DRSSTC waveform troubles / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

DRSSTC waveform troubles

previous 1 2 3 next

Move Thread

LAN_403

Goodchild

Mon Jan 24 2011, 07:35PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

The two things I notice off the top of my head are:

1)You changed the RC time constant for the dead time circuit R7 and C1 the original time constant was set to 440uS your new values have it at 30nS
2)The gate drive chips you are using are a slight bit slower and a lot less powerful than the original gate drive setup. You may not have enough power to drive the GDT and your IGBTs.

I think the best course of action to take would be to start scoping different parts of the board starting with pin 12 of the 74HC14 and compare that to your primary current and check for delay. Work you way down to the gate drive. The delay has to be in your board somewhere because there is delay in your gate drive signal before it even goes threw your IGBTs.

DoS

Mon Jan 24 2011, 07:57PM

Registered Member #360 Joined: Sun Apr 02 2006, 12:13AM
Location:
Posts: 12

Dead time circuit does not account for the time delay as it is used after the burst is over. My time constant is 33us, which I think is more than enough for 270kHz coil (3.7us period).

Did some more scoping. If the GDT secondaries are disconnected, there is negligible delay, but if I connect the secondaries, the voltage on TC4429 outputs first of all rises slowly to almost half the value (~6V) , then goes as fast back to zero and climbs up slowly to 12.
And same for high-low transition (fast from 12 to 6, then back to 12 and slowly to 0).
Sure enough the driver chips might be too weak, will it result in that kind of behaviour? Or can it be the GDT?

Goodchild

Mon Jan 24 2011, 08:33PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

I never did like those gate drive chips, Ive always liked the UCC chips more.

But this makes sense, it's not just the GDT you are driving you have to remember that the GDT is loaded with a capacitive load on the secondary. Those gate drive chips are not made for drive large IGBTs they are designed for driving small to moderate sized MOSFETs. There is a reason Steve designed that board with large MOSFETs driving the GDT.

DoS

Mon Jan 24 2011, 10:33PM

Registered Member #360 Joined: Sun Apr 02 2006, 12:13AM
Location:
Posts: 12

Without the GDT the delay is around 100ns or less (that's normal according to the propagation delays of the preceding components, roughly 20ns for Schmitt trigger, 20ns for AND and 60ns for TC4429).
When I add one IGBT directly to the driver board output, the total delay grows about 15ns and when I add the second one the, it again grows 15ns and the signal is undistorted. So I think these chips should be capable enough or am I missing something?

But when I drive the IGBTs through GDT (half bridge supply disabled and TC primary unconnected) the signal gets really distorted and delayed (>500ns).

Goodchild

Mon Jan 24 2011, 11:13PM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

What IGBTS are you using? I'm going to assume 60N60 or something similar with based on the frequency you are running at.

Here are my two bits, a GDT is probably the best way to run a system like this, what I would recommend is using the original GDT driver circuit on Steve's board those chips are just simple not cut out for driving IGBTs of this size.

axelro

Tue Jan 25 2011, 07:35PM

Registered Member #3640 Joined: Sat Jan 22 2011, 12:16PM
Location: Germany close to Heidelberg
Posts: 39

Or, another guess - the GDT material is not suitable, or the way to wound it isn't right.

Can you provide a picture of your GDT?
What GDT material and windings count prim/sec are you using?

Gregory

Tue Jan 25 2011, 07:57PM

Registered Member #2922 Joined: Sun Jun 13 2010, 12:08AM
Location:
Posts: 226

TC4429, I'm use than and I don't have problems! I think that you are using a bad core, with bad response at high frequencys

DoS

Tue Jan 25 2011, 09:36PM

Registered Member #360 Joined: Sun Apr 02 2006, 12:13AM
Location:
Posts: 12

IGBTs are not 60N60, but IRG4PC40WD, which should be much easier to drive.

The core is 3E25, wiring the CAT5 method (whites parallel, colours series, 9 turns).
Farnell link:

I found some UCC driver chips lying around, but while testing these accidentally burnt the 7408 chip, so my testing will continue tomorrow. Initial results showed that UCCs performed better than the TCs.

Goodchild

Wed Jan 26 2011, 08:40AM

Registered Member #2292 Joined: Fri Aug 14 2009, 05:33PM
Location: The Wild West AKA Arizona
Posts: 795

So I did some comparing of the IXGN60N60C2D1 and the IRG4PC40UD and found some surprising results.

This is the datasheet I found for the IRG4PC40UD:

This is the datasheet I used for the IXGN60N60C2D1:http://ixdev.ixys.com/DataSheet/ DS99177A(IXGN60N60C2-D1).pdf

First off the IRG4PC40UD is no easier to drive than the IXGN60N60C2D1, they both have about the same amount of gate change.
Second the IXGN60N60C2D1 are faster than the IRG4PC40UD by a good margin.

The IRG4PC40UD have a total full cycle switch time of 223nS ( Td(on) + Td(off) + Tr + Tf ) I generally try to keep a <10% guideline in total switch time vs frequency. At 270KHz your IGBTs are spending ~12% of the period switching from low to high and high to low. While this is within reason it may provide unsatisfactory results, particularity if you don't drive the gates of the IGBTs with a good strong gate drive circuit.

It seems clear that your main problem (extra delay) is some were in your gate drive circuitry I would highly recommend using a suitable gate drive circuit adept at handling the peak and average current required when driving IGBTs switching large currents in a DRSSTC. The better your drive the gates of the IGBTs the better your IGBTs will perform when under heavy loads. I still recommend Steve Ward's original gate drive circuit, mainly because it's perfect for the job and has already been designed and tested in this sort of system.

DoS

Fri Jan 28 2011, 04:35PM

Registered Member #360 Joined: Sun Apr 02 2006, 12:13AM
Location:
Posts: 12

I think the UCC chips are doing their job. And although the switching isn't perfect, I got about 18" sparks from the coil at 120BPS and 80us on time and 300A current limit. For bigger coil and full bridge I would certainly build a better driver.

previous 1 2 3 next

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