Greetings to all community!
Some time ago I came across a circuit board that was used to control the speed of an 11KW, 3 phase induction motor. Motor was handling a floating and free load (like a blower fan etc). While examining, I observed that is was a pretty simple circuit. Using 3 Triacs and some passive components (almost same like 3 sets of house hold fan dimmer on one board). 3 Triacs, an RC to control Triac firing angle, a snubber and VDRs for surge protection.
Since the circuit was practical and was doing the job very well and also I've a some free time now, so I'm intended to implement it on a PIC Micro-controller (to gain some experience and to add some extra functionality).

First I’ve tried it with single phase 220VAC in my lab. It was quite successful. Detect a zero crossing point from L and N and generate a delayed pulse to fire the Triac (Fig.A). Works fine.

Now I’m moving to 3 phase and stuck on a concept. In single phase, we get zero crossing point from L and N is pretty simple but for 3 phase, how to handle it? (Fig.C). From where I should get the input of ZCD?

I’ve two concepts in my mind that I’ve represented in (Fig.B) and (Fig.D) but not sure if I’m right or wrong.
Need a favor to figure it out. Please.
Thanks.

 

Usually it is a little more involved than the 1ph version.

https://www.industrial-electronics.com/emct_2e_5e.html

 

Hi, Thanks for your response. Circuit in the given link is a nice one for ON/OFF control of a 3phase motor. Working as a solid state relay. Actually I'm in search of speed control using micro-controller and triacs. By the way you pointed me to a very nice site with a lot of interesting stuff. Bookmarked it and I'm sure I'll get lot of information here. Thanks again.
Regards.

 

With motors you should be constantly pulsing the triacs and firing at a specific angle. Since its speed your after the firing angle should be based on average V over the firing intervalI think.

The triac turns off at v=0 and i=0.

 

Safety is the most important and practical aspect of power controller design includes being able to address everything that can go wrong..
An approach could follow something like this which begins with monitoring the source conditioned with uC optical input.
I would not say it is simple concidering everything that goes into a serious industrial application usually takes much computance.
https://learningbyskrai.files.wordp...ga-328-microcontroller-based-firing-angle.pdf

 

The paper suggests that they didn't know what they are doing and cannot explain what they did. A/d converter with 250 mS conversion time? Not sure if they are doing motor or power?

 

With motors you should be constantly pulsing the triacs and firing at a specific angle. Since its speed your after the firing angle should be based on average V over the firing intervalI think.

The triac turns off at v=0 and i=0.

Thanks for reply.
Yes, you are right. I've already done it with my single phase model as I mentioned before. A delayed pulse varying from 1ms to 20ms works fine for 50Hz (1ms to 15ms to be more practical).
In single phase, input of ZCD comes from between L and N. That's OK. What about 3 phase? For example considerTriac1, from where should I get a zero crossing point? This is my actual confusion.

 

I think I see your confusion and it makes sense. it might be fixed with a Δ-Y transform or transformer. See https://en.wikipedia.org/wiki/Y-Δ_transform

https://en.wikipedia.org/wiki/Delta-wye_transformer

Otherwise, you need to look at the phases independently, essentially a transformer, then you can arbitrarily move the references.

There are 3 sine waves, phase to phase, so you have some isolation to deal with. if you isolate the phase to phase, you can reference them all to ground and then you can look at zero crossings.

Opically is one way to trigger, but another is the use of trigger transformers.

 

So, I think if you bring the phases from a primary (on each phase) to a center tapped secondary, with the CT your circuit reference, you would have a zero crossing to detect, So, detect them on each of the transformer secondaries, but do the firing on the actual phase to phase voltage..

Just an FYI thing. When doing heating, you fire based on V^2. Now you can also have on option of a "tungsten" load which has a large temperature coeficient.

What I haven;t figured out, is exactly how current limiting is handled. In the snge phase world, back to back SCR's are preferred. I think, because you can fire them at the same time, thus turning them off.

When operating into an inductive load. you should be able to current limit.

 

The triac turns off at i<Ih where Ih is the holding current, not necessarily at V=0.

What I haven;t figured out, is exactly how current limiting is handled. In the snge phase world, back to back SCR's are preferred. I think, because you can fire them at the same time, thus turning them off.

That doesn't turn them off. You need a GTO to be able to turn it off - and GTOs have been replaced almost completely by IGBTs.
Two SCRs back to back withstand a bigger surge than a triac. The SCR just has to withstand the surge until the end of the half-cycle.

 

So, I think if you bring the phases from a primary (on each phase) to a center tapped secondary, with the CT your circuit reference, you would have a zero crossing to detect, So, detect them on each of the transformer secondaries, but do the firing on the actual phase to phase voltage..

Just an FYI thing. When doing heating, you fire based on V^2. Now you can also have on option of a "tungsten" load which has a large temperature coeficient.

What I haven;t figured out, is exactly how current limiting is handled. In the snge phase world, back to back SCR's are preferred. I think, because you can fire them at the same time, thus turning them off.

When operating into an inductive load. you should be able to current limit.

Delta-Y transformation seems a good idea to deal with problem. But availability, cost and mass of transformers might become hurdle. Well, need your opinion on what about a virtual ground? I'm using op-amps as ZCD. Obviously they have very high input impedance. Transforming Delta-Y might possible with higher value resistors, instead of transformers, without voltage drop or phase signal loss .(See Fig.) Zero crossing signal for triac1 comes from L1 and GND, for Triac2 from L2-GND and so on. Should it work?

 

This https://circuitglobe.com/phase-shift-in-star-delta-transformer.html is the complication.

It doesn't mean it's not possible.

Then, there is this "When the power flow is reversed, and the delta of the transformer acts as primary and star act as secondary, then the voltage phasor does not change, but all the current phasor are reverse." comment from the article.

You might need a "phase detector" for starters.

There are "potential transformers" https://www.hilltech.com/products/p...ment_Transformers.html#Potential Transformers

 

Another possibility is, for each phase, to derive a power supply (e.g. a reactive power supply) which consists of a capacitor and a resistor, Zener regulator, over the top OP-amp (analog.com) and a negative bias generator from TI. It gives you a few hundred mV for your - supply. The Over the Top OP amp accept inputs greater than the supply.

I do remember some Burr-brown) (now TI) Isolation amplifiers that were used to measure mains voltages. There are also some optical isolation amplifiers which use optocouplers which have supposedly match devices. Many isolation amplifiers create some bipolor voltages to use on the isolated side for signal conditioning.

https://www.ti.com/lit/ds/symlink/iso121.pdf?ts=1621084077093#page=12; see figure 8

AMC1100: Replacement of Input Main Sensing Transformer in Inverters with Isolated Amplifier
https://www.ti.com/lit/an/slaa552/s...27719&ref_url=https%3A%2F%2Fwww.google.com%2F

 

This https://www.ti.com/lit/ug/tidua58/tidua58.pdf might help a little.

This https://www.ti.com/lit/ug/tiduby8b/tiduby8b.pdf could prove useful.

the ACS34800. See https://sc-cd-preview.digikey.com/e...e=620&utm_medium=supplier&utm_campaign=buynow

https://www.allegromicro.com/en/pro...-amp-integrated-conductor-sensor-ics/acs37800

It has power, voltage, current and also zero cross outputs. Might be a good solution.

 

This https://www.ti.com/lit/ug/tidua58/tidua58.pdf might help a little.

This https://www.ti.com/lit/ug/tiduby8b/tiduby8b.pdf could prove useful.

the ACS34800. See https://sc-cd-preview.digikey.com/e...e=620&utm_medium=supplier&utm_campaign=buynow

https://www.allegromicro.com/en/pro...-amp-integrated-conductor-sensor-ics/acs37800

It has power, voltage, current and also zero cross outputs. Might be a good solution.

Bundle of thanks for such a detailed reply. I've visited and downloaded the links and gonna study in detail. Obviously The simplest solution would be my first choice. Thanks again. You helped me a lot.