DIY MOT spot welder for 18650 cells

[bikegeek]

Hi,
I'm building my first ebike. And I want to build the battery pack. I've ordered the cells and 0.1mm nickel strips and now I'm working on a spot welder made of a microwave transformer. But I have some questions.

What amperage I need for welding 18650 cells? I'm now at 25A powered from 220V.

What is the time the MOT needs to be powered for welding? The time for each weld.

The tips has 3mm space betwen them, it is enough?

Any advice would be appreciated.

 

[Nobuo]

at 220VAC ~20A for 0.1mm ~30A for 0.15mm and ~40A for 0.2mm. For pure nickel.

A single pulse is in the order of <20ms (corrected)

Distance bewteen electrodes for 2mm to 3mm.

Voltage from 3.5V to 5V

 

[parabellum]

:lol: First thing is to fit as thick wire in secondary winding as you can get in. Use wending wire, biggest D that fits, there will be hundreds of Amps on the output.

 

[bikegeek]

In the primari I have 220V and in the secondary I have 1.5V and 25A.

I don't need thick wires for the secondary because for 20-30A is more then enough 2.5mm cross section.

I was thinking of using AT8N timmer for the pulses http://products.autonicsonline.com/Asset/AT8N%20manual.pdf

 

[riba2233]

I'm using 1000-2000 A for welding, so go figure You're not going to weld anything with that setup, maybe alu foil

 

[bikegeek]

I want to weld 0.1-0.15mm pure nickel sheet on 18650 cells.

I think for this is ok, right?

 

[riba2233]

Absolutely not, unfortunately. Just search for other MOT welders on forum and you'll see how a successful MOT welder looks like.

 

[flangefrog]

I agree with the above, there's no way you will weld anything with that. I use three turns of 25mm2 welding cable with the outer insulation removed. I measured 750A but it might be more now I've shortened the cables. My input current is 30A at 230V. More current would be better to allow me to use a shorter pulse. If you're set on using a MOT read the links in my thread: https://endless-sphere.com/forums/viewtopic.php?f=6&t=73036#p1102899

 

[kdog]

Hey bikegeek
Good onya for the DIY! Don't get dejected but it'll take a bit of reworking.
Ditch the lights.
you need fat cables ,as in to jam that secondary coil full of copper. Go for 1.5-2.5 turns. Carry the cables as you did all the way to your electrodes. You need a good current path al the way to the electrode tip. All copper (clean surfaces in joins with big surface area) electrode must be copper.
Output amps= input amps x 240/ output voltsx efficiency of transformer. Typical (good) value might be
20x240/2.5x%70= 1344amps.
So your thin cables aren't going to cut it, plus your copper fill is very poor result in v poor efficiency.
Timing of pulses is critical. Finger on a switch varies waaaay to much. To switch the feed- Simplest way is a solid state relay with a 555timer SSR can run directly off output of 555.
Electrode pressure must be consistent
All these factors determine amount of current for what duration and repeatability. If they are not controllable then you will get arcing/melting which will destroy your battery instantly.
Also check out nobuo's repository for current carrying ability of .1 nickel. It's pretty low. Definitely aim for .15 minimum
A MOT spot welder can be really good and cheap but it'll take a bit of fiddling.
Mine will do upto .3mm nickel and .1mm copper with good reliability and 100%duty cycle.
Happy to answer any qs you have cause I've been working on mine for a while and now it's running sweet here's a few pics to insire you!

Kdog

 

[kdog]

Correct me if I'm wrong but looking at your pic, your MOT is always on and you make the contact to activate the weld?
Your MOT should be off, then you make contact with the electrodes, then apply a pulse of power ( via a 240v switch) to the MOT to make the weld.
I think any other way will fail , (but there might be another way)
Bear in mind the time difference btw a good weld (say 30ms) and blowing a hole in the end of your cell is prob 50-80ms max
Once you have it working like this- remove the shunts ( the thin stack of lams btw primary and secondary coils) to give you a decent increase in power. Might get a bit hot if you take them out now whilst it's always on.
K

 

[bikegeek]

Good morning

You all have good feedback. But I have some questions.

Firs I want you to know that I want to "copy" this type of welder: http://www.aliexpress.com/item/Hand-held-Spot-Welder-Machine-Welding-Laptop-Battery-Button-battery-Battery-Pack/594538510.html?spm=2114.01020208.3.28.cEpeCP&ws_ab_test=201556_10,201527_3_71_72_73_74_75,201560_1 And this welder works with max 40A in the secondary, right? The thing I don't know is the time of the pulse. And for that I plan using this http://www.ebay.com/itm/AC-DC-12V-0-01s-9999h-Range-Digital-Display-Timer-Time-Delay-Relay-w-Socket-/311037462460?hash=item486b4707bc which will give me the pulse from 10ms to 10s.

I see the spot weld as a consume of energy which depends on time and amperage. I see that you use high amperage with small times (1ms to 20 ms) and I want to use low amperage with longer times (20ms to 100 ms). I'm on this way because I can't control the pulse to 1ms. Is my vision wrong?

And because Nobuo gave me this real data that he uses "220VAC ~20A for 0.1mm ~30A for 1.5mm and ~40A for 0.2mm. For pure nickel. The pulses are in the order of 100ms." I think I'm on the right track. Right?

I made some tests with 16mm cross section copper wire and I get 700A but I can't control that amperage. If you can direct me to a circuit board or device that I can buy online and give me that small time of 1-2ms I can use MTO as you suggested.
I received a suggestion using Arduino, but if is involving programming a PIC is not good. I can't do that I know it is a better solution because you can use the welder on more materials. Do you have other recommendation other than Arduino for the pulse control?

I know that the welder is not finished and safe. I have to make a pedal switch, a power switch/fuse, change the secondary wires and a lot more but I wanted you give me feedback now and not after I "finish it"

I'm frustrated because my nickel strips and cells are on the way and I can't test what I think.

 

[flangefrog]

I doubt that welder uses 40A in the secondary. It's probably the current in the primary. Here's a quote from the listing:

When adjust the current to 30A, the fuse may be blown if continuous welding. It is because of the working principle and this will not do harm to the machine. The output voltage is only 3V, it is very safe.

If the secondary current was only 30A, the primary current would be only 0.4A before taking efficiency into account which is hardly likely to trip a breaker. More than likely it's the primary current - same as mine which is 30A but it should be better efficiency so it would have over 1000A in the secondary.

Nobuo was probably just writing which settings he turned the dial to on that welder, not the actual secondary current he was using.

0.15mm x 8mm nickel is rated for something like 5A continuous current. This will only make it warm. You'll need a lot more than 25A to melt it.

Using an Arduino doesn't involve standard PIC programming. Just plug it into a usb port and download some simple code like what I'm using:

void setup() {
  pinMode(13, OUTPUT);
  pinMode(8, INPUT_PULLUP);
}

void loop() {
  while (digitalRead(8)) {
  }
  digitalWrite(13, HIGH);
  delay(40);
  digitalWrite(13, LOW);
  delay(200);
  digitalWrite(13, HIGH);
  delay(80);
  digitalWrite(13, LOW);
  delay(500);
}

The relay you linked to is only rated at 5A. You might have problems with it as the start up current of an inductive device like an MOT will be very high.

 

[bikegeek]

I have to ask a friend that knows more about Arduino to explain me more because it seems is not that complicated. I've seen there are more models of Arduino, I can chose any model?

You have a good point about the amperage. I was thinking is the secondary amperage It seems Aduino all the way.

 

[parabellum]

Good morning

You all have good feedback. But I have some questions.

Firs I want you to know that I want to "copy" this type of welder: http://www.aliexpress.com/item/Hand-held-Spot-Welder-Machine-Welding-Laptop-Battery-Button-battery-Battery-Pack/594538510.html?spm=2114.01020208.3.28.cEpeCP&ws_ab_test=201556_10,201527_3_71_72_73_74_75,201560_1 And this welder works with max 40A in the secondary, right? The thing I don't know is the time of the pulse. And for that I plan using this http://www.ebay.com/itm/AC-DC-12V-0-01s-9999h-Range-Digital-Display-Timer-Time-Delay-Relay-w-Socket-/311037462460?hash=item486b4707bc which will give me the pulse from 10ms to 10s.

I see the spot weld as a consume of energy which depends on time and amperage. I see that you use high amperage with small times (1ms to 20 ms) and I want to use low amperage with longer times (20ms to 100 ms). I'm on this way because I can't control the pulse to 1ms. Is my vision wrong?

And because Nobuo gave me this real data that he uses "220VAC ~20A for 0.1mm ~30A for 1.5mm and ~40A for 0.2mm. For pure nickel. The pulses are in the order of 100ms." I think I'm on the right track. Right?

I made some tests with 16mm cross section copper wire and I get 700A but I can't control that amperage. If you can direct me to a circuit board or device that I can buy online and give me that small time of 1-2ms I can use MTO as you suggested.
I received a suggestion using Arduino, but if is involving programming a PIC is not good. I can't do that I know it is a better solution because you can use the welder on more materials. Do you have other recommendation other than Arduino for the pulse control?

I know that the welder is not finished and safe. I have to make a pedal switch, a power switch/fuse, change the secondary wires and a lot more but I wanted you give me feedback now and not after I "finish it"

I'm frustrated because my nickel strips and cells are on the way and I can't test what I think.

1st link is primary "probably" :lol:
2nd link is "just forget about"
I am looking an this https://endless-sphere.com/forums/viewtopic.php?f=31&t=68865
Nobuo is referring to primary winding, which makes at 30A primary 660A secondary at 10V

 

[bikegeek]

That battery welder is a piece of art but I have to make something that works and is not that expensive

 

[kdog]

Rough cost $70aud
SSR $15
Copper cable $25
Two heavy connectors $22
Microwave $0
Timer circuit $8
Wood, scrap copper n stuff $0

Coulda done it a bit cheaper but I was learning as I went.
K

 

[kdog]

Lower amps say 3-500 with longer pulse duration will result in severe cell heating and no weld and will destroy your cells if not immediately, then a few months down the track they'll start showing signs of heat damage.
It's critical to have high amps like ~1000+ with as short a pulse as necessary to get the weld completed before significant heating occurs.
Primarily work on timing control and then electrode set up- the rest is easy ( just my opinion)
K

 

[bikegeek]

Kdog, what type is your timer?

For now I found this on ebay:
- arduino: http://www.ebay.com/itm/NEW-UNO-R3-ATmega328P-CH340-Mini-USB-Board-for-Compatible-Arduino-/311155383820?hash=item48724e5e0c
- http://www.ebay.com/itm/24V-380V-SSR-40DA-40A-3-32V-DC-AC-Solid-State-Relay-Module-for-Arduino-/221607872157?hash=item3398dbc69d

It could work?

 

[kdog]

That SSR looks good- check out it specs before you buy but it looks good at first glance (40amp, 380v). I know nothing about arduino but it will almost certainly run off the out put cause they only need 20ma or something to activate. I prob wouldn't use 3 volts but more like 9-12
I used a 555 IC in a monostable circuit that gives a single adjustable timed pulse regardless of input range on mine 0-300ms ( anywhere near 300ms is instant destruction of test price) very simple circuit if you've done any electronic stuff.
I can dig up the circuit diagram if you want.
K

 

[flangefrog]

For now I found this on ebay:
- arduino: http://www.ebay.com/itm/NEW-UNO-R3-ATmega328P-CH340-Mini-USB-Board-for-Compatible-Arduino-/311155383820?hash=item48724e5e0c
- http://www.ebay.com/itm/24V-380V-SSR-40DA-40A-3-32V-DC-AC-Solid-State-Relay-Module-for-Arduino-/221607872157?hash=item3398dbc69d

It could work?

This is pretty much what I use. I think I'm actually using the 25A rated SSR but it's holding up fine.

 

[kdog]

Flange frog
Did I see shunts? Take 'em out for a noticeable increase in power- but you might need the 40amp SSR!
K

 

[flangefrog]

Flange frog
Did I see shunts? Take 'em out for a noticeable increase in power- but you might need the 40amp SSR!
K

Yes, I wasn't sure if it would take too much power after removing them so I left them in. I'll try it without them but first I need to replace my nickel plated steel electrode clamps which burn my fingers after several welds. The voltage in my mobile home probably goes below 100V when I'm welding (nominal 230V) so I'll have to move nearer to the source.

 

[bikegeek]

I think I will test both ways, with Arduino and 555 timer (I will test the 555 timer first because it will take some time to receive the Arduino board).

Kdog

I found this about the 555 timer: http://www.electronics-tutorials.ws/waveforms/555_timer.html

It is OK to use Monostable 555 Timer circuit? I will use a potentiometer for setting the time. What do you recommend? And how to you know that time you've set?

Flangefrog

Can you explain how the code you send me works? How can I change the pulse time? And do you gave a schematics how to connect all together?

 

[kdog]

Yeh monostable (not flipflop)is the go. This is the circuit I used from:http://www.bowdenshobbycircuits.info/page9.htm scroll down to monostable one shot. Circuit on the right is best. Good info in the text.
I changed the timing cap (22uf) to a .3uf and put a 500kpot in series with a 5k resistor to replace the 47k resistor between pin7 and positive rail. This gave me shorter pulse times and fine control.
A friend had an oscilloscope so that helped determine pulse times.
Forget the relay and diode from output(pin 3). Just plug pin three straight into your SSR.
Get some old batteries and test it thoroughly before building your pack.

 

[flangefrog]

The code within the setup function is run once after powering up the Arduino. First I set pin 13 to an output pin. Pin 13 has an on board LED so you can test without the SSR connected. The SSR is connected between pin 13 and ground. Pin 8 is set to an input. Pull-up means that an internal resistor is pulling the input high (making a high resistance connection to 5V) when it is unconnected. The foot switch is NO (normally open) and is connected between ground and pin 8. When the switch is activated it will pull the input low by connecting it to ground as it's a much lower resistance than the pull-up resistor.

void setup() {
  pinMode(13, OUTPUT);
  pinMode(8, INPUT_PULLUP);
}

The code within the loop function as you may have guessed loops continously after the setup function has completed.

void loop() {

This snippet loops while the result from reading pin 8 is true (or high). This means that it will wait until the foot switch is pressed before continuing.

  while (digitalRead(8)) {
  }

The output is turned on and after delaying for 40ms it is turned off. This first pulse is used to soften the nickel so it lies flat on the battery ready for the actual spot weld.

  digitalWrite(13, HIGH);
  delay(40);
  digitalWrite(13, LOW);

There is a small delay of at least 100ms before making the spot weld

  delay(200);

The output is turned on and after 80ms it is turned off. This is the pulse used for the spot weld.

  digitalWrite(13, HIGH);
  delay(80);
  digitalWrite(13, LOW);

The delay of 500ms before looping back to the start of the function ensures you don't accidently make two welds by double tapping the foot switch. It essentially debounces the switch.

  delay(500);
}

Start with much shorter pulses than I'm using as I only need it this long because of the low current output of my welder.