bar. Pulling the nickel off pulls a layer of aluminum right off the bus bar. I was bypassing one of the leads, and the one welding lead was jumping like CRAZY. I'm guessing 3000+ amps.
Here is the setup:
file.php

All connections are cleaned with a wire brush and clamped to hell, tested them at about 20 micro ohm.
The diode wire is way longer then needed, but still way shorter then my welding leads. The mosfets are getting noticeably less warm. I can weld at max power in rapid succession and they just get luke warm.
I'll shorten the diode wire later. Running about 8.15V, I will go ahead and push it to 8.3, then go with an A123 pack at 3s = 9.9V.
 

Attachments

  • welder diode.JPG
    welder diode.JPG
    248.3 KB · Views: 6,181
riba2233 said:
I would rather spend time developing my weldless cases idea :)
YES! Please!
 
okashira said:
Fixed the pic.
I'm using about 2000+ amps but only get a few hundred welds before FET's blow
This diode should work fine: http://www.mouser.com/Search/ProductDetail.aspx?R=DSA300I45NAvirtualkey65120000virtualkey747-DSA300I45NA
Mine will come in tomorrow and i'll hook it up and post pics/results.

Is that diode getting hot at all when you push it hard? If not I may try a smaller schottkey I already have on hand...
 
markz said:
jibberish

OMG, really? Just remove your post. you are going to confuse people. I an well aware of the physics of electromagnetism the way you changed the diagram is WRONG
Even in electrical engineering they use what you call "laymen's" current flow. Look at a a diode symbol.
 
I found this video good in explaining why circuits are designed with conventional flow, which isn't the same as electron flow. Really, best to forget about the whole electron flow because it will just confuse you.

This video he explains why we still use the conventional current, basically because we always used it that way and it isn't actually wrong to look at it that way.

[youtube]fBNLzgr9T6w[/youtube]
 
ecycler said:
okashira said:
Fixed the pic.
I'm using about 2000+ amps but only get a few hundred welds before FET's blow
This diode should work fine: http://www.mouser.com/Search/ProductDetail.aspx?R=DSA300I45NAvirtualkey65120000virtualkey747-DSA300I45NA
Mine will come in tomorrow and i'll hook it up and post pics/results.

Is that diode getting hot at all when you push it hard? If not I may try a smaller schottkey I already have on hand...

A smaller diode might work, depending on how much inductance there is. If the diode blows, it could be a problem though. It might be worth placing a fusible link in series with the diode. It will be pretty obvious if that blows.
 
ecycler said:
okashira said:
Fixed the pic.
I'm using about 2000+ amps but only get a few hundred welds before FET's blow
This diode should work fine: http://www.mouser.com/Search/ProductDetail.aspx?R=DSA300I45NAvirtualkey65120000virtualkey747-DSA300I45NA
Mine will come in tomorrow and i'll hook it up and post pics/results.

Is that diode getting hot at all when you push it hard? If not I may try a smaller schottkey I already have on hand...

No it doesn't get hot. Barely warmer then room temp. Smaller might be OK. I'll test current when I get my oscilloscope next week
 
okashira, how thick does the two wires need to be that attaches to the diode?

Can this wire can be thinner than the welding lead wire?
 
Offroader said:
okashira, how thick does the two wires need to be that attaches to the diode?

Can this wire can be thinner than the welding lead wire?
Yes, esp if short, 8 even 10 awg would be okay
 
Ok I will delete it,

In general, high school Physics and two year technician programs use Electron Flow.
But three year technologist and university engineering programs use Conventional Current.
You must realize what convention is being used because the rules change. Ex. Right-Hand rules in Conventional Current become Left-Hand rules in Electron Flow.
Many textbooks are available in both formats.


The battery negative terminal the short line, and positive terminal the long line.
From there, you either use conv. or elec. flow.
Hence, Cathode is negative Anode is positive.
So your + is on the negative terminal of EMF, as well on the Cathode negative of diode.
But we are using conventional flow though! :wink:
 
markz said:
Ok I will delete it,

In general, high school Physics and two year technician programs use Electron Flow.
But three year technologist and university engineering programs use Conventional Current.
You must realize what convention is being used because the rules change. Ex. Right-Hand rules in Conventional Current become Left-Hand rules in Electron Flow.
Many textbooks are available in both formats.


The battery negative terminal the short line, and positive terminal the long line.
From there, you either use conv. or elec. flow.
Hence, Cathode is negative Anode is positive.
So your + is on the negative terminal of EMF, as well on the Cathode negative of diode.
But we are using conventional flow though! :wink:

Yes I drew the voltage source symbol backwards :)
 
Ok I just got my fully assembled welder in today. Being a welder of many other materials I did not even think to look at what it took to run this thing.... 900+ amps @ 12VDC seems to be it. Jesus I can weld 1/2" steel plate with 900 amps. In any event, I made the mistake of not checking this out before buying it. But now that it is here...... I am reading about people blowing Fets left and right etc. I have no desire to blow this thing up. I assume everyone is welding using 12VDC for a reason.
I have followed the thread but I am still not sure if I can run the welder "as is" without diode fixes etc.
I have plenty of spare cells to build a pack to supply it. Can some one advise an S/P combination that will give me around 100 welds between charges. My cells are Panasonic CGR18650DA. I have plenty. Is 3S at 9-12.6V the magic # for voltage? What is the voltage tolerance of the welder? Can it take a 4S @ 12-15.6V ? My cells are supposed to be ok at 2C constant and up to 4-5C pulse. The cells are 2400 mAh. So I am thinking in pulse I could use 10 amps per cell as a calculator without kicking the cells ass. If I need 900 amps that is 3S 90P ?? As I don't know voltage tolerances I'm not sure. I have the 270++ cells as spares/ experimenters/ etc. anyhow. I assume it can take high voltages if it can run on a fully charged lead acid 12 volt. I would guess everyone will tell me to just go buy a car battery and have at it.....?????
If this is going to be an ultra pain in my rear I would rather just sell it and do something else......as of now it is un used and in the USA Ohio
 
hagerty1 said:
Ok I just got my fully assembled welder in today. Being a welder of many other materials I did not even think to look at what it took to run this thing.... 900+ amps @ 12VDC seems to be it. Jesus I can weld 1/2" steel plate with 900 amps. In any event, I made the mistake of not checking this out before buying it. But now that it is here...... I am reading about people blowing Fets left and right etc. I have no desire to blow this thing up. I assume everyone is welding using 12VDC for a reason.
I have followed the thread but I am still not sure if I can run the welder "as is" without diode fixes etc.
I have plenty of spare cells to build a pack to supply it. Can some one advise an S/P combination that will give me around 100 welds between charges. My cells are Panasonic CGR18650DA. I have plenty. Is 3S at 9-12.6V the magic # for voltage? What is the voltage tolerance of the welder? Can it take a 4S @ 12-15.6V ? My cells are supposed to be ok at 2C constant and up to 4-5C pulse. The cells are 2400 mAh. So I am thinking in pulse I could use 10 amps per cell as a calculator without kicking the cells ass. If I need 900 amps that is 3S 90P ?? As I don't know voltage tolerances I'm not sure. I have the 270++ cells as spares/ experimenters/ etc. anyhow. I assume it can take high voltages if it can run on a fully charged lead acid 12 volt. I would guess everyone will tell me to just go buy a car battery and have at it.....?????
If this is going to be an ultra pain in my rear I would rather just sell it and do something else......as of now it is un used and in the USA Ohio

Yes, I am sorry, don't want to mis lead people.
The welder works fine as-is. You can weld up to 0.2mm nickel easily with the kit supplied by Riba using a lead acid starter battery
You can even weld 0.3mm nickel, but be mindful that the mosfets stay as cool as possible and don't use too powerful of a battery.

The diode is an upgrade for those of us that want 2000 amps and longer welding leads and want to weld lots and lots of 0.3mm nickel.


Just use a lead acid starter battery. Don't worry about amps, S, P, whatever, just use a battery like Riba recommends and it will be fine. You can prob use the one from your car.

This is the BEST 18650 spot welder you can buy for under $2,000. Period. Your next option is a Sunstone Engineering welder. This one can work just as well.
 
Hi, Hagerty, I must tell you that you wont be lucky with using those cells as welding source unfortunately. It would be best to find a small car starting battery, that would be easiest, and least time consuming, and would work no problems. If you need any other advice please ask! It's simpler than it looks :) And BTW listen to okashira, he's right and has a lot of experience :)
 
Back
Top