AC control circuit question

[KeepItSimpleStupid]

Yep, KISS for short. Also, when i got involved in the SCR controller Variac system, work was blowing $30.00 USD SCR fuses (I^2)*t fuses left and right. I made sure they used 25A SCR units and used a 3A 3AG suitable fuse for the system wattage. We used low voltage custom tantalum heaters wound in house that mostly operated at <40 V. One operated at like 120V. When the current limiting was employed correctly with the over-rated SCR unit and the cheap fuse; failures dramatically dropped.
The heaters and wiring were all exposed in a vacuum chamber. Later, I said this is ridiculous and obsolescence of our current system forced me to re-think. We went with a recipe programmable temperature controller and a DC power supply. POWER was an important parameter to keep an eye on. A power display on the power supply would have been great.

 

[Kerim]

This is my version on how I analyse the 'original' electronic circuit of this welder (sorry if some of what I say here is already mentioned on previous posts).

The back-to-back SCRs could be seen as one TRIAC.
Their driver here is equivalent to the simple dimmer for TRIAC. In each half-cycle, it provides one current pulse to trigger the TRIAC gate; by a quick discharging of a capacitor via a DIAC. This pulse occurs whenever the capacitor voltage reaches the breakover voltage (positive or negative) of the DIAC. The delay of this pulse is controlled by a resistor (that determines the rate of recharging the discharged capacitor).

In case of resistive load, this delay could be made rather short to get the possible full power. It is limited mainly by the DIAC breakover voltage.

In case of an inductive load, at it is the case here, the TRIAC is in the ‘on’ state even after the mains voltage crosses 0 V. This is due to the magnetizing current. So if the controlling resistance is made smaller than a certain limit, the triggering pulse occurs while the TRIAC is on, and the TRIAC stays ‘off’ for the remaining time of the half-cycle, after its current returns back to zero. This is equivalent to driving the transformer primary by a half-wave rectifier! Obviously when this happens, the transformer current rises to a relatively very high value.

Solution:
The resistance of the controlling resistor should be made higher than a certain value. This minimum value depends on the primary inductive impedance and the possible highest mains voltage. So, one may start with a high value then stops decreasing it when the output power is almost full. A resistor, having the minimum resistance found experimentally (preferably with a suitable margin) could be connected in series with the main volume.

Kerim

 

[KeepItSimpleStupid]

Basically, with an inductive load, the triac turns off at both zero voltage and zero current. With a resistive load, you just have to pick when to turn it on and it works relatively smoothly.

 

[tcmtech]

Just for a heads up the gates on a SCR are for the most part bidirectional in conduction so you can simply put your opto triac between the two gates with one 10 - 100 ohm current limiting resistor and simplify your circuit.

 

[Kerim]

Basically, with an inductive load, the triac turns off at both zero voltage and zero current. With a resistive load, you just have to pick when to turn it on and it works relatively smoothly.

I am afraid that if there is a current supply that forces a current into a triac, after it was turned on, the voltage on its main two terminals cannot be zero (but close to their saturation one). The transformer primary coil acts as a current supply at the end of each half cycle (during its magnetic discharging). And when the magnetizing current becomes lower than the triac holding one, the triac voltage rises (rather quickly) to follow the mains voltage and, here, the role of the snubber comes... to limit dv/dt.

Kerim

 

[spec]

Just for a heads up the gates on a SCR are for the most part bidirectional in conduction so you can simply put your opto triac between the two gates with one 10 - 100 ohm current limiting resistor and simplify your circuit.

Is that some special SCR? I have only come across SCRs that trigger in one quadrant. Some TRIACs trigger in four quadrants , but most modern TRIACs , especially high commutation types, only trigger in three quadrants.

spec

 

[Tony Stewart]

The main cause of thermal failure I think is the PIV when power is extinguished in the arc, the SCR's are driving alternating high current pulses and high V pulse and the VI product transients exceed the SOA for the device even though both parameters may be well below their maxima. Snubbers absorb the voltage transient so that when the SCR transitions from low to high impedance that they absorb the inductive turn-off current rather than cause a larger power transient than the steady state arc pulse.
- low inductance 100mV Current shunt and voltage monitor across SCR is best to see this.

PLastic caps have. very low ESR but can only handle a few Amps RMS , which is easy to measure with differential probes that are balanced and very short gnd clip.across the subber resistor. So many parallel caps are needed or a large oil filled cap. if the RMS current exceeds 3A
Oil only has a dielectric constant of 2-3.

the peak to average current ratios could be as high as 10 or more. This depends on duty cycle or pulse width to rep rate

 

[tcmtech]

Is that some special SCR? I have only come across SCRs that trigger in one quadrant. Some TRIACs trigger in four quadrants , but most modern TRIACs , especially high commutation types, only trigger in three quadrants.

spec

Just do a resistance test on any common power SCR you have and you will see the junction between the Gate and Cathode looks like a near dead short which means that if you have two SCR's wired back to back with their gates tied together the reverse feedback/conduction from one gate will trigger the forward conduction of the other SCR and so on.

Back in the old days before higher voltage solid state optoisolators were cheap and reliable it was a pretty common practice to use a simple mechanical relay that just tied the two SCR's gates together through a bit of resistance to work as a semi-solid state switch.

 

[spec]

Just do a resistance test on any common power SCR you have and you will see the junction between the Gate and Cathode looks like a near dead short which means that if you have two SCR's wired back to back with their gates tied together the reverse feedback/conduction from one gate will trigger the forward conduction of the other SCR and so on.

Back in the old days before higher voltage solid state optoisolators were cheap and reliable it was a pretty common practice to use a simple mechanical relay that just tied the two SCR's gates together through a bit of resistance to work as a semi-solid state switch.

(By back to back I take it that you mean both SCR anodes connected, rather than both cathodes.)

Well, I have never heard of that before- very clever.

Hell, the number of times that would have been useful to know!

spec

 

[Kerim]

Oh, sorry. Me too, I didn't hear yet of an SCR type that could be triggered by a negative current pulse.
But in case such a type does exist now, I wish I know the name (label) of one SCR, in the least, of it.

Some power modules (usually called AC solid state switches) have, instead of a triac, two SCRs (connected anode to cathode), but each has its own pin, on the module, connected to its gate.

Kerim

 

[spec]

Oh, sorry. Me too, I didn't hear yet of an SCR type that could be triggered by a negative current pulse.
But in case such a type does exist now, I wish I know the name (label) of one SCR, in the least, of it.

Some power modules (usually called AC solid state switches) have, instead of a triac, two SCRs (connected anode to cathode), but each has its own pin, on the module, connected to its gate.

Kerim

If what I assumed was correct and the two SCR anodes are connected together, the current following through the opposing gate/cathode resistor is positive. You would need to ensure that you did not exceed the -VGC absolute rating of the opposing SCR though.

spec

 

[tcmtech]

(By back to back I take it that you mean both SCR anodes connected, rather than both cathodes.)

Reverse parallel would be another term for it.
Anode of SCR one tied to Cathode of SCR two and Cathode of SCR 1 tied to Anode of SCR 2.

Oh, sorry. Me too, I didn't hear yet of an SCR type that could be triggered by a negative current pulse.
But in case such a type does exist now, I wish I know the name (label) of one SCR, in the least, of it.

A cheat for getting a negative triggering SCR is to use a triac oriented in a specific way. One way it will work like a normal SCR but when reversed the gate will work as a negative triggered SCR. Minor downside is it needs a diode on the gate to make its turn on monodirectional instead of the normal bidirectional function of a traic.

If what I assumed was correct and the two SCR anodes are connected together, the current following through the opposing gate/cathode resistor is positive. You would need to ensure that you did not exceed the -VGC absolute rating of the opposing SCR though.

That's what the resistor in series with the optoisolator or mechanical switch is for. It limits the current going through the gates to the safe level they can handle at the devices maximum forward voltage drop.

 

[ShawnR]

I have once again redrawn the circuit. I replaced the PUT symbol with the proper one instead of a unijunction one and connected the optotriac as suggested by tcmtech. In the sp100 diagram, 2 inverse parallel diodes are shown in the gate circuit so I added them here with question marks beside them. But, in the previous post, I think these are what tcmtech is referring to

Minor downside is it needs a diode on the gate to make its turn on monodirectional instead of the normal bidirectional function of a traic.

As always, any input is appreciated. I think it is almost ready to breadboard. I plan on breadboarding it with a 120 volt transformer on the bench as I do not 240 there. If all good then I can replace the pc board in the unit with this new one.

A question I have.... is the snubber component values suitable for 240 vac operation? They are the same values as the sp100 which is a 120 vac unit so I would imagine that I have to make some changes, just not sure which values outside of max voltage range on the caps...

Attached also is an excerpt from the sp100.

Thanks all

 

[tcmtech]

Yes that should work just fine. The snubber you have should be okay as well.

 

[Tony Stewart]

If what I assumed was correct and the two SCR anodes are connected together, the current following through the opposing gate/cathode resistor is positive. You would need to ensure that you did not exceed the -VGC absolute rating of the opposing SCR though.

spec

2 SCR's BACK-BACK = 1 TRIAC by design

 

[ShawnR]

I think that I drew the PUT connections wrong. If I understand PUTS correctly, the anode voltage has to exceed the gate voltage by 0.7 volts to fire. I was reviewing the circuit before I started building the prototype when I noted it. I will upload a new drawing later.

 

[ShawnR]

Attached is the latest version. I am wondering if D6 is a bad idea, an LED to indicated trigger on. Any input appreciated.

Changed attached image. I had uploaded same one as previous.

Thanks, Cheers, Shawn

 

[Kerim]

Attached is the latest version. I am wondering if D6 is a bad idea, an LED to indicated trigger on. Any input appreciated.


Thanks, Cheers, Shawn

Since your circuit provides one triggering pulse in every half cycle, one of the two SCRs may miss it and stay off after its (inductive) lagging current returns to zero (soon after the voltage zero-crossing). This is likely to happen at a small triggering phase. A transformer doesn’t like being driven by half-cycles of the same polarity... I guess you know why and what happens to its input current in this case.

It seems I couldn't explain this point well on my previous post #42.


Kerim

Edited:
But it the output triac of MOC3020 will turn on, even at the SCR saturation voltage, perhaps there will be no problem.

 

[alec_t]

What purpose do D4 and D5 serve?
Won't you need a snubber across SCR1?

 

[ShawnR]

Kerim - thanks for the reminder. Before I put any circuit into use, I had best review many of the good advice I have received here. There was a post by schmidt trigger in #34 regarding ensuring symmetry as well with a resistive load. To do a trial for your advice, I will need to be welding for a load on the transformer? Or is simply at idle ok?

I tried some test circuits last night with the PUT. It's hold on current (valley current?) is in the microamps so I am wondering how it ever shuts off. A tutorial on PUTs addresses this in picking resistor values. I may have to change the pot in my foot pedal for another value. Combine that with your input and I have many things to consider. I think I have to build some circuits on the bench to get a better handle on it all. I know pretty simple for many of you who work in this field but kind of new to me. There are many things I was not aware of. tcmtech steered me in this direction for a more stable circuit than what I started with so I want to make sure I do it up right. My biggest concern is not to blow up another SCR module in 30 seconds or less....

Considering what I knew of AC control circuits a month ago, I have come a long way, but still more to learn, obviously.

alec, see post #52 and a few before. Also, the sp100 circuit has them. I am not going to pretend I fully grasp them, but am trusting....
SCR1 is in parallel (reverse parallel) with scr2 so one snubber circuit serves both.

Cheers,