Gate Voltage / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Gate Voltage

1 2 next

Move Thread

LAN_403

Graham Armitage

Sat Jan 10 2015, 03:39AM

Registered Member #6038 Joined: Mon Aug 06 2012, 11:31AM
Location: Salado, TX
Posts: 248

This is the voltage pattern at the two gates of the IGBT. No ringing which is good, but a little worried by the rise time. 2uS/div and 20v/Div on the display. Running UD2.5 driver. This is driving a CM300DY-24H IGBT but I still fee like it should be a little faster. No bus voltage applied. H-Bridge design attached. Interested to hear anyone's feedback on this and if it's acceptable?

Kolas

Sat Jan 10 2015, 09:55AM

Registered Member #102 Joined: Thu Feb 09 2006, 08:15PM
Location: Raleigh, NC
Posts: 169

As far as I can tell, the driving of the gate to the maximum voltage is to insure complete saturation, and the datasheet says that the VGE(th) is 7.5v max. It appears to attain that voltage far faster then your switching frequency. so without any other knowledge then that, I'd be willing to believe that it is acceptable.

Kizmo

Sat Jan 10 2015, 12:51PM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

Its easily fast enough. General rule of thumb is that you want to switch as slowly as possible. I had CM600 coil which ran at around 3000A and gate rise/fall times were little bit 1Âµs at 41kHz. No problems.

Graham Armitage

Sat Jan 10 2015, 03:27PM

Registered Member #6038 Joined: Mon Aug 06 2012, 11:31AM
Location: Salado, TX
Posts: 248

Thanks all - glad to hear this looks ok.

Graham Armitage

Sun Jan 11 2015, 03:39PM

Registered Member #6038 Joined: Mon Aug 06 2012, 11:31AM
Location: Salado, TX
Posts: 248

Ok, maybe it's been a long night huddled over the coil, but this is not making any sense. Trying to tune for ZVS and getting some plain weird results as shown in the image. I have one probe on one of the gates and the other at the E1/C2 bridge output. The square wave is the gate signal. What is confusing is that the frequencies of the two signals don't match. I can shift the phase with the variable inductor, but the rising and falling edges of both signals should be in sync. Any ideas what's going on here?

Well, it seems my scope may be the problem. with no grounds connected it's picking up some strange freq. By turning up the bus voltage to around 40v I was able to eliminate that and now see two signals in sync, but slightly out of phase. What is still odd is that the one gate voltage is 1/2 the other one. When I disconnect the GDT the outputs from the UD2.5 are both +-24v. Could this be an IGBT issue?

Another question - when setting ZVS phase, which gate should one probe, or does it matter?

Mads Barnkob

Sun Jan 11 2015, 06:59PM

Registered Member #1403 Joined: Tue Mar 18 2008, 06:05PM
Location: Denmark, Odense C
Posts: 1968

Graham Armitage wrote ...

Well, it seems my scope may be the problem. with no grounds connected it's picking up some strange freq. By turning up the bus voltage to around 40v I was able to eliminate that and now see two signals in sync, but slightly out of phase. What is still odd is that the one gate voltage is 1/2 the other one. When I disconnect the GDT the outputs from the UD2.5 are both +-24v. Could this be an IGBT issue?

Another question - when setting ZVS phase, which gate should one probe, or does it matter?

It is the inverter voltage output versus the primary current you have to sync in order to obtain ZCS.

Graham Armitage

Sun Jan 11 2015, 07:27PM

Registered Member #6038 Joined: Mon Aug 06 2012, 11:31AM
Location: Salado, TX
Posts: 248

Mads Barnkob wrote ...

It is the inverter voltage output versus the primary current you have to sync in order to obtain ZCS.

I understand that we want to switch the transistors when the current is zero, but there is a phase difference between the voltage and the current in the primary circuit correct? So where should those current measurements be taken? I see some people scoping bus voltage and gate voltage.

I have been looking for a simple test procedure to figure this out.

Too complicate things, I have a 100MHz analog scope and taking measurements on the primary circuit without ground connections is causing very noisy signals so I am not sure what I am seeing is an accurate representation of what's really happening. Any suggestions on reducing this? I am testing with an isolated DC power supply - 30-40V

Steve Ward

Sun Jan 11 2015, 07:54PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

If you are on isolated power, you can connect the scope ground clip to the -DC bus as your reference.

Your remark about gate voltages changing sounds like you are just measuring the gate voltage plus the source voltage on the top IGBT, so of course it looks funny. You cannot easily scope a high-side and low-side gate simultaneously, under power.

My suggestion is to look at the switch node (connection between top/bottom IGBT) along with the low-side gate, with the scope clips on the -DC bus as reference (preferably as close to the IGBT terminal as possible). Ideal switching conditions will give minimal ringing on both Vge at turn on, and Vce in both directions. I usually observe the Vce (switch node) vs primary current to tune the phase lead inductance for optimal switching. I find optimal timing to be such that the IGBT begins turning off before the primary current has hit zero current, perhaps some 10-20 degrees ahead of that. Going too far from ZCS will result in large overshoot as you are "hard switching", and so a voltage is produced across the switch inductance proportional to dI/dt. Switching more than Ic-nominal should be avoided.

Its extremely helpful to have a current transformer to scope out as well. Watch out for induced voltages on home-made CTs, keep all external wiring loops to a minimum by using co-ax or twisted pair techniques. A simple wire loop near a high-current RF circuit will experience significant EMF that adds in with your desired signals.

Graham Armitage

Sun Jan 11 2015, 08:32PM

Registered Member #6038 Joined: Mon Aug 06 2012, 11:31AM
Location: Salado, TX
Posts: 248

Steve, thank you very much. This is super helpful. First off, I referenced the scope clips to the neg DC bus and that solved the noise problem. As you suggested I connected the one probe to the E1/C2 connection and the other probe to the G2 terminal. The resulting wave form is interesting. The switching on the falling edge looks perfect with minimal ringing on the primary. On the rising edge is where things seem to go wrong. It looks like the gate is turning off way to soon and we get a lot of ringing. As you can see from the first image on this thread the Vge pattern looks great without any bus voltage. What would cause the one side to behave different from the other?

Kolas

Mon Jan 12 2015, 03:44PM

Registered Member #102 Joined: Thu Feb 09 2006, 08:15PM
Location: Raleigh, NC
Posts: 169

I personally cannot fathom why that waveform looks the way it does, but I can be reasonably sure that you DO NOT want to operate it with much power when it's doing that right there. Shoot through is a killer. If i was seeing that, i might look at the length (or difference in) your GDT wiring from igbt to igbt.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob