capacitive commutation of thyristor

[davidhillsgangy]

Good Morning

If you used a N channel MOSFET to switch a capacitor across a Thyristor to force Commutation(switch off) do you connect the capacitor to the source or the drain of the MOSFET.

thanks

David

 

[(*steve*)]

It depends on how you're driving the gate of the mosfet.

Probably safer since we don't know if you put it in the drain. However it actually makes no difference for what you're doing.

I assume its an SCR.

 

[davidhillsgangy]

scr forced Commutation

thanks Steve

the gate is driven from 555 timer with common source.

do you know where I can find an explanation of why you get -2Vc across the SCR
when the capacitors negative terminal is switched through the mosfet to the anode(positive rail) of the SCR.

what's happening to the charge on the cap to cause the reverse voltage across the SCR
I can't see why you get a minus 2 times Vcc across it!

thanks

David

 

[(*steve*)]

You'll have to show us the circuit you're using.

 

[davidhillsgangy]

rough idea

Thanks

here is the rough ides of what I am trying to do

edit:

Moderators note : converted image and uploaded it to the post

 

[(*steve*)]

I'm not 100% sure you have the mosfets exactly right, but let's assume for the sake of argument that the SCR is turned on, the SCR's gate is floating and the other mosfet is off.

The load resistor will have most of the voltage across it and the capacitor will charge to the supply rail.

As the mosfet attached to the capacitor turns on, the supply voltage appears across the SCR (but reversed!). The load sees twice the supply voltage for an instant until it discharges and then recharges the capacitor across the SCR (which is now off).

When the mosfet turns off, the full supply rail now appears (reversed) across the load (and thus twice the supply voltage appears across the SCR) and the resistor from the capacitor/mosfet junction to ground. It (relatively) slowly discharges as the voltage across the SCR falls.

The full supply voltage appears across the SCR when the capacitor is discharged.

So, at various stages, the load *and* the SCR see both twice the supply rail *and* the supply rail reversed.

 

[KrisBlueNZ]

That's a pretty weird-looking circuit. It raises more questions than it answers!

1. Can you add some circuit references to the resistors and semiconductors so we can refer to them by name?
2. Can you add ballpark values for the resistors and the capacitor?
3. What's the voltage on the positive rail?
4. The resistor between the SCR cathode and the 0V rail, is that a load?
5. If so, is it an actual resistor or something else?
6. Does the load need to be connected like that?
7. Are all the FETs N-channel enhancement mode MOSFETs?
8. What are you trying to achieve?
9. Does the circuit work, or is there a problem?

You might want to redraw the diagram using a free EDA program such as LTSpice IV (see http://www.linear.com) (which can also simulate it) or kicad.

 

[(*steve*)]

That's a pretty weird-looking circuit. It raises more questions than it answers!

I think I put it more politely

But yeah, I ignored a lot of it, assumed that the resistor in series with the SCR was the load, and that the mosfet attached to the capacitor would act as a switch.

Whether the circuit works or not will depend a lot on the values for the various components and your ability to switch (on) the mosfet quickly enough.

You did say it was "an idea" so I tried to go with that rather than spend time pointing out the problems. But rest assured, there are some and Kris has done a good job ar enumerating some of them -- or at least asking questions that will allow him to enumerate some more...

 

[davidhillsgangy]

Thanks for reply's
positive rail 12volts
MOSFET are FQP50N06
Load resistor as Steve says, in series with SCR (24watt 12volt bulb)
cap 1uf
other resistor is 1k

circuit came from Rex M.David Power Diode & Thyristor circuits
chapter 7 capacitor turnoff: chopper regulators and series invertors

555 timer running 50% duty square wave at 1 to 400 hz

purpose of exercise to understand forced capacitor turn off with simple practical circuit using available parts

scr is a
Stud Phase Control Thyristor SCR 100A 1600V KP100A

thanks again

 

[KrisBlueNZ]

You didn't answer all my questions so I'm not sure whether this will help.

Connecting a load in the cathode path of an SCR is not normal. Normally, the cathode is connected to the 0V (negative) rail of the circuit, and the load is connected between a positive voltage and the anode. I am going to assume that you can rearrange your circuit into this topology.

I've done a bit of Googling and discovered that normal SCRs cannot be turned off by bringing the gate negative relative to the cathode. A special type, called Gate Turn-Off (GTO) SCRs, can be turned OFF in that way, but your 100A device is not one of these. One of the documents I found said (I think) that GTO SCRs are limited to about 10A current ratings. Since your load only draws 2A continuous, and its cold current isn't likely to exceed 10A, you may be able to use a smaller SCR that can be turned OFF via its gate (i.e. a GTO SCR).

Assuming you stick with a standard SCR, in a DC circuit where there is no automatic commutation due to the supply dropping to zero on each half-cycle, the normal way to turn an SCR off is to force its anode negative relative to its cathode for long enough for it to stop conducting. I think the obvious way would be using a capacitor connected to the anode. The other side of the capacitor would be connected through a resistor to the positive supply, or perhaps a somewhat lower positive voltage. When the SCR is ON, the capacitor would charge up to that voltage (minus the SCR's anode-to-cathode voltage drop).

To turn the SCR OFF, the side of the capacitor that's connected to the resistor to the positive voltage needs to be brought quickly and firmly to 0V. By an N-channel MOSFET, presumably. When that happens, the SCR's anode will be pulled negative. The capacitor needs to be large enough that the SCR's anode will stay negative for long enough for the SCR to turn OFF, while the anode is being pulled positive by the load. If you can guarantee a certain minimum ON-time for the SCR, you may be able to assume that the bulb has warmed up and the load current has dropped to around 2A when you come to turn the SCR OFF.

The resistor from the capacitor to the positive voltage needs to be low enough to charge the capacitor fully (or nearly fully) within the shortest time that the SCR will be ON for.

I hope this helps.

 

[jpanhalt]

Here's a link to an informative text on Thyristors and Triacs. Page 128 discusses Kris's approach with the capacitor to the anode.

http://www.onsemi.com/pub_link/Collateral/HBD855-D.PDF

John

 

[davidhillsgangy]

thanks working on prototype circuit
sort of works.
switch off of scr is a bit haphazard

 

[(*steve*)]

Have you considered switching your load on and off with a mosfet or igbt?

 

[davidhillsgangy]

My sort of long term goal is to build a full H bridge using the 4 stud scr's that I have.
they are rated at 100amps and I already have them.(seem cheap these days compared to equivalent mosfets or IGBT's etc )

I want to build a small AC square wave generator at 400hz for tig welding light aluminium.

you need AC to weld aluminium because a crust forms on weld puddle that acts like a diode. the current reversal cleans this away.

So I was starting out trying to understand some of the basics for controlling forced commutation of a scr using parallel cap method.

so I am using a 555 to drive a couple of mosfet's one to turn the SCR gate on and off the other to switch in and out the commutating cap.

The principle is well established but the practical building of a simple first attempt is a steep learning curve.


Shoot through is almost certainly going to be a problem with the H bridge

thanks for your help

David

 

[mursal]

Thanks Kris, very interesting