3-phase Controlled Soft Starter

[schow]

Dear all,

 

We are now developing a 3-phase anti-parallel SCR controlled soft starter for our own use.

 

We have 3 zero crossing circuits to detect the zero cross of L1, L2 & L3 and 3 SCR firing circuits to trigger the 3 sets of anti-parallel SCRs.

 

The idea is to fire these 3 sets SCRs at an angle just before zero cross and gradually increase the firing angle and eventually fully turn on all the 3 sets SCRs.

 

The problem is when we started to fire these SCRs, the motor started to ramp up slowly at CW direction at very low speed, after a short while, the motor started to jerk and then the motor started to ramp up at opposite CCW direction smoothly until it reached full voltage (SCRs fully on).

 

Does anyone have any idea what is the cause of this problem and how to overcome this problem?

 

All comments are welcome!

 

Cheers!

[marke]

Hello schow

 

You need to alter the initial firing angle to a point where the initial reversal does not happen. If your initial firng angle is too low, you get uncontrolled current flow.

 

Look at the actual current flows through the SCRs. current that flows in through phase 1 must flow out through phase 2 and phase three. You must have an overlap between the phases.

Also consider when the curent stops flowing through the SCRs. This happens when the current passes through zero, so it is influenced by the inductance of the motor.

 

I would suggest that you monitor the firing angle and determine the angle where the motor begins to rotate in the correct direction and start from there. You may still need to alter it a little.

 

Best regards,

Mark.

[schow]

Hi Marke,

 

Thanks for your reply and explaination!

 

Will try it out and see if it works but I am a bit curious as we do have a 2-phase controlled SST that works fine with the same control algorithm, pls can you tell me why it works with the 2-phase controlled but not 3-phase controlled SST?

 

Many thanks!

[marke]

Hello schow

 

With single phase and two controlled phase three phase controllers, current only passes through one SCR, therefore have one SCR on at a time is not an issue.

With a three phase three wire system, current must pass through two SCRs, therefore you must have two SCRs on at any one time.

 

Best regards,

Mark.

[schow]

Hi Marke,

 

Thanks for your reply.

 

I had tried your advice to increase the firing angle of all 3 sets SCRs, but it seems that it only works fine after the firing angle had been increased more than 50% of the norminal voltage (start at around 260V of the 3-ph 415 system), anything below that still experienced the last ratation problem. Hence, it looks like a "kick-start" rather than "soft start".

 

I had tried to fire the 3rd sets of SCR at all time along the ramp-up process (just act like a 2-phase controlled sst) and by doing this, the motor can be ramped up smoothly from almost 0 volt.

 

Are all the 3-phase controlled sst in the market not starting from almost 0 volt? or there are some other control algorithm to make it starts from close to 0 volt?

 

Cheers!

[marke]

Hello schow

 

How are you "triggering" the SCRs? You need to ensure that there is a return path for the current from one SCR to the next. This means that at least two SCRs must be on for current to flow.

 

Best regards,

Mark.

[schow]

Hi Marke,

 

The 3 sets of anti-parallel SCRs are control by 3 opto-triac (each opto-triac control 2 anti-parallel SCRs).

 

The firing sequence is:-

 

fire the L1, U phase opto (after detecting the zero cross angle and some delay to push the firing angle to near next zero cross) ->

 

fire the L2, V phase opto ->

fire the L3, W phase opto (control algorithm same as phase L1, U)

 

Cheers!

[marke]

Hello schow

 

This raises a number of questions.

1. What is the supply voltage (phase to phase)?
2. What is the voltage rating of the opto triac?
3. How do you trigger the opto triac? How long is the trigger pulse?

the firing angle had been increased more than 50% of the norminal voltage (start at around 260V of the 3-ph 415 system)

Use the current as an indication of the effective voltage. The locked rotor current is typically 6 - 7 times the rated current of the motor. If the current is much less than the LRC, then the voltage is down also. At 50% voltage you would draw around 300 - 350% rated current.

In reality, it is rare to start a machine (not an bare motor) with a start current less than 250% so it is not important to have a start current much below this.

 

Best regards,

Mark.

[schow]

Hi Marke,

 

Pls see below my answers:-

 

[*]What is the supply voltage (phase to phase)?

Ans: 415VAC

 

[*]What is the voltage rating of the opto triac?

Ans: 600V (I put 2 in series to build up 1200V)

 

[*]How do you trigger the opto triac? How long is the trigger pulse?

Ans: they are control by a PIC MCU, the trigger pulse is 1 msec

 

In your previous reply, you mentioned that 2 SCRs need to be turned on at the same time in order to create a current loop but I wonder how could this happend as the L1 SCR would be turned off when it crossed the zero point before I turn on the L2 SCR and the same thing happen to L2 & L3 SCRs.

 

Is this the reason why the motor could be ramped up smoothly when the firing angle of each phase is more than 50% so that the L1 SCR is still on when I turn on the L2 SCR (as you said phase to phase should overlap) and so on?

 

Cheers!

[marke]

Hello schow

 

1. With the opto triacs, ensure that you are using opto triacs that have a very high dv/dt. The standard opto triac will turn ON with low dv/dt and this will cause erratic behaviour. You should be looking at devices with a dv/dt of around 5000V/uS or better. At 500V/uS, the tric and SCRs will be turned ON by another phase being controlled or transients on the supply.
2. Ensure that the gate cathode of each SCR is terminated by a low value resistor to overcome the capacitance of the opto triacs, otherwise you will have dv/dt triggering.
3. If you use a single 1 ms pulse at the time you want the SCR to turn ON, then if there are lo leakage currents or capacitive currents, and if the SCRs are not dv/dt triggered, you can not get current to flow.

   If you trigger on phase 1, and then 60 degrees later, trigger on phase 2, and then 60 degrees later, trigger on phase 3, with 1 ms pulses, the SCR will be held ON for the dureation of the pulse, and then as there is no where for the current to go, it will turn OFF. When the next SCR is triggered ON, there is no where for the current to go so it turns OFF.

   You need to extend the triggering so that there is an overlapp so that current can flow through two SCRs.

Best regards,

Mark.

[schow]

Hi Marke,

 

Thanks for your advice.

 

The situation improved after I increased the duration of the SCR trigger pulses as per your recommendation.

 

The motor can now ramp slowly from station condition up to full voltage quite smoothly but I would not know if the 3-phase output are balance during the entire ramp up process, can we tell by measuring the 3 output phase current? How can we ensure that the current of 3 output phase are evenly distributed?

 

Another problem we are facing now is that after the motor ramped up to full voltage or before reaching the full voltage, we stopped firing the SCRs and fired the SCRs again before the motor was completely stop (something like flying start), the motor some time will jerk, stop and ramp up again but some time it will ramp up again from the speed, which we started to fire the SCRs again (flying start), any idea what's wrong with this?

 

Pls kindly advice, danke!

[marke]

Hello schow

 

we stopped firing the SCRs and fired the SCRs again before the motor was completely stop

The motor will generate a voltage imdeiately after the SCRs are stopped firing and this voltage will decay to zero as the rotor field decays. This can take some time. If you reapply voltage too soon, you have to resynchronise the field in the motor and this can lead to current transients.

The main problem however, is probably related to the firing pulses. It is critical to have the start and the end of the firing pulses correct.

It is also critical to ensure that the dv/dt of the design is able to withstand > 1000/uS without triggereing the devices.

 

Best regards,

Mark.

[marke]

Hello schow

 

How is progress??

 

Best regards,

Mark

[schow]

Hi Marke,

 

The software development is almost done, tested the ramp up, down time (with a small motor) and everthing seems fine and the small motor could be ramp up and down smoothly.

 

The initial torque setting is done as we just need to determine the first SCR firing angle and ramp from there until full voltage.

 

Have one question about 'kick-start', I noticed that some sst in the market have this feature, which full voltage will be applied to the motor for certain period and then ramp up from there based on the initial torque setting. Do you have any idea how this works? Which of the following is the right 'kick-start' procedure:-

 

1) supply full voltage for a fixed time -> ramp down to the initial torque setting -> ramp up to full voltage again OR

2) supply full voltage for a fixed time -> turn off all SCRs for a fixed time -> ramp up to full voltage from intial torque set point

 

Not sure if I need 'kick-start' for my mixer machine (45kW) but would like to have this feature integrated into my sst in case the initial torque is insufficient to start my machine, btw, what is the optimum time for 'kick-start' to supply full voltage to the motor? I found some products with says 200msec up to 2 sec, can the SCRs withstand 7 or 8 times FLC for 2 sec???

 

Now, we are looking at the hardware design of the sst for our 45kW machine. I found some ssts in the market are applying opto-triac to turn on and off the SCRs and some with 3 sets of snubber circuit (built up from some power capacitors and resistors) each for each SCR. From your experience, do we need the snubber circuit for our 45kW mixer machine?

 

You have a nice day!

 

BR,

Schow.