Thyristor used as comparator for short circuit protection?

[cupoftea]

Hi,
Our client is using the P102AA thyristor to give short circuit protection in their AC mains circuit.
Its a trailing edge phase cutting heater regulator.
(The phase cutting is simply done with the switching of the back-to-back NFETs.)

As you can imagine, short cct protection involves promptly turning off these NFETs.

The thyristor, is simply used as a "comparator", with its VGT voltage (0.8V) as the reference.
There is a sense resistor in the mains neutral, and this sense voltage is fed to the thyristor's Gate-cathode terminal.
If a short cct current occurs, the thyristor breaks over and pulls current through an opto which results in a comparator
getting tripped and turning off the back2back NFETs.

Advantage is that its obviously an AC power circuit, and the thyristor can trip whether the current is going one way or t'other
in the sense resistor.

As can be seen from fig13 of the datasheet, the VT of the thyristor can be anywhere from 0.8V to 1.4V from 25degC to 125degC.
Indeed, from batch to batch, this can be different. As such, would you agree, this method of protection is
simply unwise?

..The poor , back2back SiHG70N60 NFETs will simply not be tripped in a timely fashion when a short cct occurs...and they will explode.
Would you agree, what is in fact needed is to use back2back IGBTs instead, and simply use a proper window comparator with +/-
supplies to detect overcurrent in the sense resistor. (obviously its AC mains current in the sense resistor, hence the window comparator with +/- supplies.)

P102AA thyristor datasheet

https://www.farnell.com/datasheets/2308722.pdf

SiHG70N60 FET

https://www.vishay.com/docs/91598/sihg70n60ef.pdf

 

[D.A.(Tony)Stewart]

As can be seen from fig13 of the datasheet, the VT of the thyristor can be anywhere from 0.8V to 1.4V from 25degC to 125degC.

No it doesn't.

That's the output Voltage, VTM vs output current ITM maybe when used as a trigger for a bigger Triac.
The input is VGT, IGT.

Fig. 7 shows that Vgt typical above shows consistent NTC of -3.57 mV/'C or a 357 mV over a 100'C rise with Rgk = 1k

This also shows by mental calculations the gate to trigger impedance rises rapidly with temperature and becomes more sensitive.
The plot above shows @ 25'C crossover Vgt/Igt= 1V / 1mA (?) = 1 kohm, although the incremental impedance from 25 to 35'C = -35mV/-200uA = 175 ohms/10'C but much higher at 80'C .

Yet below,
At 25'C the worst case is 800 mV /200 uA or 4k ohms

The guaranteed specs @ 25'C are 0.8V max { (1) says at 25'C} seems to be inconsistent with the typical plot at 1V.
But other than that I don't see your point yet without a proper tolerance sensitivity analysis compared with the design spec.

I expect the VGT & IGT to slightly increase with load (current after triggering) which is another variable which can be simulated to verify.

 

[cupoftea]

I expect the VGT & IGT to slightly increase with load (current after triggering) which is another variable which can be simulated to verify.

Attachments​

Thanks, as you know, the trigger for this Thyristor is intended to be short cct....so the load current will be enormous, and you kindly tell that the trigger voltage increases with load current.
There is the obvious chicken and egg situation here.....if the trigger voltage goes up with load current, then how did the thyristor "know" what the load current was going to be, in order to set its trigger voltage?......unless it kind of "quickly turns back off again" if a really large load current initially flows...and it kind of then waits for the trigger voltage to get higher before it re-triggers?

BTW, our customer has not included any series gate resistance.

Ayk, a BJT is often used in current clamps, the VBE getting breached, then it turning on...but these are always noted for their wide tolerance and always create very rough threshold levels......a thyristor is basically a npn and pnp BJT, so i would expect it to be equally "rough" in its use in current limiting circuits.

 

[Easy peasy]

For an AC ckt - you can use a CT to get a decent signal - then use this ( rectified ) to trigger the FETs / IGBT's off, e.g. a 2 xtor latch,

correct turn off speed and heaps of suitable protection are needed to keep the semi's alive.

 

[cupoftea]

I reckon a good robust way would be to switch on a Triac in parallel with back to back FETs.....then when the triac is turned on...then switch off the nfets....as the triac is much more robust. ...then switch off the triac.

 

[Easy peasy]

you are of course free to pursue any method you may care to dream up