400Hz operation

[NickStilwell]

I have a problem with an SSR when driving a three phase fan at 115V, 400Hz, max 440Hz. The powerfactor is not very high typically 0.3 . The problem is that occasionally two of the three SSR's feeding the fan do not switch off when commanded to do so via the inbuilt opto. The problem is worse as the supply frequency is increased. Looking at voltages across the SSR the dV/dt is well within limits so something else presumably must be happening. Does anyone ahave any insight into what might be happening ?

Cheers, Nick

 

[waross]

Try three capacitors to correct the power factor. I would start with fairly small capacitots and then increase the frequency to determine the effect.
This should determine whether the problem is related to the poor power factor and provide a cure if it is. If there is no improvement with increased power factor, you will have eliminated one possible cause.
respectfully

 

[Skogsgurra]

Are those SSRs rated for 400 Hz operation?

Inductive loads are tricky and current remains even after voltage has gone through zero. It is a problem at 50/60 Hz - and even worse at 400 Hz. Make sure your specs are right before trying anything else.

Gunnar Englund

http://www.gke.org

 

[NickStilwell]

I will try the capacitor idea out. The relays are Clare PS2401 devices which are specc'd to 500Hz, doesn't mention at what power factor though. Data sheets talks about minimum power factor for guaranteed turn on but turning on is not the problem !

 

[Skogsgurra]

It may be good to include current-limiting resistors in series with the capacitors. Thyristors and triacs do not like switching a capacitor directly - to high a di/dt. Can create local heating around gate electrode if current allowed to rise too quickly. Select resistors that will produce something like 10 or 20 percent of rated load.

Gunnar Englund

http://www.gke.org

 

[itsmoked]

Hello Nick,

Are you using these SSRs to do speed control (PWM) or to just turn the fan on and off?

Are the SSRs being "failed" or do they seem to be happy even after not turning off the fan?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[sreid]

I had this problem one time and fix it by changing to another SSR vendor. The original manufacturer fixed the problem subsequently by changing the internal snubbing network.

 

[NickStilwell]

These SSR's are just for on/off control. They do not fail, they just "fail" to turn off when commanded to do so. The problem appears worse when the fans have just started to spin up. If you get them up to full speed then turn off they generally do.
I finally found another vendor with the same pinout and am investigating that route also.

 

[Warpspeed]

Most of these SSRs use triacs for load switching, and they usually work fine, as long as load power factor is near unity.

However, some SSRs use inverse SCRs and these have much better zero crossing commutation. With problem loads, try fitting the inverse SCR type of SSR, it will almost certainly solve your problems.

 

[NickStilwell]

The part in question, does use dual SCR's.
If the problem is solely dV/dt then the rates I can measure using a scope of approx 1V/us is way down on the manufacturers spec sheets of 1000V/us. I wonder if the 1000V figure is under ideal conditions at say 50Hz and unity power factor and maybe at 400Hz and PF of 0.3 the critical dV/dt rating is much less ? Certainly if I drop the frequency to say 300Hz the incidence of not turning off is much reduced, If I raise the frequency to 500Hz I can induce the fault every time. Obviously it is probably a combination of actors malking the difference, eg power factor changing with frequency, available turn off time reducing at higher frequency, higher di/dt etc. Snubbers are probably the answer and 0.1u + 100ohms has been suggested. Just wish I could do the math to prove it !

 

[itsmoked]

You will always have a problem with that one because you really don't know what the circuitry inside is. There is more than just the triac and an opto isolator as you probably realize.

You just need the PF/snubber correction that barely fixes the problem then improve it by 50% or some other value to assure production success.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Warpspeed]

Try running it with just a pure resistive load and see if the problem remains. dV/dT should not change, but power factor certainly will.

Is there any possibility that the input control to the SSR is picking up noise or has some leakage current in the "off" condition ? Try placing a dead short across the control terminals and see if that helps.

 

[Skogsgurra]

"0.1u + 100ohms has been suggested. Just wish I could do the math to prove it "

Doing the math is fairly simple. The problem is what data you shall plug into the equations.

You need to know a lot of things and most of those things can be found by using a dual (at least) channel scope, a differential probe for voltage across SSR and a DC currnet clamp to observe load current. It may also be useful to have two more channels to observe on/off signals and the 400 Hz mains voltage.

When you have this, you will probably see what is going on - even without the math. If you are really careful, you can use a shunt resistor between load and ground to measure current and a 10:1 or 100:1 probe to measure voltage.

Gunnar Englund

http://www.gke.org