anywhere can find a simple IGBT or MOSFET switch circuit? I mean just simply turn on / turn off the large current, and the on/off time adjustable?

Why? Do you understand how fast a capacitor discharges compared to how fast a mosfet of IGBT can switch? The SCR is the correct thing to discharge the capacitor(s). You adjust the amperage/current into the weld by the number or value of the capacitors you use for the weld, not the time that the switch is on. Max I think already told you this.

Or you could use a GTO, gate turn off SCR if you really believe that your way is best and what commercial built machines use. Do some more reading on this, I did about a year or so ago. But decided the welds were only worth while for jewelry makers/artists that's who they are marketed to.

 

Why? Do you understand how fast a capacitor discharges compared to how fast a mosfet of IGBT can switch? The SCR is the correct thing to discharge the capacitor(s). You adjust the amperage/current into the weld by the number or value of the capacitors you use for the weld, not the time that the switch is on. Max I think already told you this.

Or you could use a GTO, gate turn off SCR if you really believe that your way is best and what commercial built machines use. Do some more reading on this, I did about a year or so ago. But decided the welds were only worth while for jewelry makers/artists that's who they are marketed to.

Thanks.
I didn't go that detail yet, I don't have any SCR in hand now, and I do have quite many IGBT and MOSFET handy.
Sure, it's kind if I can have SCR circuit.

 

SB is correct about the fast current rise time, which is limited by the inductance of the current loop from the capacitors, through the switch, the weld, and back. And that is a quite small inductance so it does not limit the rise time very much. The problem with any other switch device is the power dissipated in the switch device before it reaches full saturated conduction. That means that it can burn open the very first time. And that time is less obvious from the data sheet. The use of a suitable FET to trigger the gate of the SCR seems like a good idea, because it can ease the burden on the trigger circuit.

 

SB is correct about the fast current rise time, which is limited by the inductance of the current loop from the capacitors, through the switch, the weld, and back. And that is a quite small inductance so it does not limit the rise time very much. The problem with any other switch device is the power dissipated in the switch device before it reaches full saturated conduction. That means that it can burn open the very first time. And that time is less obvious from the data sheet. The use of a suitable FET to trigger the gate of the SCR seems like a good idea, because it can ease the burden on the trigger circuit.

Thanks.
I am reading the datasheet of '70TPS_VishaySiliconix', I didn't find and parameters about time yet.
The IGBT - 'FGH60N60' - has a 22 ns 'turn-on delay time ' and 134 ns 'turn-off delay time', that I think will be OK in my case?

 

I did look at the data sheet and I see that the lead current limit is 75 amps. and looking at the picture of the device I do not think that it would handle 75 amps for very many milliseconds. And there is no way that package can withstand 200 amps for a length of time that would be useful in welding. Pulsing a diode laser with femto-second pule, perhaps, but not melting steel. It could work for the welder putting tab on battery cells, maybe. Look at some picture of pulse welder oof the range that you are considering and see the size of those witching devices. That will help clarify the realm to be searching.

 

I did look at the data sheet and I see that the lead current limit is 75 amps. and looking at the picture of the device I do not think that it would handle 75 amps for very many milliseconds. And there i no way that package can withstand 200 amps for a length of time that would be useful in welding.

Yeahbut...

Hi,
The cold welding machine, its principle is to use a charging capacitor and discharge in an ultra-short time of 10-6 (10^-6) to 10-5 (10^-5) seconds

All these pulses will be .01 to .001 mS supposedly. I am holding my breath to see it.

 

Really, those pulse lengths seem more like laser welding pulse lengths. There are some physical limits to what can be done with capacitor discharge welding. Reality is seldom negotiable. And the internal resistance of real capacitors is never zero. That has a serious effect on RC time constants.

 

I did look at the data sheet and I see that the lead current limit is 75 amps. and looking at the picture of the device I do not think that it would handle 75 amps for very many milliseconds. And there is no way that package can withstand 200 amps for a length of time that would be useful in welding. Pulsing a diode laser with femto-second pule, perhaps, but not melting steel. It could work for the welder putting tab on battery cells, maybe. Look at some picture of pulse welder oof the range that you are considering and see the size of those witching devices. That will help clarify the realm to be searching.

Thanks.
I'll dig into more later.
In face most of the cold welding machine on market used MOSFET, which I checked few of, the IRF3205 has turn on delay 14 ns, and tuen off delay 50 ns, the other one W9NK90Z has almost same.

 

Yeahbut...

All these pulses will be .01 to .001 mS supposedly. I am holding my breath to see it.

Thanks.
May I know where the " .01 to .001 mS supposedly" come from? It's a huge amount relative to the talking I guess?

 

Really, those pulse lengths seem more like laser welding pulse lengths. There are some physical limits to what can be done with capacitor discharge welding. Reality is seldom negotiable. And the internal resistance of real capacitors is never zero. That has a serious effect on RC time constants.

Thanks.
In fact, lot of people call the cold welder as 'imitate laser welder' even they are totally different stuff.
I agree with you that the capacitor discharge may not be able to even us magnitude in that mount current.
The ' discharge in an ultra-short time of 10-6 to 10-5 seconds' in the main topic is just a theoretical statement I guess?

 

In face most of the cold welding machine on market used MOSFET

I would like to see the links to those that do. Most I saw use a SCR.

 

I would like to see the links to those that do. Most I saw use a SCR.

Thanks.
I didn't find a complete article about 'cold welding machine', just a few words here and there.
Anyway, as long as I got suitable driver circuit for anyone of SCR/MOSFET/IGBT, I'll start to test them, my SCR order is coming.

 

Thanks.
May I know where the " .01 to .001 mS supposedly" come from?

It came from here:

Hi,
The cold welding machine, its principle is to use a charging capacitor and discharge in an ultra-short time of 10-6 (10^-6) to 10-5 (10^-5) seconds

10^-6 seconds = 0.000001 seconds. Move the decimal 3 places to the right and you get .001mS

10^-5 seconds = 0.00001 seconds. Move the decimal 3 places to the right and you get .01mS

It's a huge amount relative to the talking I guess?

Unless I have a tragic misunderstanding of scientific notation or the metric system or both, it's exactly the amount relative to the talking. And it's not huge, it's infinitesimal relative to the actual pulse width of actual welders. I mean I can count 3-5 video frames worth of ON time per pulse in the video you linked. Assuming it's shot at 24fps like most videos, that's 125 to 200 mS per pulse, which is 20,000 - 125,000 times huger than the amount of the talking.

 

It looks quite accelerated.

Are you suggesting the video is played back at fast speed? If that's what you're saying, and if you're correct, then it would mean that the ON pulse is well above 125-200mS. Maybe half a second.

This is what I was talking about when I made the suggestion to just use reasonable numbers and make something that works, then make outrageous and obviously false claims about its performance. That's how China do.

 

It does take a finite bit of time for that delivered power to generate some heat and for the weld metal to flow. Quite rapid but not instant.

 

Look at the finger movement in this sequence.

Yeah it's pretty obvious now that you pointed it out.

So @LAOADAM are you still committed to microsecond pulse widths or are we permitted to discuss real world values that actually work?

 

If this is going to continue we should start to call it by it's real name/term. While Laoadam is calling it "cold welding" that is a marketing name. The real name is "pulse TIG welding". And he will get more information if that is the name he searches for, instead of cold welding.

Here is a link to a company that is the supposed gold standard(by many that use them) of pulse tig machines -
https://sunstonewelders.com/orion-jewelry-welder. There are a few others making them but people that are serious usually end up with one of their machines as their work gets more intense and start to make money at it. Click one the resources button at the top of the page to get more info on the processes. But be prepared when looking at the prices of the machines unless the price has gone down since I looked.

There are less expensive ones if you just want to look into it to see if it will do what you want. These are just a few links to the less expensive machines, a Google for "jewelry spot welder" gets more. One of these in my opinion would, at the price be much better and do more than trying to DIY one, since there is so little information on doing it that way. And you get all of the welding hardware with most of them. And don't forget you will need an argon tank which is fairly expensive to buy.

https://www.gesswein.com/p-13745-pu...kv66_UPRUTSuS87jWCQy0J4gdWvPsgzBoCP5QQAvD_BwE

https://www.ebay.com/itm/1532990483...yLx74PLlbfm80_lH191Q6LmIPzBickkxoCr6MQAvD_BwE

https://www.amazon.com/Carejoy-Spar...ocphy=9015438&hvtargid=pla-944548195365&psc=1

 

If this is going to continue we should start to call it by it's real name/term. While Laoadam is calling it "cold welding" that is a marketing name. The real name is "pulse TIG welding". And he will get more information if that is the name he searches for, instead of cold welding.

Here is a link to a company that is the supposed gold standard(by many that use them) of pulse tig machines -
https://sunstonewelders.com/orion-jewelry-welder. There are a few others making them but people that are serious usually end up with one of their machines as their work gets more intense and start to make money at it. Click one the resources button at the top of the page to get more info on the processes. But be prepared when looking at the prices of the machines unless the price has gone down since I looked.

There are less expensive ones if you just want to look into it to see if it will do what you want. These are just a few links to the less expensive machines, a Google for "jewelry spot welder" gets more. One of these in my opinion would, at the price be much better and do more than trying to DIY one, since there is so little information on doing it that way. And you get all of the welding hardware with most of them. And don't forget you will need an argon tank which is fairly expensive to buy.

https://www.gesswein.com/p-13745-pu...kv66_UPRUTSuS87jWCQy0J4gdWvPsgzBoCP5QQAvD_BwE

https://www.ebay.com/itm/1532990483...yLx74PLlbfm80_lH191Q6LmIPzBickkxoCr6MQAvD_BwE

https://www.amazon.com/Carejoy-Spar...ocphy=9015438&hvtargid=pla-944548195365&psc=1

OK, and thanks, shortbus. for pointing out that it is a TIG machine. I was thinking about the impulse spot welders that I have seen used for fairly small stuff, primarily welding splices in wire harnesses and welding tabs on batteries. None of the big stuff. Big stuff is a different game completely.

 

The impulse spot welding machines I worked on, also in the automotive industry were Seam welders, they were identical to a large spot welder, instead of single electrode they had 12" dia revolving copper wheels, top and bottom.
The two metals to 'seam' were drawn through the wheels and a weld was continuously pulsed. Sealing the tank etc.

 

It came from here:


10^-6 seconds = 0.000001 seconds. Move the decimal 3 places to the right and you get .001mS

10^-5 seconds = 0.00001 seconds. Move the decimal 3 places to the right and you get .01mS



Unless I have a tragic misunderstanding of scientific notation or the metric system or both, it's exactly the amount relative to the talking. And it's not huge, it's infinitesimal relative to the actual pulse width of actual welders. I mean I can count 3-5 video frames worth of ON time per pulse in the video you linked. Assuming it's shot at 24fps like most videos, that's 125 to 200 mS per pulse, which is 20,000 - 125,000 times huger than the amount of the talking.

Thank you for the detail.
Sorry about my asking cause of I misunderstand the unit...
I agree with you there is a big difference between theory and practice, in fact, I don't think it is necessary to focus on the (10^-6) level, and even it may not work actually. cause of the arc is a such a complex process.