Phase Lead_Inverter Output Waveform / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Phase Lead_Inverter Output Waveform

Move Thread

LAN_403

zzz_julian_zzz

Tue Jun 16 2015, 06:55AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Hi Guys,

Just want to hear your comments regarding my results/adjustment on Phase Lead driver for my QCW (401 khz primary)
310 vdc bus, 140Amp peak, 60us on-time, 50 bps

*yellow - inverter output
*blue green- primary current

On the pics below, I don't see any high voltage spikes except for the "ringing" in between before changing polarity of the inverter's output (in red circle)

I swept the whole inductor range and this is the spot where the spikes are gone, but shows ringing in between transition. I wanna know if these ringing are dangerous for the system? and what causes this? also what would be the solution?

Thanks in advance guys.

Blackcurrant

Tue Jun 16 2015, 12:46PM

Registered Member #2989 Joined: Sun Jul 11 2010, 12:01AM
Location: UK
Posts: 94

this is just a guess
I have seen this type of thing in spice simulations with large IGBTs but do not understand it yet
I would look at the gate drive signal to your switches could it be due to not supplying them with enough current?
If you are using a gate drive transformer you could use higher voltage drive on the primary and use a step down transformer

zzz_julian_zzz

Wed Jun 17 2015, 02:00AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Hi Blackcurrant,

Thanks for your advise, problem was solved. It is due to the DC blocking capacitor in one of my GDT primary circuit that was accidentally removed during debugging, causing a slow ramp of voltage in the 2 IGBT gates - same behavior like an overly damped Gate Resistance / same response if the driver is not sufficient to drive Gate capacitance quickly enough. (BTW, I'm using 2 N30 gdt to drive 4 IGBT - conventional full bridge, 2 IGBT : 1 GDT) It is now corrected (please see photo below) 340vdc, 150Amp peak, 2nd image L5 MRMV

As you can see, the ringing between transition of inverter output were removed, BUT there are visible spikes at the top of the waveforms and whatever inductance I use/sweep ranging from 9uH to 30uH the spikes are still there, Notice that with the overly damped circuit (original issue of my system - see my 1st post above) It doesn't present any spikes at all on the INV-output, I thought I was able to find the sweet spot of the inductor to tune the phase lead, turned out that the IGBT is not turning on fast due to overly damped Gate resistance. With this, I will be trying to increase gate resistance of IGBT from 10 ohms changing to 15 ohms and see what will happen to the spikes.

This is the waveform without Phase Lead inductor L4 NLMV

Current Leads Voltage output of Inverter

Thanks! will be posting the result soon.

EDIT: P.S., Anyone have a comment on why is it that the top waveforms have more pronounced Spikes compared to the bottom waveform in terms of magnitude?

zzz_julian_zzz

Thu Jun 25 2015, 07:36AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

Work continues,

2nd slide(PDF) is the diagram that I am using currently in the bridge. Plus, a 4400 uF capacitor is connected at the 340v rail with about 6 inch wire for both pos and neg from the IGBT (FGA60N65SMD) bridge rail and the cap's terminals. (I didn't get the inductance here - later we'll see why I need)

NOTE: all tests/waveforms were taken while the bridge is running at 340vdc with 160 Amps primary tank current (without the secondary coil in place)

3rd slide (caption with it)
4th slide (shows where I see the problem)
5th slide (shows the problem in zoom mode)

Observation:
There are a bit spikes on the switching inverter output. This cannot be covered with just adjusting the phase lead as I have tried this before, values are from 9 uh to 30 uh (different inductors) - I have LCR meter to measure the uH range value. Getting the advise of Loneocean, he said that this is in fact a shoot through condition that's why I see spikes in the inverter output and cannot be fully removed by just Phase Lead.

I tried using different values of gate resistors(of all igbts) from 7 ohms to 33 ohms (originally 10 ohms) to see whether the spikes will diminished or ring further. With 7 ohms in the bridge, the spikes went even worse. I then tried to use 15 (still ringing), and then 33 ohms (still ringing). BUT in 33 ohms, I am pretty sure that this is not because of cross conduction. Why? because with 33 ohms at the Gate, I can see again those ringing in between transitions. If we look again at the first post of this thread, the ringing between transitions were actually due to excessive dead time (cause was cited above) - which is pretty similar(or near)

NOTE: all the tests on varying Gate resistance value were feeded by a signal generator (I also checked if there are changes when running at diff frequency 100 khz - 500 khz) - nothing changes. And this were tested using much lower voltages (~30vdc) unfortunately, I was not able to take photos of the Vge waveform. To describe, the higher the value, the longer the ramp compared to a lesser value. It takes longer time to reach Vth of IGBT to turn it on, hence "a bit of a dead-time" -of course, it is more lossy due to longer linear period of IGBT.

What I am thinking now, is that the spikes were cause by the inductance presented by the 6 inch wire from filter capacitor. I can't test it now as I am far from the test area, so, I will be posting the results when I got it. I just want to share this so that If any other member also experiences this, at least they'll have a guide :)

]phase_lead_debug.pdf[/file]

BigBad

Fri Jun 26 2015, 01:09AM

Registered Member #2529 Joined: Thu Dec 10 2009, 02:43AM
Location:
Posts: 600

I'm not absolutely sure without seeing your circuit and layout, but those kinds of spikes are often caused by the combination of stray inductances around your switch device and the capacitance within it.

What frequency is it? If it's around 20 Mhz or so, then it's almost certain.

The gate resistance can help a little, but you need to mostly solve it in other ways.

In my circuit a resistor on the earthy side of the IGBT to the driver circuit can help in conjunction with gate resistance; it helps to damp out the resonance by reducing the Q.

The reason it's better on the low side may be because there's clamp diodes hidden somewhere in the drive circuit. However, the clamp diodes speed will likely be marginal, and things can easily blow up if the voltage goes too low; drive circuits are particularly sensitive to undershoot. The high side is less sensitive, and there's not going to be clamp diodes limiting that.

But as I say, could well be wrong, haven't seen your circuit.

zzz_julian_zzz

Fri Jun 26 2015, 03:42AM

Registered Member #3964 Joined: Thu Jun 23 2011, 03:23AM
Location: Valenzuela City
Posts: 332

BigBad wrote ...

I'm not absolutely sure without seeing your circuit and layout, but those kinds of spikes are often caused by the combination of stray inductances around your switch device and the capacitance within it.

What frequency is it? If it's around 20 Mhz or so, then it's almost certain.

The gate resistance can help a little, but you need to mostly solve it in other ways.

In my circuit a resistor on the earthy side of the IGBT to the driver circuit can help in conjunction with gate resistance; it helps to damp out the resonance by reducing the Q.

The reason it's better on the low side may be because there's clamp diodes hidden somewhere in the drive circuit. However, the clamp diodes speed will likely be marginal, and things can easily blow up if the voltage goes too low; drive circuits are particularly sensitive to undershoot. The high side is less sensitive, and there's not going to be clamp diodes limiting that.

But as I say, could well be wrong, haven't seen your circuit.

Hi Bigbad,

Thanks for your reply. Here are my comments:

Yup, will be posting my bridge layout as soon as I get back to the Test Area.

Yes, I believe this ringing is caused by stray LC in the layout. Actually that is the reason why I need to check the L of "6 inch transmission line" from terminal cap to the IGBT rails. I needed the transmission line because the capacitor is not really intended to be there since the bridge is for QCW and the test setup is for Phase Lead test only.

The ring frequency last time I check is at around a couple of Mhz region. I didn't record the number, because obviously that time I am not thinking the possibility of that until I post another one (I'm away from test area for couple of days so I went on some thinkering :) )

Yeah, I tried to vary the gate resistance but I find it is not the correct solution.

"In my circuit a resistor on the earthy side of the IGBT to the driver circuit can help in conjunction with gate resistance; it helps to damp out the resonance by reducing the Q."

I didn't get this very clearly, you mean to say you have a resistor that is coming from the Bridge Negative rail going to the UD (driver) to reduce the ringing problem?

"The reason it's better on the low side may be because there's clamp diodes hidden somewhere in the drive circuit. However, the clamp diodes speed will likely be marginal, and things can easily blow up if the voltage goes too low; drive circuits are particularly sensitive to undershoot. The high side is less sensitive, and there's not going to be clamp diodes limiting that"

I have no idea on this one... could you elaborate a bit more?

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