# Capacitor discharge welder

Status
Not open for further replies.

#### Jacounet

##### New Member
Hello.

I'm in France , I try untill 2016 to build a Welder with discharge of capacitor , the target ,to solder 2 sheets of aluminium ( 2x2 mm)
I have used 20 xIXYS 100A/100V in // , I used 0.2F under 100Volts ( 900 Joules)
So aluminium have a very low résistance if comparated to iron ( X4 lower) .
I have soon burned 10 MOSFET ..upper than 80 Volts power supply...
I try last week with 32 X IRFP4468 ...in // without success.(32x195= 6240 A pic , Under 1 ms of , "on" conduction ).
That work from 20 to 80 Volts without problem with 20 IXYS ...but without soldering , ...and from 5 to 20 Volts with IRF but without soldering too.

So I remake my calculations , 2 sheets of 2x2 mm of aluminium have a résistor of "800µOhms" Under pin copper of 2 mm diameter Under contact .

So Under 100Volts pic is 100/0.0008 Ohm = 125 000 Ampères pic for 6240 A possible with 32 MOSfet.

For having 1 k.Joule ( energy for solder 2 sheets of aluminium 2x2mm) , I want to uncrease my capacitors to 1000.Farads with 2.7 Volts super capacitors , and I want to put,... 50x MOSFET IRF3711 in //.(50x110=5500 Ampères pic max ).
Curent Under 2.7 Volts for solder to sheets 2x2 mm aluminium Under contact surface = 2.7/0.008 =2750 A making 55 A by MOSFET .

This calculation and this "third project" seam to be good in the real live .

Oh my Mosfet are équilibrating with a 1 milli Ohms (DIY resistor) serial with "source" of each MOSFET.

Is somebody follow my brain storming ...

Is somebody have tried this sort of experience ?...
Jac
PS.
If you add the rDS"on" ...5mOhms (by MOSFET's r DS"on" /32=0.156 milli Ohms)... +3 milli Ohms of "ESR" capacitor 0.2 F , we have till 100/0.003956=25 278 Ampères pic , and only 0.8/3.156 = 25.3% of the energy for the soldering ...intensity too huge and soldering impossible .

Last edited:

#### shortbus=

##### Well-Known Member
I try untill 2016 to build a Welder with discharge of capacitor , the target ,to solder 2 sheets of aluminium ( 2x2 mm)
I don't think you will ever be able to spot weld, which is what you seem to be talking about, not soldering, aluminum. Aluminum has a very thin layer of oxide that needs to be removed before it can be welded, and I have never in all the years I've worked with metal, seen aluminum spot welded. Even when TIG welding it is done with AC to remove the oxide layer so welding can take place.

#### JimB

##### Super Moderator
Note:
Jacounet is in France.
In the French language the word "souder" can be translated to both of the English words "solder" and "weld".

JimB

#### Jacounet

##### New Member
Hello.

I have forgoten to present/introduce who I am.
I'm an retired technician of électronics equipments .I'm 68 years old .
I hope my English/american is correct, or understanding .

Yes I well want to weld by fusion 2 sheets of 2 mm aluminium.
I have a DIY spot welder ( 5 kWatts consumption , 70 % efficiency) build with 2 mecanical assembled micowaves transformer ( 12 kgs), certainly "m.t." in English , "m.o" in french .
It is possible to weld/solder 2 sheets of aluminium ( oxide of aluminium eliminate ) , with "spot welder" with 2 points of copper of 10 mm or 20 mm diameter( 5 mm at extremity)...if you put your 2 sheets of aluminium( Under) and beetween 2 sheets of iron .
So the solder/welding is not perfect because aluminium is very liquid/fluid , it fuse and quasilly explose if the thermic énergy is too hard/quick.
One factory , where ? good question ...in USA,Ausralia or Canada ...I don't remember where , I sick is name , make A welding machine , big monsters , of 10 to 100 kW , with frequency generator of 20 kHz of 10 to 100 kWatts so , and drive ( with SCR ...nothing is said ) the sinus wawes of 50 µs accross a big transformator 20 kHz ( perhap's with ferrite) , this waves are choosen perhaps 2 waves "on"or or 4,8,16 , and 1 or 2 ms seconds "off"...for making a goof fusion without "splash" explosion .

But they say they have abandon the idea of a welder with capacitors , because of the impossibility at this moment to contrôle the duration of the discharge .

With a SCR 5000A/1600 V, the best on the market( in China of course ) , to discharge a capacitor , seam to me impossible because of the dI/dt of silicium which is only of 150/200 A by µs (200A/µs) for the best...even if the driving power is present ( hundreds kW) , ...but ...there is a but , only with sinus wawes of 50 or 60 Hz .
With a capacitor discharged , we have a crazzy "pic" of current impossible to locate with the most performant oscilloscope of laboratory .
We are waiting for SiC with a 4000A/µs anonced , but in qualification since 2015 or 2016 . ...nothing produce and sailing on the market .

So with MOSFET I have the same problem of the dI/dT , wich is given for the IXYS IXFK100N10 for 100A/µs and 100A max Under 100 Volts .
With 100 Mosfet in // , even if not easy , we can rise 10 000 Amperes .
With the command and an appropriate gate resistor we can turn this problem , with a signal of 1 or 2 µs "0" to "1" .( 0 to 15 Volts ).

That's all for today .
By.

Jac .

Last edited:

#### shortbus=

##### Well-Known Member
In the French language the word "souder" can be translated to both of the English words "solder" and "weld".
Didn't know that, just assumed when it was spelled solder that was what was meant.

@Jacounet, I did some more reading and found out aluminum can be spotwelded, so I learned two things today. But I've never been able to do it even with the welders we had at work. But don't know if all of the needs can be met as a DIY type thing. Some of the parameters that need to be met -
Electrode force:
Weld time:
Welding current: 3 to 5 kN
2 to 5 cycles (40-100ms)
22 to 28kA
From, https://www.twi-global.com/technica...w-do-i-resistance-spot-weld-aluminium-alloys/

#### Jacounet

##### New Member
Hello shortbus.
I have found the name of the society/factory making the spotwelder for all metals as aluminium, gold,... : it's Matushek .

For the parameters you give that seam correct for me , I have read the same informations .
For 22 kA the need would be of 220 MOSFET in //...impossible for me .
(That's begin a "gaz factory", in french It's a joke translated by "
une usine à gaz"...an impossible work...have you the same expression .)

But I can play on a longer timing of the soldering with 50 MOSFET in // and 2000 to 4000 Ampères , instead of 2 to 5 cycles passing to 8 to 20 cycles...for have the same energy.
I calculate 130 Joules by cycle of 40 ms ...and technical/mathematic caculation give 1 k Joules aproximatively for soldering 2x2mm of aluminium Under a cercle of 1 mm ==> 0.785 mm square ...so 10 cycle of 40 ms .
We are in the same play as I ( and you) mentionned upper .

I think I have the contrôle , (but perhap's not ), of the driver ...able to drive hight current ( 5 to 20 A ) because of 50 gate in // ...aproximatively 150 nF ( 1 gate= 2980 pF ==> data sheet IR ) , and with a slope of 2 µs .
I simulate a scheme Under Spice with "ISIS" software . ..could be OK for me .

A big problem : the resistor of the 2 sheets of aluminium 2 mm, I aim by calculation at 200 to 800 µOhms Under a surface of 1 mm/circle ...(0.78 mm square).
under a so thin surface I suppose a bad soldering in final process...

All this is to try .

A+.

Jac

Last edited:

#### Pommie

##### Well-Known Member
Are you talking about sheet that is 2mm thick?

Mike.

#### Jacounet

##### New Member
Hello Pommie.

Yes exactly.
But a sheet of 2 mm thik is twice more resistive than a sheet of 1 mm , and the sumit of curent is twice lower , even if that need more energy ( longer time of soldering with a same intensity )...
That's why I'm going on sheets of 2x2 mm.

Jac .

#### Pommie

##### Well-Known Member
Yes, higher resistance but the ability to dissipate heat is much much more.

Mike.

#### Diver300

##### Well-Known Member
I used to use a capacitor discharge projection welder. It was used to weld the tops on DIL-14 crystal oscillators. Example here.

The welding had to be very quick in order to reduce the heat loss, as the weld was the whole perimeter of the oscillator, all welded simultaneously. To do that the welder had 2500 J of stored energy, at about 2 kV, discharged through a thyristor, into a welding transformer. The welding transformer was very close to the weld, had just one output turn, and was connected with around 500 mm2 of copper.

The power of the weld was controlled by adjusting the initial voltage of the capacitors before the weld happened.

I estimated that the peak welding current was about 100 kA.

I think that trying to control the weld current directly is very difficult, and using a transformer might be a lot easier. That technique was also used where we had a 5 V, 200 A heating load. It was controlled with a 0 - 240 V variac, followed by a 230 - 5 V fixed ratio transformer.

#### Jacounet

##### New Member
Hello Diver 300.

For me for soldering aluminium there is 2 possibility ( perhap's it's not wright ) :
- by discharge of capacitors under 2 to 20 pulses ( perhap's more) with MOSFET , and you are obliged to survey the dI/dt not upper than 200 A/µs.
- by a sinus wawe ( 2 to 20 so ) from 50Hz to 20 kHz into a transformer(230/5 V) controlled by VCR or Variac with an acceptable dI/dt less than 200 A/µs .

I don't believe into a discharge of capacitor across one transformer 230V/5V because of the " -dphi/dt" ( more than "- 10 kV" if 200/300 Volts power supply )make by the brake into the inductor of the primary which can distroyed the VCR, Variac ( in french thyristor , triac) , and the saturation of the transformer quasily obliged at the first discharge , and the impossibility to contrôle the time of discharge .
Or find/search the good diode able to dissipate this energy without problem for electronic power command ...and put a significant snubber .( R and C)
Or under 5 Volt 500A ( for 2500 j if one second welding)...or Under 5 Volts Under 200A with a longer time welding , across an special unsaturated transformer ,and ... suveying an lower 200A/µs level of energy into the silicium waffer .
It's not easy this building .

I have no schemes real or princep ,of this machines ...certainly under patent .

I wait and see , and made my tries , began in 2016 .

By.

Jac .

Last edited:

#### Diver300

##### Well-Known Member
What I wrote may have been confusing.

The variac and 230 / 5V transformer and variac was for a heating application. That was not for welding.

The welding machine had a single discharge from capacitors to the welding transformer. It wasn't a sine wave.

#### Jacounet

##### New Member
Yes, higher resistance but the ability to dissipate heat is much much more.
Mike.
Hello.

Yes of course a bigger energy dissipate Under 2x2 mm of soldering than Under 2x1 mm , but for this exercice/project , I am blocked on 50X100 A= 5000A "pic" of current .
That want to say with 2.7 V with 800 µ Ohm of aluminium( 2x2 mm with 0.78 mm sqare surface électrodes) ===>2.7/0.008=3375 A.
If resistor lower than 800 µ Ohms that's mean upper than 3375 A pic and blocked at 5000A max , limited at 2.7/5000 = 540 µ Ohm charge .
In this case we have so the ESR ( 0.3/0.4 milli Ohms for C= 1000F ) , and the rDS"on , = 0.1 milli Ohms for 50 MOSFET IRF3711 in //.

Have a good day .

Jac.

Last edited:

#### Jacounet

##### New Member
The welding machine had a single discharge from capacitors to the welding transformer. It wasn't a sine wave.
Hello.
The variac if I understand well is for an other application.

So the transformer of your machine is a special one ...certainly with ferrite ( a big one) ...and in this application you are not on aluminium but on steel resistorx4 more ...and the contrôle of the energy is only made with the contrôle of the voltage of C ...not with the duration of burst of pulse of energy .
So the problem for me for soldering aluminium is always present .

By.

Jac .

#### Diver300

##### Well-Known Member
Yes, the welding transformer was very special. The output was one turn of 500 mm2 of copper bar.

#### Jacounet

##### New Member
Hello .
Yes one turn of 500 mm2 can pass 7500 ampères without problem and 4 or 5 more during a short time T<1 second ..
My DIY spot welder with transformer ( 12 kgs) 50Hz have one turn of copper bar 10.5mmx51 mm...and can weld 2 pièces of steel 2x6mm thik .
By.
Jac

#### dr pepper

##### Well-Known Member
Shortbus, as far as I remember the aircraft industry spot welds ally, however the type of alloy and the prep is critical, its not like snotting a couple of bits of steel together with some 6013's.

#### Jacounet

##### New Member
Hello.

I am on a forum in France which is named "usinages.com" , and we have done ,10 of our members, a DIY spotwelders with microwaves transformers .
The power consumption was 1500 to 2000 Watts for a simple "m.t." and 4500 to 6000 Watts for a mecanical doubled "m.t.".
We have all solder sheets of 2x1mm minimum with simple "mt" , and 2x6 mm max with double mecanical "mt" .
And we have all tried the method 2 sheets of aluminium between to sheets of steel .
All of us have fuse the aluminium without problem , but many of us with a hole in the 2 sheets , and for the others a very bad welding .

So I think the machine with 20 kHz Under a power of 10 to 100 k Watts , of the factory "Matuschek " , are near 20 to 30 kAmpères ...as we have said upper in this subject .
The last ISO Norm for aluminium soldering , indicate what sort of solders are correct an what kind of defects are acceptable or not .
And it's not free of charge more than 1 xhundred dollards or Euro ...I don't buy it .

Matuschek turn the problem of the dI/dt for their VCR , but they added an other problem , build a monster of 10/100kW oscillator .
I think that is this kind of machine that make soldering in aeronautics factory in the world since tenths years .

All that don't resolve my problem of my DIY spotwelder with discharge of capacitors and MOSFET .
I have bought in China 50x IRF3711 and 2 super capacitors of 500Farads,2.7 V ...for my tries ...a cost of 25 € only .
It's enought for a third try .

Have a good day.

Jac.

Last edited:

#### Jacounet

##### New Member
Hello.

If sombody interested by a DIY capacitor discharge welder , I can give you the schèmes ( all are very simple ):
- the driver that command 32 MOSFET IRFP4468 ( I max 50 A possible ,t "on" et "off" setting at 2 µs because of the dI/dt Mosfet capability )
- the MOSFET in // ( easy on scheme , not easy to mount )
-the power supply 100 V/5 A , 500 Watts with 2 SCR (50A/400 V) ...limited on my prototype ,only by the transformer .

By.

Jac .