Parallel Full Bridge (8 switches) / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Parallel Full Bridge (8 switches)

1 2 3 next

Move Thread

LAN_403

zzz_julian_zzz

Mon Dec 22 2014, 03:11AM

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

Hi Everyone,

I'm finished building my new QCW DRSSTC bridge. This new parallel bridge was created after I blew my last bridge consisting of 4 IGBT (FGH40N60SMD). Per my records, the last bridge was only capable of handling 150 Amps @ 10ms up +2ms down ramp creating a meter long QCW sword spark (could be longer if branching is minimized/or absent) Fres of my coil is around 300 khz (maybe lower when loaded due to streamers)

CONFIG:
Parallel bridge is consist of 8 FGH40N60SMD IGBTs, estimated to handle not more than 270 Amps max @ 10ms up + 2ms down ramp

SET UP:
When I first tried this parallel bridge, I was getting about HALF of the expected output spark. This was concluded because I changed the specification of the primary circuit by decreasing the impedance of the tank to about half (e.i. old = 60 Ohms, new = 30 ohms) to my assumption that it will double the primary current & hopefully double or at least increase the spark output, BUT, the primary current is still playing at 130 Amp & spark output is still the same (a meter long). My first response is to still lower the impedance, I even went low to a 4 turn primary coil (tank impedance dropped to about 20 ohms) which I think is already extremely low for a QCW DRSSTC system.

I noticed that I'm having problem when I switched back to my original (old) full bridge configuration consisting of only 4 IGBTs. With this old bridge (4 IGBTs), even at low voltage, it easily peaked the current (reaching very high current of about 100 amps only @ 4 ms), I am viewing this primary current waveform via a 100 ohm burdened 100:1 CT & a oscilloscope. Hence, I am convinced that there is something wrong with the parallel full bridge (8 IGBTs).

I suspect that the culprit is in the GDT. As you can see in the images in this thread, I am using only 1 GDT to drive 8 IGBTs, Maybe the GDT cannot drive 8 gates. All of other QCW coils i've seen on the internet have 2 igbts to drive parallel bridges.

Now, I am winding 2 GDTs to drive 8 IGBTs equally (4 gates per GDT).

Question is, I have only 1 power output from my UD, (recent UD versions have 2 outputs for 2 GDTs) can I simply parallel my GDT primaries to get the GDTs working ok?

Any answers/comments in this topic is highly appreciated, thank you in advance.

Sulaiman

Mon Dec 22 2014, 11:35AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Maybe the GDT driver is too 'weak' to hard drive 15 nF of combined gate capacitance?

Goodchild

Mon Dec 22 2014, 12:28PM

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

I'm not sure why you came to the conclusion that it was the GDT. If the GDT was not driving that gates hard enough you would have other problems like exsesive heating or destruction of the IGBTs. You are likely Q limited on your resonator ie your losses match the resonant rise in the primary & secondary circuit.

Options to try:
Higher bus voltage
lower coupling
secondary MMC to keep fres shift to a minimum and Q Max

zzz_julian_zzz

Mon Dec 22 2014, 04:47PM

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

Thanks for your comments.

To Eric,

V input is same. Coupling is same. If sec capacitance is low, then why is that when I tried to use single Full bridge the primary current easily went up(not changing anything else)? so sec parameter is i think not the problem in this case. The problem is only in the bridge. Again, will the 2 GDTs work if their Primaries are connected in Parallel? Thanks in advance.

loneoceans

Mon Dec 22 2014, 08:06PM

Registered Member #4098 Joined: Fri Sept 16 2011, 09:26PM
Location:
Posts: 236

Answering your question, primaries in parallel should work fine; make sure you get them phased correctly. That said, I don't see why one GDT won't work even with one GDT output from your UD (one output is good enough to drive a full bridge of CM300s which is harder to drive than the 8 transistors you have.) If your drive was insufficient, your IGBTs would desaturate and get very hot and sad quickly.

Other than the parameters suggested by goodchild, I noticed you have the IGBTs have their gates wired with separate resistors. If I'm not mistaken, the proper way to do this is with a common shared resistor with individual tuned series resistors to make sure the switching waveforms are matched. If you're using cheap 5% resistors they could be up to 10% different from each other. I'd scope the gate waveforms as well to make sure I get the switching I expect. I use 0.5% individual gate resistors for my 8 bridge and they seem to work fine.

I think the reason why people (including myself) use 2 GDTs is probably because it's easier to find small ferrite cores. Well, at least that's the reason why I used two GDTs since I have a bunch of small cores around and I can't fit so many wires around a small core. :) But I might use a single bigger core the next time.

zzz_julian_zzz

Tue Dec 23 2014, 04:14PM

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

HI loneoceans,

thank you. I presume all the percentage (%) mentioned in your post refers to tolerance?

What do you mean by common shared resistor with individual tuned series resistor?

Can I also use 1% tolerance resistor @ 0.25 watts?, 0.5% is not available in my country?

thanks for the help guys..

Dr. Dark Current

Tue Dec 23 2014, 05:41PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

I think the gate resistor tolerance plays little role here. The difference in switching times is your main enemy. The best way would be to use 2 primaries (can be wound bifilar) with their tank caps, or small matching inductors in series with each half bridge (small air-cored inductors should do the trick).

If each two paralelled transistors are very close to each other and interconnected by wires as short as possible, then I'd just connect them directly in parallel (done this in a half-wave CWDRSSTC, gates directly in parallel with a single gate resistor, worked all right).

Steve Ward

Tue Dec 23 2014, 11:05PM

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

Make sure the coil is operating in the same mode in all tests. If you tuned to operate at the upper pole frequency, the tuning will be awful for the lower pole frequency performance. So just check that the operating frequency is what you expect in all cases, it may be that your different bridges have different delay times causing the driver to lock to a different mode.

Are you using a driver with phase-lead? Have you checked the IGBT switching vs the load current? QCW can be more sensitive to ZCS because the frequency is high and the switch time is significant. Any delay in switch time results in lower power factor from the bridge, so you are getting less watts per ampere.

teravolt

Wed Dec 24 2014, 03:26AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

hi julian. perhaps your transformer cores are saturating what does the rise time of your gates look like?

zzz_julian_zzz

Thu Dec 25 2014, 03:30AM

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

I think the gate resistor tolerance plays little role here. The difference in switching times is your main enemy. The best way would be to use 2 primaries (can be wound bifilar) with their tank caps, or small matching inductors in series with each half bridge (small air-cored inductors should do the trick).

If each two paralelled transistors are very close to each other and interconnected by wires as short as possible, then I'd just connect them directly in parallel (done this in a half-wave CWDRSSTC, gates directly in parallel with a single gate resistor, worked all right).

hi Dr. DC, I think too that the Gate resistance is not the culprit here. at the moment, i want to stick with the parallel switches, once I perfected working with them, then I'll try the other driving methods like in one of your recommendation about 2 primaries. Thanks again!

Make sure the coil is operating in the same mode in all tests. If you tuned to operate at the upper pole frequency, the tuning will be awful for the lower pole frequency performance. So just check that the operating frequency is what you expect in all cases, it may be that your different bridges have different delay times causing the driver to lock to a different mode.

Are you using a driver with phase-lead? Have you checked the IGBT switching vs the load current? QCW can be more sensitive to ZCS because the frequency is high and the switch time is significant. Any delay in switch time results in lower power factor from the bridge, so you are getting less watts per ampere.

Hi Steve,

unfortunately, I don't have DSO to measure /record frequency shift or F run. about the bridge's delay times, will it really differ much if the same IGBT were being paralleled from just singular? all i use is FGH40N60SMD, so switch times are same i think? and yes, I do use Phase lead driver. Thanks!

hi julian. perhaps your transformer cores are saturating what does the rise time of your gates look like?

Hi teravolt,
now that you mentioned it, I noticed something from my GDT output. this waveform can be seen on both of my GDTs ( i already implemented dual GDT) The total 10 ms burst drops from 24v to 10 volts. I already added some capacitance 4700uf(adjacent to the power stage mosfets) on the 24v rails - which makes the drive beefier, now it drops to only 15v. But how do I really address this? I have complete filtering on my GDT driver, DC input, regulator's in & out.

Is it on the blocking capacitor? does anybody experienced this also?

1 2 3 next

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