TRIAC gate circuit... Probably missing the point but could do with a little help

Thread Starter


Joined Sep 4, 2010
Hi folks,
I am trying to design a crowbar circuit that will open an MCB to protect equipment from high voltage.
Yes this is work but I am in a bind. All test will be performed using a variac fed via an isolation transformer.

I have a site with a turbine, hydro, driving a SEIG. The capacitors are switched to manage the speed and voltage as this alters the frequency to voltage relationship. There ere 2, well technically 3, loads, heaters and an isolation transformer, with 2 separate secondary's, the output of which is rectified and fed to grid tie inverters.

There is an over voltage relay to disconnect the generator, SEIG, from its caps and the loads.

With just the heaters on all is well but if the inverters are running, and abruptly disconnect, we are seeing a voltage spike big enough to damage the energy meter. I do not know the origin of the spike but I suspect it has something to do with collapsing flux in the transformer.

Whatever the cause of the spike I cant muck about with caps as that would mess with the voltage regulation, the generator being an islanded SEIG, and the secondary voltage is none standard to avoid excessively high DC voltages after the rectifiers.

That is a summery of the issue, assuming I have to mitigate the spike as opposed to preventing it n the first place then ...

I need a way to protect both the primary circuit, to protect the energy meter and the two secondary circuits as these are blowing rectifier fuses when the inverters shut down, presumably because the voltage spikes and the current associated with lifting the voltage of the caps on the DC side.

The energy meter can be protected with a 1A MCB, the rectifiers will require at leas 10A and if I decided to protect the main circuit anywhere it would need to be at least 25A, preferably 32A

I would rather not use MOV's because of the ageing. TVS diodes are a possibility but voltage matching will likely be an issue as will current limiting. I looked as SIDAC's found they were not easily available in the voltages I need and based on that decided a TRIAC would probably do the job well. Or at least I originally thought it would, perhaps not

The problem I have is that running in QI and Q3 I need an IG of 50mA or more and if I do that with a simple divider and a DIAC the power requirement of the resistors it quite high.
I may be missing something obvious but if the diac VBO is 45V any voltage decider could be dropping 210V, at the top and 40V, or so, on the ref side whilst supporting a current in excess of 50 mA when the DIAC is off.
Obviously I can use higher value resistors to limit current in the voltage divider, and thus power, but as this would also limit IG I do not think that is an option, hence the question.

In short how do I drop 250VAC to 45VAC in a way that can provide sufficient current to turn on a TRIAC, Qxx30LH5, without the resistors having to handle 15W between them?
this feels like an asked and answered thing and must be a common design issue so what am I missing?

I am planning to have a current limit resistor that will open the MCB in its MAG region but also limit the pulse current through the TRIAC.
I realise that the IG will only be a short pulse and because the supply voltage will fall as soon as the TRIAC turns on, even before the MCB opens.

I also appreciate that I have probably answered my own question RE power dissipation over the restive divider and the 'normal' off state current needs to be significantly lower than the required IG with a cap providing IG when the DIAC is pushed above VBO...

What I don't know is how to calculate all that to actually decide if the plan is viable specifically with respect to the delay that will be added to the trigger pulse. It is also fair to say that there is likely quite a bit more that I don't know I don't understand yet...
Ignorance is not bliss in this case.

To be honest although I have been talking about RMS voltages here I expect that I need to be thinking in terms of peak voltages and transients for everything but the power dissipation...

My idea may be unworkable but if it is I would like to understand why so I can come up with a workable one...

Any and all comments are welcome and don't feel the need to pull any punches... I can live with feeling silly if I am learning something.

Delta Prime

Joined Nov 15, 2019
Hello there :) are you using a modified sine wave or pure sine wave inverter.
Can you use rcbos instead of MCB's.
crowbar circuit keeps the track of the input voltage and it only acts when it exceeds the limit. When the limit is exceeded, the circuit causes a short circuit across the power lines and the fuse connected made of a low melting point metal, melts breaking the circuit. The value of voltage at which the short circuit happen is depend upon the Zener Voltage. The SCR in the circuit is directly connected across the input voltage and the ground of the circuit. However, this SCR is kept turn off by grounding the gate terminal of the SCR. When the Zener voltage is exceeded, the Zener diode starts conducting and the voltage is applied to the gate terminal of the SCR. The voltage applied at the gate terminal of the SCR makes it conduct and there is short circuit between the input voltage and the ground. This short circuit draws the maximum possible current from the circuit and blows up the fuse isolating the power supply from the load.
This arrange of the circuit saves the components and the circuit itself form overshoot of voltages, or voltage transients by blowing up a sacrificial fuse


Joined Jul 11, 2016
why running the whole bunch together (at/under the same protection node)
has something to do with collapsing flux in the transformer.
? how - if the primary is "shorted" . . . self check ::
returns (if it's the case) (the full-text - can be downloaded)
Last edited:

Thread Starter


Joined Sep 4, 2010
How... I dont know and essentally still dont know after much reading, hence the aproach of dealig with the spike as opposed to trying to prevent it. That said Iwill take what I can get advice wise, thanks.

The inverters are type approved grid connected units so pure si but that is the mains side so not actually relivent.
RCBO's could be used in fact I considdered that as an option with the 'crowbar' connected to earth via a curent limit, essentally letting the earth fault current do the tripping as opposd to a short circuit current.
However this has no clamping effect and is more complex where the circuit is IT on the transformer sccondrys.

I am pretty sure the TRIAC/SCR whatever will do the job, my issue is working out how to trigger it, potentally on two concecutive cycles, without needing a voltage devider that is capable of supplying 50mA whilst dropping some 200V.
Obviously I could build that but the resistors would be disipating some considerable heat in normal conditions, circa 10W!

Thread Starter


Joined Sep 4, 2010
Just to clarify the SEIG is not doable fed...
It is a standard induction motor with a bank of switched capacitors attached to provide the required reactive current.
The system runs just fine islanded. In fact it exports just fine too for the bulk of the time.

The isolation transformer, between the generator output and the rectifier input, is always connected, even when the inverters are not active, and in this state, with the generator feeding directly connected heater loads, an over voltage event, caused by a load disconnecting or even oscillation in the control system, is not leading to a damaging voltage spike.

The problem only seems to be evident when the inverters are exporting and a G59 event occurs.
in this case both rectifiers are affected as is the energy meter which is on the generator circuit so isolated foam the rectifier input by the transformer.

I do not know, and cant think how to test, if the issue is common to a trip cause by transient events on the grid as well as total loss of mains. This may or may not be relevant, as I currently don't understand the mechanisms in play I cant begin to postulate.

Having said all that I would prefer to concentrate on mitigating the spike to protect equipment first, even if after this is done I then go looking for the cause. At leas that way every test will not destroy an expensive energy meter and two sets of FF fuses.

The industry standard solutions seem to be MOV based, where there is a significant ageing problem.
SIDACs combined with MOV's to lower the clamp voltage is also a thing and I had wondered about SIDAC/TVS combination to achieve the same but with less of an aging issue. Obviously some form of current limit would need to be in play to protect the TVS.

But for the power dissipation of the trigger circuit the TRIAC seems like a reasonable plan.

Perhaps I should start with a more direct question...
How small of a capacitor can I use to provide the gate current to turn on a TRIAC, assuming the resistor network charging it is not capable of supplying the current.
I realise this is likely device specific and my potential device is listed above. Post#1.

Unfortunately I do not know which part of the datasheet tells me the minimum time that 50mA, or more, needs to flow, or for that matter how to work out what value of capacitor would be required to sustain that current.

I should probably learn SPICE but the requirements to model SCR's seem complex and as an absolute SPICE beginner I would not trust myself to get it right.


Joined Jul 11, 2016
The capacitors are switched to manage the speed and voltage as this alters the frequency to voltage relationship.
since there is no one else reacting yet

i don't understand this sentence = i don't "power electronics" (i only know what the crowbar does with single phase)

call your local retired electrical engineer ask how "it" 's done (obviously for a reasonable consultation price)
// "it" - is quoted coz i quite can't imagine your exact grid layout
: i re-checked me again coz i'm a programmer (= code fast / debug later . . . the reason is practical/obvious -- you can't test the system that does not exist ... otherwise irrelevant in the context of this thread)
so - my "primary shorting" option seems to cut the current at the 2-ry (as i thought and as it probably !does!) . . . if it's floating ( ! no Earth ! ) /!\
1597698057323.png revised simulation
e-Fuse - Dev - sm-10.png //// dropped the original idea to simulate a fuse coz it's too slow anyway (wanted to check the transformer part only)
some links i sought out for fuse simulation . . . perhaps they are interesting reminder/read anyway
it likely doesnot reach up to your case (but anyway)
Last edited:

Thread Starter


Joined Sep 4, 2010
Not that it matters much but a SEIG (Self Excited Induction Generator) is essentally just an induction motor with capacitors providing the reactive current, essentally the field current.
The mathmatics behind induction motors/generators is complex, and not something I understand well to be fair, but in very basic terms the differenc in frequency, which translates to a slip angle, determines, speed, torque and impedence, thus real power.
When used as a generator, the voltage that a motor will produce, at any given speed, is govened by the capacitance attached, so hanging the amount of capacitence changes the voltage, well attempt to.
In practice this will usually create more load, requireing more torque, which isnt available so the speed will fall and the voltage with it.
Within reasonable limits things ballance out and the power remains constant, not witstanding losses and turbine performance.
For a resistive load altering capacitence alters both voltage and speed. If an active load that maintains the voltage is applied then only the speed changes.

Obviously if you try to run way to slowly or quickly the motor starts to do odd things electrically and magnetically and the turbine efficiancy starts to be a big issue... The upshot it just dosnt work or go's way over voltage and breaks something.

Applying a load, or changing the load will afect the voltage even if th prime over compensates to maintain the speed but as I said earlier the primerry relationship coverning voltage is speed, for a given capacitence. Small changes are required to trim the voltage to significantly different load conditions but big changes are required to run at significantly different speeds.

I hope that help you... This is not what I am asking about but I am very happy to give a bit back.
If somone want to try and explain the maths and the air gap thing to me I would be very greatful but it isnt gong to fix my issue.

I either have to find the cause of the spike and eliminate it or eliminate the spike without using a component that will age and die... any and all suggestions more than welcome.

BTW I/We have several of these [ SEIG > Rectifier > Inverter > Grid ] setups.
Only the site organised as [ SEIG > Transformer > Rectifier > Inverter > Grid ] is causing me a problem, and then only when the inverter shuts down unexpectedly.

If you are looking at this and finding the whole generator thing confusing but know what you are going with TRIAC circuits please ignore the generator and feel free to help with the triac gate pulse question.

Last edited: