SPOT WELDER WITH PIC16F88 tested - YOUTUBE video

[kalpakchiev]

In this case the SOURCE of weld FET will receive GND from DRAIN of charge FET that's why I think that welding current will have to pass trough both the transistors. Am I right?

[borberk]

No

Posted on: April 12, 2010, 11:46:29 11:46 - Automerged
You solved the problem without knowing it...

[kalpakchiev]

I tested it in PROTEUS but I am still not confident that the problem is solved.

[borberk]

I used Microcap to show start and end currents during weld time. At end time it is supposed that C is totally discharged.

[kalpakchiev]

So it appears that this is the solution. I'll give it a try. For charging FET I am going to use  2xIRL1004 and for weld 5xIRL1004.

[kalpakchiev]

What about this schematic as final decision ?

[borberk]

What circuit is connected left side to pins 1 and 2 of CON4 .

[kalpakchiev]

Pin.1 - charge control from PIC16F88
Pin.2 - weld control from PIC16F88
Pin.3 - analog input of PIC16F88
Pin.4 - GND

[borberk]

I know it and expect that you have 15V voltage swing on them.

[kalpakchiev]

No, 5V when corresponding port of PIC16F88 is HIGH and 0V when LOW

CHARGE MODE                                 PIN.1=5V PIN.2=0V
READY(capacitor is charged) MODE     PIN.0=5V PIN.2=0V
WELD MODE                                    PIN.1=5V PIN.2=1V

[solutions]

You are right but the capacitor I use in these project is rated 20VDC/24VDC Peak so 20V is highest limit. When this project is finished I can make some tests with these capacitors. It will be fun .

Nice job, buddy. Very nice.

A suggestion for V2.0 - two caps in series.  This will half the capacitance and double the voltage.

Here's my evil plan....

E (energy) is C*V*V/2

If you half the cap value and double the voltage by placing two caps in series, you DOUBLE the energy that you can dump into the weld....

Bonus is punchthrough of crud on the metal, as mentioned by a prior poster, due to higher voltage.

[kalpakchiev]

Nice job, buddy. Very nice.

A suggestion for V2.0 - two caps in series.  This will half the capacitance and double the voltage.

Here's my evil plan....

E (energy) is C*V*V/2

If you half the cap value and double the voltage by placing two caps in series, you DOUBLE the energy that you can dump into the weld....

Bonus is punchthrough of crud on the metal, as mentioned by a prior poster, due to higher voltage.

That's a very interesting idea. I want to clean up schematic for charge and discharge and when it is ready I certainly WILL give it a try. IRL1004 is rated up to 40V.

[jzaghal]

Hi,

Do not forget to add a resistor across each of the caps in series. (say 1K ohm).

Cheers.

[kalpakchiev]

Hi,

Do not forget to add a resistor across each of the caps in series. (say 1K ohm).

Cheers.

You mean these capacitors I suppose. I usually use 2.2k.

[borberk]

What about a "ohmmeter" which first detects that electrodes are well pressed on object and then enables firing.

[kalpakchiev]

It is cool  but I prefer to control the process myself. What about last power schematic? What is Your opinion - is there something missing or wrong? I am going to make the PCB tonight.

[borberk]

It looks OK. It is better that resistors in C voltage divider are smaller than 1k. Actually a C leakage currents difference at full voltage determines the resistance. Idivider=10*Ileak.

[kalpakchiev]

Tonight I'll start the PCB. Thanks for the verification. About C voltage divider - I will have it in mind.

[sadman]

hi

1F audio cap hard to find in my country if found price is two much can i used alternative such 47000uf cap in parallal to achive value that required

regards
sadman

[kalpakchiev]

Yes, you can.

[Ichan]

Do you solder over all the power track on the pcb? I think standard copper foil alone is not enough to carry the peak current, with the software below 10mm width of standard copper foil can only carry 7.7 Ampere current.

-ichan

[Ichan]

http://saturnpcb.com/pcb_toolkit.htm

It's Free 

-ichan

[kalpakchiev]

I also had some embarrassment about thickness of the power tracks and I decided to flood them with tin. Many thanks to Ichan.  This software will be very helpful to check the tracks. Best regards

[oldvan]

I prefer most of the older version of the circuit, it seems inherently safer and more correct.  

As far as I can see, a common ground for the charging supply, the - weld electrode and the cap
is the best way to go.

With this version neither of your electrodes is at GND potential and one is always live to your + supply
through 4 Ohms of resistance.  This design risks an avoidable fire and shock hazard.

When not charging, your voltage measurements measure your charging supply voltage because the - of
your weld capacitor is not at GND potential, but left floating.  When charging they will be somewhat off
because of the voltage drop across your charging transistors.

I'd also add a bleeder capacitor resistor to the circuit so the weld capacitor isn't left charged for days and surprises you by erroneously welding a screwdriver, ring, etc.

Edited to correct a brain fart.

[kalpakchiev]

And all those negatives that Oldvan talks about are result of hunting for LOW SIDE to improve the circuit.  
Now when the new PCB is ready to test it I am also wandering does it really worth's or Oldvan is right?