SCR firing in an inverter

Thread Starter

recca02

Joined Apr 2, 2007
1,214
I am supposed to give a presentation on VFD
so i thought of having a Load commutated inverter circuit simulated.
so i first tried to make an inverter circuit using matlab simulink
the problem that i am encountering is when i turn on a SCR it is supposed to remain on until it gets reverse biased (right?)
so how do i turn it off in case of an inverter which has constant polarity DC at its input?

i gave firing using a pulse generating circuit which had an on for 10 % of cycle and i kept the firing pulse period = o/p voltage frequency?
but i only got the desired o/p for first cycle(only monitored one phase )
after which the SCR remained on, any ideas on what i might be doing wrong?
did i set the firing times incorrectly?(i'll post all details if required -amplitude,Ton
circuit i designed and waveform that i obtained )
 

Attachments

pebe

Joined Oct 11, 2004
626
If you are going to give a presentation, take a tip.
Red lines on a blue background looks awful! Stick to a white background.
 

Thread Starter

recca02

Joined Apr 2, 2007
1,214
oh they are not the ones i'll use for presentation
they were taken from another presentation though ;)
anyways thanks for the tip will keep that in mind.
 

beenthere

Joined Apr 20, 2004
15,819
Once triggered, SCR's will remain in conduction under DC conditions (assuming foreward bias). I recall a Coufelt VVP-15 electrofishing controller that used a second SCR to pull the main one out of conduction for pulsed operation. The secondary SCR pulled charge onto a capacitor that in turn dropped the voltage across the conducting SCR and dropped it out of conduction.

This is a very kludgy way to do things, but it worked. I don't have the schematics available to see just how the arrrangement was made.

It would be helpful to see the details of the trigger circuits to see if they are keeping the gate voltage up.
 

Thread Starter

recca02

Joined Apr 2, 2007
1,214
here is my pitiful try,
i will be able to get thru with the presentation since there is a million times better 'premade' circuit with a customized pulse generator to do the trick in the matlab library but i wud like to get mine working as well.
if any other detail is reqd please ask there seem to be many parameters i have no idea how to configure :(.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,210
Recca, have you considered MOSFETS?

There are some mighty beefy ones out there nowadays, with amazingly high current capabilities - even in a TO220 package. On-state resistance in small fractions of Ohms. Gate current ridiculously small.

Kick the tires on some N-channel MOSFETS. Try looking around International Rectifier's site in the IRF7xx series as a starting point.
IRF730; N-Channel 400v, Rds(on) 0.75 Ohm, Id 5.5A, TO-220 MOSFET
IRF3415 has a max Vdss of 150, can do 43A, Rds(on)=0.042 Ohms. Those are some mighty impressive numbers for a TO220 case.
 

beenthere

Joined Apr 20, 2004
15,819
Why didn't I think of that? Also, consider IGBT's (insulated gate bipolar transistors). They don't switch as fast as FET's, but handle more power.
 

Thread Starter

recca02

Joined Apr 2, 2007
1,214
thanks MR Sgtwookie and MR beenthere,
yeah most inverters now use IGBT's and pwm is possible i'll try that too. i'lll post back if i get it working somehow with the the goof up that i made :D
 

SgtWookie

Joined Jul 17, 2007
22,210
200V is about the threshold for IGBT's; below that use MOSFET/HEXFETs.

Power MOSFETs are a snap to parallel, because of the positive temperature coefficient. If one MOSFET is carrying more current, it heats up, which raises it's on-state resistance, which causes it to carry less current. Just about everything else has a negative temp coefficient, which makes it much harder to parallel them efficiently; without careful planning, you'll wind up with runaway thermal conditions or lots of wasted energy.
 

Thread Starter

recca02

Joined Apr 2, 2007
1,214
actually i m studying applications of VFD for power plant auxiliaries so they must handle large power and voltages. right now i m studying about IGBT's .Will try to work out how MOSFETs compare to them for such applications ,thanks.
BTW i need one clarification about the carrier frequency for IGBT's : does the o/p voltage
for an inverter decrease with increase in carrier frequency-(will post back more info if required).


since i would like to simulate the simple circuit using scr (for satisfaction sake) .
here is a quote from wiki(on thyristors) please comment (i need your helps since i have no experience in power electronics-but will gain enough soon)
It should be noted that once avalanche breakdown has occurred, the thyristor continues to conduct, irrespective of the gate voltage, until either: (a) the potential VG is removed or (b) the current through the device (anode−cathode) is less than the holding current specified by the manufacturer. Hence VG can be a voltage pulse, such as the voltage output from a UJT relaxation oscillator.
so if i keep the pulse off after triggering for some time the SCR will turn off?
if so i think all i need to do is some rethinking about the firing pulses.
thanks for all the help thus far.
 

SgtWookie

Joined Jul 17, 2007
22,210
The on-state resistance is one thing - the power dissipation will certainly generate some heat.

What REALLY generates heat is the power dissipation during transition from state to state. The more time the device spends in transition, the more waste heat you'll have to dissipate. Switching a slow device at high frequency = bad news; you'll spend porportionately more and more time in transition, thus decreasing available output current while increasing waste heat. You need to find the "happy medium" or the optimum frequency to switch to maintain proper output while minimizing time spent in transition. That will be device-dependent.

Once the SCR is conducting, it's going to stay that way until the current through the device drops behold threshold. The gate just serves to trigger the device to an ON state, providing there is sufficient potential across it.

If your load is inductive, you might do something like Beenthere suggested in reply #4, or possibly use a UJT through an inductor to "spike" the SCR off.
 

Thread Starter

recca02

Joined Apr 2, 2007
1,214
so no change at gate i/p pulse wud turn the SCR off?

or possibly use a UJT through an inductor to "spike" the SCR off.
where do i give that spike?(forgive my ignorance)
right now all i want to do is turn the scr's at the desired instances so that the bridge starts behaving.
 

SgtWookie

Joined Jul 17, 2007
22,210
Well, if you managed to "spike" it between the SCR's output and the inductive load, you might be able to momentarily cause the potential across the SCR to fall below the conductance threshold.

I haven't worked this out; it's just a SWAG. ;)
 

Thread Starter

recca02

Joined Apr 2, 2007
1,214
ok i get what u are saying will try.thanks.
whenever i see an inverter bridge they only show a simple scr one so i thought it wont be
such a big problem,
i'll post the corrections if i find what i was missing(hopefully not my head)
for now i stick to IGBT's.
 
to turn off the scr, reduce the anode current to zero or below 'threshold level' and reverse the voltage across its terminals. this process is termed as commutation.
DIAC's are uncontrolled devices and and they cant be used in VFD as they require variable frequency control.
SCR's are used in power systems as their power handling capacity is higher than MOSFETS, IGBT, UJT, etc. well most.

SCR can never be turned OFF using a gate signal.

look into commutation circuits.
 

BillB3857

Joined Feb 28, 2009
2,534
For those curious about capacitor commutation , here is an example. This is from an early to mid 1970's vintage control made by General Electric. This is a simplified drawing of only one phase of a 3 phase drive. The Commutation SCRs are of much lower capacity than the inverter SCRs since they only have to handle a short pulse of current. In this application, the main SCRs are of the "hockey puck" design.
 

Attachments

Top