AC control circuit question

[ShawnR]

I installed a ciruit in my welder to control the output current via a foot pedal.

**broken link removed** (the only difference is the potentiometer is a 500 K, not 250..he comments on this in the write up)

I completed the installation last night and tried it out today. It worked great for about 30 seconds, then went to full power. I found one of the scrs in the module shorted out on me. The SCR module is this one http://www.digikey.ca/product-detail/en/ixys/MCC56-16IO1B/MCC56-16IO1B-ND/1652739

So, a couple of questions.

The fella who did the website suggested a capacitor be installed into the welder to protect the welder when the hi frequency box is running. Could the HF have taken out the SCR? I cannot find "bathtub" capacitors on Digikey. They sound like very old technology. What would be a suitable capacitor to protect the welder circuitry from the HF? Or a better way of doing it? Any .05 uf 1000 volt capacitor? The HF box spec out at 3500 kv so I would think that is what I should be looking to protect from or am I missing something?

The circuit breaker feeding the receptacle is a 50 A. The welder (250 AmpAC/DC) is probably rated for less but I am running the welder down around a 100 amps output (max set by front panel control) so input current should have been much lower, guessing 15-20 amps (ish?) no where near the current rating of the SCR.

I am concerned about ordering a replacement SCR and have the same thing happen. Any input for me? I have never done much AC power control so not too familiar with the circuit and general rules (ie rules of thumbs)

Thanks

 

[ronv]

I installed a ciruit in my welder to control the output current via a foot pedal.

**broken link removed** (the only difference is the potentiometer is a 500 K, not 250..he comments on this in the write up)

I completed the installation last night and tried it out today. It worked great for about 30 seconds, then went to full power. I found one of the scrs in the module shorted out on me. The SCR module is this one http://www.digikey.ca/product-detail/en/ixys/MCC56-16IO1B/MCC56-16IO1B-ND/1652739

So, a couple of questions.

The fella who did the website suggested a capacitor be installed into the welder to protect the welder when the hi frequency box is running. Could the HF have taken out the SCR? I cannot find "bathtub" capacitors on Digikey. They sound like very old technology. What would be a suitable capacitor to protect the welder circuitry from the HF? Or a better way of doing it? Any .05 uf 1000 volt capacitor? The HF box spec out at 3500 kv so I would think that is what I should be looking to protect from or am I missing something?

The circuit breaker feeding the receptacle is a 50 A. The welder (250 AmpAC/DC) is probably rated for less but I am running the welder down around a 100 amps output (max set by front panel control) so input current should have been much lower, guessing 15-20 amps (ish?) no where near the current rating of the SCR.

I am concerned about ordering a replacement SCR and have the same thing happen. Any input for me? I have never done much AC power control so not too familiar with the circuit and general rules (ie rules of thumbs)

Thanks

Old oil filled caps.
I think I would try these.
**broken link removed**

 

[spec]

Hi Shawn,

I can't really figure out what is going on with the welder modifications but here are a couple of observations for you.
If a power semiconductor, like an SCR, has gone short circuit, it is normally sign that the junction inside has overheated and literally melted. Generally this can be for two reasons. Either the SCR current is simply under specified or the SCR does not have enough heat sinking- you would need a massive heatsink for your application. If the SCRs are handling the full welding current, I would expect you would need SCRs with a current rating of more like 300A rather than 60A, assuming a maximum welding current of 250A that is.

The fact that the SCRs worked OK for a short period and then failed supports the above theory.

But don't take what I say as gospel because I have not analyzed your welder circuit, just some food for thought.

spec

 

[tcmtech]

Okay? That's a pretty crude and very poorly protected circuit.

Yes you need a good sized snubber and spike protection circuit on that.

As it is designed you are trying to use a very crude phase angle delay based voltage regulation circuit (giant but cheaply made light dimmer) on a highly inductive and less than stable load which although the basic numbers may say that it running under 50 amps and 240 volts the reality is you are likely hitting those SCR's with massive peak currents when they switch on under certain conditions which in turns has the effect of creating very high inductive pulse effects that could easily be topping their 1600 volt ratings being there is no snubber or spike suppression anywhere in the primary power circuit.


I would like to explain it in detail but it would take too long so here is a link to the actual service manual for the Lincoln SP-100 which although it's a MIG unit has a very simple yet very well designed dual SCR line side based power and current system that is easily scaled to work with higher voltage and current applications with minimal modification other than just using higher voltage a current rated parts on the power circuit.

https://www.manualslib.com/manual/95205/Lincoln-Electric-Sp-100.html?page=60#manual

Around the back 10 or so pages is where you need to look for the circuit schematics and parts lists to get a good idea of what you need to design around. SCR 1 and 2 on the last models made with that circuit were something like 60 - 75 amp 600 - 800 volt units despite this being a 120 volt 20 amp input rated machine.

 

[ShawnR]

Ronv, thanks for the lead on the caps.

Spec, the scrs are on the primary side of the transformer so I figure very little current, relative to welding current. I had the max welding current set to 100 of a possible 250 so guessing the scrs were only rectifying up to 20 amps max... .? They are rated for rms of 100 (if I am reading specs correctly) . I mounted the block to the metal chassis of the welder but did not note how hot it was right after it failed. I will need to look at the heat sinking again though, for the replacement.

tcmtech, Thanks too. I have an SP-100, interesting it came up. It looks like a very similar circuit so this must be a standard configuration for controlling AC..? I will look for some more info on "snubbers". If I understand correctly, in the attached excerpt from the schematic of the sp100, C25, C2 and R1 form the snubber?

If I have that right, then yes, I can see how the circuit I used has no protection.
Can I copy those values or is there a way for me to figure out what value caps and resistors I should be using ?

Thanks all.

 

[spec]

Spec, the scrs are on the primary side of the transformer so I figure very little current, relative to welding current. I had the max welding current set to 100 of a possible 250 so guessing the scrs were only rectifying up to 20 amps max... .? They are rated for rms of 100 (if I am reading specs correctly) . I mounted the block to the metal chassis of the welder but did not note how hot it was right after it failed. I will need to look at the heat sinking again though, for the replacement.

Apologies Shawn. I didn't have time to analyze your circuit as I said. The SCRs are rated at 1600V so there will not be a voltage problem. But even at 20A there is liable to be in the order of 20W dissipation. If the metal chassis is hot due to other heating that could over temp the SCRs. Also, if the welder chassis is steel, especially thin steel, that would not give much heat dissipation.

Just a check, are these the SCRs you are using: https://ixapps.ixys.com/DataSheet/MCC56-16io1B.pdf

spec

 

[ShawnR]

yes spec, that is the one (I was using, for 30 seconds... )

Thanks

 

[spec]

yes spec, that is the one (I was using, for 30 seconds... )

Thanks

Oh dear- expensive.

You could probably find cheaper replacement SCRs though. 1600V is not required. 700V would be safe, off the top of my head, and 350V if you were definitely not going to use 240V mains.

Bed time for me now, but I will have a closer look at what is going on in about 9 hours time (9am GMT)

Can you let us know where you are- or have you already stated that.

spec

 

[ShawnR]

Physically? Ontario, Canada. ;-)
Electrically speaking? Learning about protection in AC circuits I guess. I have not moved on it yet as I simply don't know enough ....

I did not want to order parts yet as I am not sure what all I will be needing. I am thinking the SCR module I had was fine and will order the same again, but need to add a snubber circuit. I have been reading some articles on line. Some are more heavy reading than I care for. I am leaning to copying the component values from the SP100 welder circuit as that is similar, although my old transformer welder may have somewhat larger currents due to the larger transformer thus inductance. I am thinking that this is one of those situations where something is better than nothing.

There are many articles online about people doing similar projects using an Arduino so I am thinking that they must have installed a snubber of some sort in their power control circuit. I will keep looking. Another factor is that I have the HF box on the welder. Not sure if this was a factor in the failure. It induces a 3500 volt frequency on the welding leads to help start an arc, (and maintain it on AC) Right now, with the info you all have provided, I think my problem is the lack of a snubber in the power switching, as pointed out. I don't think the HF box was the issue so I am leaning that way right now.

 

[tcmtech]

So what's a larger fixed rate USPS shipping box stuffed full of mini and regular Hockey puck plus assorted stud and mini brick dual SCR's from industrial power control system take offs worth to you? 50 - ~600+ amps @ 350 - 1200+ volts.

I was digging around my shop and found one of my 5 gallon buckets of assorted take off's that I probably won't put to any good use anytime soon if ever and thought about your thread.

They all work.
Some are pretty dirty and rough while others look like brand new but a bit dusty and everywhere in between.
I'll probably toss a few other high-powered goodies in there as well. Snubbers, diodes, maybe a huge power transistor or IGBT block or two. I get pretty generous on shop clean outs of this sort of stuff.

Send me a PM with an offer if you're interested. I take Paypal.

 

[spec]

Another general point. I noticed that the chap in the article bought his SCRs from Ebay. If you did the same there is a chance they might be fakes and that may account for the failure, if they have adequate heat sinking that is. Just to check, has the faulty SCR gone short circuit/low resistance between the anode/cathode.

spec

 

[ShawnR]

All parts were from Digikey.

Here is the diagram I used with the 3 components (circled in yellow) drawn in with values from the sp100 drawing. Is this something that looks ok or should I change the values of the components? Any other suggestions? Some of the reading I was doing on snubber design got into determining inductance in the circuit before doing any calculations and since this transformer arc welder looks like nothing but inductance to me, giving a go at some actual math is a little overwhelming.

Back to those bathtub capacitors mentioned earlier, every article I found on installing a high frequency box mentions installing the same lincoln kit into the welder first regardless of make....something I did not do. Since they are only available at a surplus centre, I would think that the technology has been replaced by something such as ronv suggests so will give the metallized prop capacitors a shot.

These are the caps I figured on using

**broken link removed**

I am thinking that the SCR was not taken out by the HF as much as the lack of a snubber but I am just guessing at this point. Regardless, both mods need to be installed before I gamble another SCR module.

Thanks all

 

[ShawnR]

"Send me a PM with an offer if you're interested...."

Not sure how to do that..is that the same as "Start a conversation"?

 

[spec]

"Send me a PM with an offer if you're interested...."

Not sure how to do that..is that the same as "Start a conversation"?

To start a conversation just click on a members avatar and then click on 'start conversation'
spec

 

[tcmtech]

I got to thinking more about your circuit and if it was me I would use a dedicated phase angle delay circuit to do the work.

To do that you would use a small step down transformer to power a simple zero crossing detect circuit that resets a 555 timer that is configured as a delay timer that in turn powers an opto coupler triac as is used in the SP-100 control circuit.

By doing so the zero cross detect would reset the 555 timer IC every half cycle and the 555 timer would be set up to be able to fire after ~ .1 - 10 milliseconds giving you a nice clean consistent phase angle firing delay of 0 - 100%.

Now as for your 'Bathtub capacitor' I am guessing that is nothing more than a common AC motor run or power factor correction type which yes I would definitely recommend having on the transformer side of things.

 

[spec]

Shawn,

I have had a good look at the circuit for the SCR control of the input voltage to the welder transformer and I just can't see how it could work. At the very least it appears as though both of the thyristors are the wrong way around. The other problem is that there is not enough energy to trigger the thyristors reliability. That is my feeling at the moment.

An alternative approach along the lines that tcmtech suggested above would give much better control and reliability.

If you were interested, it would be possible to incorporate current control rather than voltage control. This is the way it would work:
(1) you set the weld current that you want. The weld current could be controlled by the footswitch.
(2) the welding voltage stays at maximum so that you get a good striking ark but the current is limited to the value that you set.

I have also found some thyristors to do the job that are a lot cheaper than the one you first used. This is probably the best choice:
ST TH5050-12WY, 1200V @ 50A ave, 50mA gate trigger @ £6.19 UK from Mouser
https://www.mouser.com/ds/2/389/DM00131778-541736.pdf
https://www.mouser.co.uk/ProductDetail/STMicroelectronics/TN5050H-12WY/?qs=XhO1W%2b5LdJ/YQjCLqctTPA==

spec

 

[ShawnR]

tcmtech and spec. Thanks again for your continued interest and support. Much appreciated. You both have me interested in redoing the circuit for a better control but at this point, I have installed it all in the welder and want to move on. I tried to quit being lazy and went through the circuit. I think I see what you mean spec with regards to the scr's being backwards. With my limited recall of old theory, it seems to make sense. Could I use the circuit components as shown, but rearranging the scrs as shown in the image attached to this post? I know the diacs break down at a given value, I think 24-30 volts. Rather than ask you(s) to type in the circuit workings, I will walk through my reasoning and maybe you can just correct me a little if I get off track...
As the power starts going positive, (line going direct to transformer), it cannot complete the path due to SCR1 being off. A charge cycle starts through D3, R5, R4 and C2. As C2 charges to the breakover voltage of D1, at which time it triggers SCR1, completing the path... and the cycle repeats on the other side. Is that a proper analogy or did I miss?

Because I am basically impatient,.......I ordered the snubber components as per the SP100 post above and a new SCR so hopefully, it all can be used. If not, another lesson learned....

And spec, if I have to order more parts, I will try the ones you suggest. Certainly much cheaper. Not available through digikey and I have not used Mouser in Canada yet but could try them. Or I could spec something similar out with Digikey.

Thanks.
Shawno

 

[ShawnR]

A photo of the installation for those interested. And one of the foot pedal. It contains the pot R4 and a microswitch which initiates (or will ) events such as hf start and gas flow.

 

[tcmtech]

I seriously doubt those To-247 case 50 amp SCR's are going to hold up to 240 VAC high inductive load service.

The problem is when you have a load on your secondary of the transformer the inductive impedance or the primary drops off proportionally plus that impedance is heavily dependent on the relative magnetic field motion and thus its counter EMF effects throughout the whole period of 50 - 60 HZ sine wave.

When you change to a phase angle delay control there are points in the transformers core magnetic field flux transitions where if the core desaturates fast enough due to its inherent magnetic properties plus secondary load that the SCR's may be firing into nothing but the primary coils native resistance due to the counter EMF being extremely low or possibly moving in the same relative direction as the magnetic field the primary coil generates as it starts to conduct.
Given that it's entirely possible that the SCR's could be firing into a less than 1 ohm load at the ~340 volt peak of the 240 volt RMS sine wave.

Basically it's like trying to forcibly hit a nail with a hammer while the nail is moving away from you at the same speed as you swing. Lots of energy involved but nothing to absorb it on the overshoot but your own arm.

That's why the old SP-100's despite only working ana 120 volt 20 amp input rating and the transformers having a ~ 2 - 3 ohm resistance needed to have a 50 - 75 amp capable SCR sets.

Given your unit I would not be trying to run with anything under a 300 amp rated SCR set.

Your 50 amp units might work for a while but the first time you stick your electrode and step on your control peddle I would not be surprised to see them pop.

BTW years ago I used to be a service tech at a local welding supply store and I worked on a lot of units that used SCR based primary control circuits and they always ran with devices that had substantially larger current and voltage ratings compared to their power source volts and amps input numbers. 3 - 4X on voltage and 6 - 8 X on current were common on the larger units.

The heaviest ones I ever regularly worked on were some old Western Arctronics commercial spot welders. They were rated at an input of 240 VAC ~150 amps and only fired for 3 - 10 cycles on a spot weld (but did up to 40 a minute while using a dedicated zero cross detection firing system) and used 1000+ amp 1600 Volt SCR's but blowing one or both to bits 1 - 3 times a year was still common.

That's why I have concerns that your 50 amp units won't take the abuse for very long snubbers or otherwise.

 

[spec]

tcmtech and spec. Thanks again for your continued interest and support. Much appreciated. You both have me interested in redoing the circuit for a better control but at this point, I have installed it all in the welder and want to move on. I tried to quit being lazy and went through the circuit. I think I see what you mean spec with regards to the scr's being backwards. With my limited recall of old theory, it seems to make sense. Could I use the circuit components as shown, but rearranging the scrs as shown in the image attached to this post? I know the diacs break down at a given value, I think 24-30 volts. Rather than ask you(s) to type in the circuit workings, I will walk through my reasoning and maybe you can just correct me a little if I get off track...
As the power starts going positive, (line going direct to transformer), it cannot complete the path due to SCR1 being off. A charge cycle starts through D3, R5, R4 and C2. As C2 charges to the breakover voltage of D1, at which time it triggers SCR1, completing the path... and the cycle repeats on the other side. Is that a proper analogy or did I miss?

Because I am basically impatient,.......I ordered the snubber components as per the SP100 post above and a new SCR so hopefully, it all can be used. If not, another lesson learned....

And spec, if I have to order more parts, I will try the ones you suggest. Certainly much cheaper. Not available through digikey and I have not used Mouser in Canada yet but could try them. Or I could spec something similar out with Digikey.

Thanks.
Shawno

Hy Shawn,

No worries about investigating your circuit; once again, I have learned a lot from the exercise, especially as you have been a good poster and provided the technical data needed and responded quickly to our questions (not always the case on ETO with some posters).

It is reassuring that you agree with my thoughts about reversing the SCRs. The moment I saw the circuit something bothered me about it, but I couldn't believe that a circuit published on the web could have such a gross error. That was what threw me.

Apart from that though, due to the high value of the foot control resistor (250K) and the corresponding low valve of the caps (100nF) to get the phase angle right for 50Hz/60Hz, there simply does not appear to be sufficient energy to trigger the SCRs on reliably- lets hope you get SCRs that require much less gate current than the spec sheet worst case. The alternative SCR that I suggested is better in that respect because they only need 50mA gate current, worst case, to trigger. I found around five suitable alternative SCRs, but the one recommended is the latest generation from ST and has some improved features.

In any instance, I would suggest that you put 5K6 resistors between the gate and cathode of both SCRs. That will improve commutation by providing a path to discharge parasitic capacitances and also drain any gate leakage current. It will also make the gates more immune to spurious triggering due to the general hash, both transmitted and conducted, from the welding process.

One of the problems with SCRs, and more so with TRIACs, is the commutation performance. This means how sensitive they are to high speed transients. This is the reason why snubber circuits are fitted, rather than just for protection. Most modern SCRs are the equivalent of triac HiCom types which do not need snubbers. The other, related characteristic that has been improved is the dV/dt and dI/dt ratings which, if exceeded might damage the SCR, but more likely would cause spurious commutation.

About the operation of the circuit, yes you are correct, that is how it works.

I can understand how you might feel about not perusing another approach. I expect you just want to get on and do some welding, and good luck to you.

spec