[cnc2]

Hi all,

I have built a 400Vac 3ph solid-state relay using BTA41-600 triacs as shown in the attached schematic, to drive a 6.6KW motor which has a LRA of 80Amp, but draws around 5Amps while running...

During the first test run the motor started, the current was normal 5A, but when I wanted to switch it off the motor kept running, but the current went up to 7.5Amp, I guessed that only one phase had been switched off, so, I switched off the main braker to stop the motor then switched it back on the motor wouldn't start with a weird grinding sound so I switched it back off and unplugged everything. After inspection I found out that two out of three triacs were shorted, and the "golden_rules" app note said that it was a result of a too high current rise rate dI/dt.

Well I am lost between the app notes, some said that a 10n 300ohm snubber resulted in 400A/µs, knowing that these triacs are rated for 50A/µs I am not even sure that the snubber killed the triacs instead of protecting them.

Should I add inductors in series with the load as suggested in the "ten_golden_rules" by winding each phase supply wire or is it the snubber that needs redesign ?

Please help, I am too green at all this & definitely not an EE, all data around there is for 240Vac and too many formulas that seem to be conflicting and I can't locate any Transil diodes rated above 400Vac. don't wanna fry other triacs.

Thanks !
cnc2.

[cnc2]

Up !

[KOC62]

How can you tell if it is the snubber?
What is your PWM signal for?

[cnc2]

Hi KOC62, sorry for the very late reply.

I'm not sure it's the snubber, but it might be the cause, because a bad snubber can create very high dI/dT current rise rates that will fry the triacs which will become shorted...(from the ten golden rules). and two out of three triacs have become shorted after the first try, the third one which came online after the motor started moving survived the test.

The PWM signal is supposed(no oscilloscope) to be increasing from 0 to 100% duty cycle, so it will act as a soft start for the motor and reduce the LRA which is rated at 80Amp max.

Thanks !
cnc2.

[KOC62]

I haven't dabbled with PWM circuits for 3 phase motors, but thought it odd that your circuit used the same PWM signal to all 3 phases. I thought that each phase had its own PWM to create a 3 phase control, hence my question.
The snubber is a passive device and doesn't increase the di/dt, it shunts the voltage spike created by a dv/dt due to current being cut suddenly, as I understand it. Designing a controller without proper test equipment is asking for trouble as you can't see the possible damaging signals. Even a circuit board layout can create extra capacitance that could assist spike generations,. You are then left to guess - and I never did have a crystal ball that worked for me. Using a known working circuit is a start, as you are relying on a proven design. But scratch building your own without an oscilloscope is a risk for failure, as I see it.

[cnc2]

Hi,

if you think about it, the PWM signal doesn't need to be 3 phase related, it's just like a stupid dimmer/chopper turning on&off to linearly increase the current & try to suppress the 80A current spike of the locked rotor (LRA).

Thanks !
cnc2.

[KOC62]

I did think about it (briefly ) and thought that each triac would turn on at the same time since the PWM is exactly the same for each triac. That part didn't make sense to me.
But, ... I then looked up the MOC3061 and it has a "zero crossing" circuit inside, meaning it waits for that phase to reach zero volts before triggering the "big" triac..Since each phase has it's own zero crossing time the firing is also "3 -phase".
So, as I see it, your controller can only do a "burst fire" mode and not a "phase control" mode. See this link as a concept view for each control mode; https://openenergymonitor.org/emon/mk2/switchdev

[cnc2]

It's not even a controller, so, no phase control. it's supposed to be a very basic soft starter to cut down on LRA current spikes...

[Al_The_Man]

See the application note from fairchild AN-3006
Also ..
Al.

[KOC62]

Have you tried using an 800V device instead of the 600V version? If your using a 400VRMS power supply you get a peak voltage of about 565V, which is not far from the 600V rating of your triac, especially if there is also a voltage spike from the motor being turned off.

[clmenz]

N

Sent from my SAMSUNG-SM-N920A using Tapatalk

[cnc2]

Thanks for the reply Al, I'm happy you showed up

I've gone through too many application notes but missed that one, and I have to say that I'm lost :drowning: at least the one you pointed shows that it is possible to have the snubber off board so I might try higher capacitance values according to the table, with no space constraints. This app note says to use non inductive resistors while another one says that using wire wound resistors (6µH) limits dI/dT current rise rate.

All of them app notes & white papers bring in useful information which add to my confusion, I was about to add in Transil diodes but didn't find any with the needed rating, locally.

A little list of papers I've been through (not that I fully grasped everything):

TRIAC Specifications and Recommendations (Turn‐On Behavior)
AN-3003
AN-3004
AN-3008
AN-437
AN-392
AN-439
AN-8008-D
AN-Golden-Rules
AN-328/0393

Thanks !
cnc2.

[cnc2]

Have you tried using an 800V device instead of the 600V version? If your using a 400VRMS power supply you get a peak voltage of about 565V, which is not far from the 600V rating of your triac, especially if there is also a voltage spike from the motor being turned off.

I used 600V triacs because it was the only option available, but after the fail the local supplier had a new stock of 800V devices so, I now have 800V triacs but a little investigation about the failure mode showed that according to the golden-rules app note, the shorted triacs mean too high dI/dT current rise rate...and two triacs were shorted which correspond to the two leading phases, the third triac seems fine (not shorted) and I suspect that it is ok because during my only test, when I turned off the circuit the motor continued turning but the current went up from ~5A to about 7.5A which means it was running on two phases(the shorted ones) while the third phase got turned off by the still working triac.

So, I didn't try the 800V ones yet.

Thanks !
cnc2.

[KOC62]

Well, if you don't have the right parts then yes, you may be in trouble. Without test equipment to give us the actual current flow, we are only guessing what it might be. I suppose you could put a resistor in series with each triac to limit the di/dt somewhat for experimental purposes. The snubber can only work on dv/dt (voltage), so I don't see how it would help to control the di/dt. Engineering paper designs have their limitations.

[cnc2]

Without test equipment to give us the actual current flow, we are only guessing what it might be. I suppose you could put a resistor in series with each triac to limit the di/dt somewhat for experimental purposes. .

Current flow was measured with a digital clamp meter (AC amp meter) at around 5A of course it doesn't show the LRA nor a graph of dI/dT....good luck finding resistors rated for 5A@400VAC.

The golden rules say that an inductor of few µH in series with the load reduces dI/dT but I'm not quite sure about any side effects which could put more strain on the motor or fry it....Seems to be an easy(ugly) fix when space isn't a problem...I thought about coiling each phase wire according to an online inductor calculator but I really have no idea about the side effects of such a setup (not an EE with very limited electrical knowledge)

From an other paper, the discharge of the snubber capacitor might exceed dI/dT of the triac and be in the range of 1000A/µs if I recall....look at Fairchild's AN-3008 page 7 ....and of course I have no way of measuring the load's(motor) inductance.

Thanks !
cnc2.