SPOT WELDER WITH PIC16F88 tested - YOUTUBE video

[Ichan]

Putting the business end at the source of the N-Channel mosfet will need high side mosfet driving which will add an unnecessary complexity.

-ichan

[kalpakchiev]

Anyway the new PCB is ready, tonight I'll test it and will post the results

[oldvan]

Putting the business end at the source of the N-Channel mosfet will need high side mosfet driving which will add an unnecessary complexity.

I beg to differ with the word "unnecessary."  High-side MOSFET drive indeed adds a bit of circuitry.

Driving the N-Channel MOSFETS on the high side of things needs little additional, a couple opto-isolators
and a +30V supply capable of a hundred mA or so would do the trick.,

The small additional amount of circuitry provides SAFETY which should always be a driving force in any project.

How many injuries or deaths does it take to negate saving a couple of dollars?

[kalpakchiev]

In another project I have successfully used this schematic for Driving the N-Channel MOSFET on the high side, but this requires a free I/O with PWM.

[Ichan]

I beg to differ with the word "unnecessary."

Granted. 

Kalpakchiev, i really want to see you are successful with this as i will copy it if so  .
My old tab welder is only a crude wiring of caps, scr, relay, and switches. I need to upgrade it.

-ichan

[kalpakchiev]

You are not going to copy it - I can say that we are making it together 

[MAXPAYNE]

You are not going to copy it - I can say that we are making it together 

Thats the spirit we r herewith ............

[kalpakchiev]

PCB is ready. Tests are done. Here some pictures. Next I have to write a conclusion

[kalpakchiev]

So here come conclusion based on experiments with both variants:
Second circuit works, but there is no difference in performance compared with first one.

As Oldvan mentioned about second schematic : ''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. "

It appears that the first variant is better.

[Ichan]

From your picture:

- Looks like the spark is good, but it happen at the wrong place.
- The material is too thick, try to use battery tab taken from old battery pack, even a small cellphone battery has it.
- Do not use plated material, you may try to find 0.1 - 0.2 mm stainless steel sheet, nickel is the best but harder to find.
- Not enough pressure, spot welding need some pressure so the material bent down at the weld point.
- The electrode material is not good enough.

Please try again, i think you are already close.

-ichan

[kalpakchiev]

- Looks like the spark is good, but it happen at the wrong place - these 2 sparks at the left side are from previous try.

- The material is too thick - I want it to be able to weld up to 1mm material, this one was 0.8mm

- Do not use plated material, you may try to find 0.1 - 0.2 mm stainless steel sheet, nickel is the best but harder to find. - thanks, i will have it in mind

- Not enough pressure, spot welding need some pressure so the material bent down at the weld point. - Right, that is important

- The electrode material is not good enough. - It is cooper, but it is not electrode at all - just a piece of thick wire - Do You think a 10mm cooper rod with conical tip will be enough?

Please try again, i think you are already close. - I will, but with first circuit design, it needs some refinements but I consider the entire experimets as successful.

Best regards
 

[Ichan]

I meant burnt mark at the right side with copper color, the spark shouldn't be there but between the welded material.

1 mm is to thick for a welder like this, there is a good DIY project on the web which use modified microwave oven transformer, more appropriate for thick material.

For the electrode, pure copper will be to soft but i do not know what alloy it should be, need to try many copper rod available.

I just quick draw a layout for power section if i will build it later. The schematic is still on my head, it use my proposed driver for discharging mosfet and a P Channel mosfet driven with one npn transistor for charging. For R-Charge i would prefer to use 24V 120W halogen bulb for 5A charging current, yes it will be a very bright charging indicator  .

-ichan

[kalpakchiev]

If you want when you are ready with schematic send it. I can make a protopype and test it.

What is the program with which you draw this PCB?

Regards

[Ichan]

Sure i will be glad if you willing to try it, i will draw the schematic tomorrow.

It is drawn with Sprint Layout, if you haven't tried it this software is good to quick draw a pcb layout without drawing the schematic first, you can find it here in Sonsivri.

Attached the Sprint Layout file of the WeldPower, well.. need to rar it as the forum software doesn't allow .lay extension to be attached.

-ichan

[kalpakchiev]

Sprint Layout supports direct output for milling and drilling 

[Ichan]

Kalpakchiev,

Here is the schematic, you may change anything you want.

-ichan

[kalpakchiev]

I will make it. Ichan, do you think one single IRF9540 will be enough

[Ichan]

I am thinking about 5A charging current set by R-Charge, a 120W 24V halogen bulb to give 5A at 24V Vin is on my mind, so single IRF9540 should be enough as long as it is not a bad quality "clone one".

But you are free to modify it as you like, thanks for the will.

Note: I forgot to put R-Bleeder on the schematic, they are exist on the pcb drawing.

-ichan

[kalpakchiev]

This sounds to me like plan

[sadman]

Sprint Layout supports direct output for milling and drilling 

which software you used for cnc job

regards
sadman

[kalpakchiev]

which software you used for cnc job

regards
sadman

At work : Autograv
At home : KellyCAM

It works with both of them 

Best regards

[titi]

Hello,

To ichan about your drawing "powerweld.png",

there is a problem in the weld driver part.

When WELD input is high, Q7 is closed, and Q6 too, but the current from junction Base-Emetter of Q6 is only limited by R15, so either (Q6 or Q7), or (Q6 and Q7) will by destroyed.

To avoid problem, put a resistor 1K between Base Q6 and Collector Q7.

It whould be better to replace Q6 and R9 by a push-pull with BD139 and BD140 by exemple.

As explain before, this give a strong drive to turn on, and a strong drive to turn off, that give a very precise driving for time to weld and use less power to drive Mosfets.

I hope this help you.

Best Regards.

[waicon]

....and change R9=4k7, and parallel with ziner 12V if VIN>30V. possible remove R1~R5? i prefer less components count.

[Ichan]

To avoid problem, put a resistor 1K between Base Q6 and Collector Q7.
It whould be better to replace Q6 and R9 by a push-pull with BD139 and BD140 by exemple.

The resistor is there on the proposed driver (reply #17) but then i was thinking to saturate the pnp transistor harder to make a faster turn-on. I do not use totem-pole driver as i think a slow turn-off would be better for this application(reply #19).

....and change R9=4k7, and parallel with ziner 12V if VIN>30V. possible remove R1~R5? i prefer less components count.

With R9=4K7 the turn-off will be too slow, will it? Yes adding a zener diode would be safer, R1-R5 to reduce the risk of "Gate Ringing" caused by parallel mosfet and/or stray inductance.

-ichan

[borberk]

What turn on-off time is suitable comparing to 10ms to 200ms weld time? 100us is acheived with 6k drive resistance.