AC control circuit question

[Tony Stewart]

D1 purpose? redundant
R7 current is 170mA is excessive, and can be driven direct from 555 Q out , use recommended current.
D2-3 purpose?
missing ferrite beads on drivers.
arc trigger and short circuit current control missing.
magnetic air gap current limited transformer missing
.... actually a lot more missing. try to understand how a good arc welder design works first.

 

[ShawnR]

Thanks Tony
I thought I would need D1 to prevent the filtering effect from C1 interfering with my zero detector components. I will experiment with taking it out.
R7 was mislabelled. It shound have been 120 ohm. A typo. R5 should have been 22 ohms
The Q, forgot to label it too, is intended as an inverter, but yes, would make it simpler. I will reconfigure that and drive it direct off the timer.
D2 D3 & R5 of 22 ohms, based on input received here, post #44
I have never used or researched ferrite beads as I have not done any of this type of circuit before but yes, great suggestion. I don't think I need the other arc welder info as I am only trying to control the primary of an old, transformer stick welder with the intention of reducing the output to allow me to weld aluminum more easily. Not much beyond that.
I am learning a lot. That is a big part of why I am doing this rather than just buying a TIG welder. Thanks for your suggestions

 

[tcmtech]

You will need D1 to do exactly what you think it's doing.

You also don't need the mosfet between the 555 IC output and the optotriac.

The 555 timer ICs have totem pole outputs so they can sink or source up to 200 ma either way. If you need to invert an output signal just set it up to sink power from the postive side of supply (Optotriac on when output is low) rather than source it from the output to the negative rail (Optotriac on when output is high.)

 

[ShawnR]

Hey tcmtech, thanks for coming back. I was hoping you would chime in again since you wanted me going in this direction in the first place.

Attached is the latest. Not many changes other than the timer output, added a led indicator for power and put the correct part number in for U1. I ran it too, on the bench with a resistive load for the welder primary and the voltage varies quite nicely from near zero to near max. Perfect for my application.
Thanks
Cheers,

 

[tcmtech]

Good to know. I may have to copy it someday.

Nice clean schematic drawing too! You came a long ways in less than a year.

 

[ShawnR]

tcmtech, you steered me in the direction whereby the gate triggers for the SCR's is simply two inverse parallel diodes and a 22 ohm resistor (10- 100). I am not sure I fully understand how that works but I am wondering what kind of current that resistor sees. ie what power resistor do I need...? I am assembling a pc board. I upgraded the circuit to 12 volts instead of 5. The reason being that the transformer I have is 12 volt so there will be less power dissipated in the regulator and the foot pedal cord is about 15 feet long so I figured it would be better with a 12 volt driver versus a 5 volt one...not sure if it is even relevant but it did not affect the circuit much....I hope since I did not breadboard the 12 volt version.

 

[ShawnR]

I think I might get it. ....the gate signal is very brief and essentially the gate circuit is shorted out by the scr when in the on state so the gate resistor does not have any appreciable power dissipation. Close? So a half watt would be fine. ...

 

[tcmtech]

looks like you're doing just fine without me!

I don't think the back to back diodes on the AC side of the opto triac are absolutely necessary. I've seen as many circuits with them as without.

The 22 ohm resistor to limit the firing inrush current through the optocoupler is about the only somewhat important part in my opinion and even then I have seen many circuits without that as well.

 

[ShawnR]

I installed the circuit in the welder today. Unfortunately, the scr module shorted just like the last one.

I just went back to the lincoln diagram posted here back in the beginning, post #4, and my circuit is almost identical. One thing I don't have is d1-r2 and d2-r4 ( https://www.manualslib.com/manual/95205/Lincoln-Electric-Sp-100.html?page=60#manual ). Do they add a level of protection for the gates?

Other than that, could it be the scr module is just not big enough? The next one I order will be much larger but I am looking for a better excuse for them shorting so quickly. I did not see any control today. Last time, I had about 30 seconds of good power control before it shorted.

Digikey is liking me.....

 

[tcmtech]

I would try it being at no load if your running 1600 volt 100 amp SCR pair they should be handling the power just fine until you load it up heavy.

Without O-scope readings to see what sort of voltage spikes you are getting and where they are at it's just guesswork this point.

I'm assuming you're running the 1600 volt 100 amp units now?

It may be worth putting a good sized MOV set across the transformer's primary as well as across the SCR's. If you getting severe over voltage spikes they will pop first and give you an idea about what needs redesign.

If you were killing them with overcurrent through the gates the Opto triac should have blown itself to bits if the 22 ohm resistor didn't go first.

 

[tcmtech]

Just a thought.

It may be worth your time to set up a large heater or light bulb array to work as a current limiter for experimenting with this setup.

That or run the welder on a much lower input voltage while trouble shooting things rather than going directly to line voltage and current levels first shoot after a circuit redesign.

Or both.

Ya shoulda took me up on that offer to send you a box of surplus SCR's and whatnot. Burning up near free good used ones would have been way cheaper than doing it with new ones.

 

[ShawnR]

The unit was what I had, the same one. I was so optimistic that the addition of the snubber circuit would protect it. I will be ordering larger ones, obviously.

I can set up some resistive light bulb loads and maybe get some scope images.

I have never bought MOVS before. Would this one be OK. ..?

https://www.digikey.ca/product-detail/en/littelfuse-inc/V271HF34/F5391-ND/2501444

I was hoping to get some units from you but did not hear back. (2nd pm) Figured the offer closed. My bad.

The welder was set to 100 amps so primary should have been around 20ish max? Well within the specs of the scr but obviously still not enough.

Fascinating project for me....and frustrating too!

Thanks again

 

[tcmtech]

Sorry. I more than likely missed reading it and the conversation got cleared out along the way.

I still have them but I am taking off to drive truck for sugar beet harvest tomorrow for the next two weeks ans will be gone.

Those big MOV's should do well.

 

[ShawnR]

No problem. I appreciate the help and offer. I will still order another one from Digikey and we can chat when you are back. I need to order the MOV's anyways. As I said, I have not used MOVs before so pardon my ignorance but one across the AC in and one across the SCR's or do they go from supply to chassis ground? I can go google it too I suppose...

 

[tcmtech]

Typical procedure now is to have multiple devices. Line to line, line to neutral and across the switching devices that way there is no route a HV spike can take that's not covered.

 

[ShawnR]

I made some progress! But at the cost of some embarrassment...

I took the circuit board out of the welder and set up a test fixture on the bench, using two 120 light bulbs in series for the 240 v load (what I had handy)
I also made up a small circuit board with two small SCRs to replace my power module. (which is shorted)

The circuit still did not work but I did note that when I checked the voltage on pin 3 of the timer, the light bulbs lit up. (odd) I finally realized (after staring at the circuit board...) I had forgot to install R8, the resistor in the zero detector circuit.

After replacing that, I had great success. I did find that the load voltage is around 50 v with the pedal not pushed, then as soon as I move it, the voltage drops to about 22 (better) and then climbs with proportion to the pedal position. Not a big deal and probably something to do with the mechanics in the pedal but for now, happy it is working. I can now plow forward with installing some MOV's as suggested and ordering SCR module #3

I have also been in contact with the manufacturer of the SCR module and they are giving me some input so I am becoming optimistic again....

Here is video for those so inclined...

Cheers,

 

[ShawnR]

A question for those handy with the 555. In the last circuit variation, the foot pedal delays the trigger point of the scrs. But the once the scr gate signal goes on, it stays on until the next zero crossing. What would be the simplest ( because I have built the circuit into a pcb already) way of having a 1 msec trigger pulse instead of a constant on trigger signal? Do I need a second timer chip? I have done some reading online but the 555 configurations mostly result in a timed pulse as a result of a trigger. I am using the zero trigger signal to start a delay so I suspect i need a second time chip.

Having asked that, does having the gate signal on for the duration of the scr on time potentially cause an issue with the scr?

Another thing I am considering is using the timer reset function to shut the output off. This might be the easiest modification. Perhaps the level of the zero crossing signal could cause a timer rest. ..? Just thinking out loud. ....

Thanks all

 

[tcmtech]

If the zero cross detect part of the circuit is resetting the gate triggering timing cycle there is no issue with how much of the cycle the gate is powered on for.

Once the SCR fires the gate signal being on or off until the zero crossing event resets thing has no specific effect.

 

[ShawnR]

OK, good to know. Thanks. I managed to get some more scope signals. I also found out why the load is on when the foot pedal is not depressed. I think it comes back to the value of R4. I exchanged it for a 100K pot, expecting to have to increase the value beyond 10K and was able to "tune" the zero foot pedal point (no pressure on pedal) to almost zero power almost. I was surprised to find the new pot is set at 4K. Opposite of what I expected but as I worked through it later, it seems to make sense. C3 is probably not what I expected, exactly. Video links are below I suspect that the extra 10K (R4) with the actual C3 value lead to timing delay extending into the next half cycle. Once I could get the actual delay down to within the period of a half cycle, it works much better.

These videos are before I replaced R4 with a pot. Sorry, but I can't seem to rotate the video but here is a link to the signal at pin 6

And here is the resulting signal at pin 3

I am stumbling forward.....most days. ;-)

 

[ShawnR]

I have made some headway. I was welding today with great success....but in a 2 steps forward, one back kind of way.

Changes to the circuit include increasing the value of R5 from 22 to 300 ohms. In the lincoln manual for the sp100, this resistor is 150 ohm so I figured with twice the supply voltage, this needs to increase. The engineer from IXYS felt that the problem with blowing the SCR's the way I was (going to a short vs an open) was a result of the gate trigger current too high but admitted it was a hard value to calculate. Even at 120 ohms, I blew scrs but at 1K, the circuit did not work.

He felt the timing circuit was ok but the trigger circuit was the issue. Normally, these scrs open on a fail and I was consistently getting a short.

I also changed C3 to a .012 capacitor which limits my delay to well under 8.3 mSecs, ensuring I do not trigger past the zero point, into the next cycle. I am not sure how close to the next zero point I can delay my on time but with practice, I will be able to set the welder for the values I need.

I have also installed MOV's across the SCRs and one from each line side to ground (earth)

I also upped the size of the SCR module. (tcmtech, you will be happy to finally hear... ) Again, the engineer felt like I had lots of "headroom" with the other one but the cost difference was not much more for the added security.

So, welding went well....given my ability levels. Control was good.

The new problem is blowing the main breaker (feeding the garage) or sometimes the welder breaker. These breakers can easily handle the welding current it self. I have welded at much higher currents before so I believe this new problem is to do with the high frequency box. I think I need to install a filter or filters on the line. I am guessing MOV's either in the panel or in the welder? These units were built to work together (welder and hi frequency box) and the high frequency box is meant to do exactly what I am doing with it but today, I turned it on after the welder was running and the breaker blew immediately. I reset the breaker and proceeded to weld for about 10 minutes. Sometimes the breaker goes when I am not welding so I am pretty sure it is something to do with the hi freq. Any suggestions?

Picture of installation is attached for those interested in the hardware end.

Thanks all.

Shawn