Capacitance on Drive side of Gate Drive Transformer affects current counterintuitively. / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Capacitance on Drive side of Gate Drive Transformer affects current counterintuitively.

Move Thread

LAN_403

borderline128

Sat Apr 25 2009, 09:36PM

Registered Member #1531 Joined: Tue Jun 10 2008, 08:37PM
Location:
Posts: 8

Has any one else noticed that the value of the capacitor that is placed in series with two Drive IC's and a GDT has an very strange response. If I use a value below 1uF for this cap I tend to see a dramatic increase in current which quickly causes my drive IC's to overheat. My wave form also becomes highly distorted and useless for driving FETs. If I go bellow 20nf (or so) the current drops back down and the signal adopts a slightly rounded but distinctly binary appearance. The magnitude, however, drops significantly and again is useless for switching FETs. My question:
What is causing this behaviour?

My theory is that these undesirable affects are a product of the capacitor value corresponding with the value needed for resonance at the specific drive frequency and transformer inductance. In my case I measured the inductance of the drive side of my GDT to be about 930 uH.

So far my solution to this observation has been to use capacitance values above 2uF which seems to produce a binary signal with rounded edges (representing a delay in charge and discharge of the gate) with a magnitude high enough to drive FETs. But the delay in switching seems exagerated and I'd like to produce a more square wave from the GDT.

Z28Fistergod

Sun Apr 26 2009, 09:31PM

Registered Member #2040 Joined: Fri Mar 20 2009, 10:13PM
Location: Fairfax VA
Posts: 180

When you measure the inductance of the primary alone you are not actually measuring the "true" impedance. You are measuring primary winding impedance. What you are looking for is the impedance of the primary while it is in the circuit, wich is called driving point impedance. Driving point impedance will vary with the load on the secondary. So if your calculations did not come out correctly as far as capacitance required for resonance, that is why.

It is possible to calculate driving point impedance when using sine wave signals, and I'm sure it's possible with square wave signals too but most likely more complicated. I don't think your actually interested in the driving point impedance but it is still good to know.

Typically DC blocking capacitors are between 1uF and 10uF.

The "delay" is called rise time. Many things can affect rise time. Driver power, driving point impedance, gate capacitance, gate resistance, number of windings, windings ratio, etc. Describe the GDT to me and also your drive circuit.

borderline128

Tue Apr 28 2009, 06:54AM

Registered Member #1531 Joined: Tue Jun 10 2008, 08:37PM
Location:
Posts: 8

Here is the datasheet to the GDT. link I want this project to be as repeatable as possible so I decided to use parts that can be purchased directly from a retailer.

So long as using any value above a certain point is a OK then I'm fine with that. Ill look for a non polarized cap above 4uF.

For the drive circuit Im using two TC1427's which are being controlled by a 3525 pwm. The GDT is already connect to the continuous wave class D amplifier made up of some FDH44N50 power FETs I picked up on ebay. Below shows a picture of my setup. I want to use TC4452's to drive my GDT because they are TO-220 and can be heat sinked. However they are also very expensive ($3.20 from mouser) and, in my experience, very sensitive to anything that might give them a reason to die. Can you see any other grounds I shouldn't use TC4452's?

Thanks a lot,
JJ
mod edit: oversized images! Nice gear though :)

Edit: BTW the waveform you see in the pic is from the TC1427. You can see on th epower supply to the right that the whole circuit is using about 130mA 110 of which are to power the TC1427's. This makes them heat up quite a bit and is the primary reason Im interested in using TO-220 and a heat sink.

Z28Fistergod

Tue Apr 28 2009, 12:22PM

Registered Member #2040 Joined: Fri Mar 20 2009, 10:13PM
Location: Fairfax VA
Posts: 180

Will this device be run CW or pulsed? When running CW I also found those FET drive IC's to heat excessively. I've used the TC4451/2's and I've had no problems with them, I can't remember ever blowing one up. I hate to tell you this but you really should split your ribbon cable and twist your wires together, really all of your signal wires should be twisted.

GeordieBoy

Tue Apr 28 2009, 12:56PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Keep the total DC blocking capacitance large. I would recommend 100nF ceramic in parallel with something between 1uF and 10uF depending on the operating frequency.

It is likely that your undersized DC blocking capacitor is resonating with the magnetising inductance of the GDT. It sounds like either the GDT has a low permeability core or not enough turns on it's primary. Either that, or you are operating at quite a low switching frequency for that GDT. Usually the series resonance of the DC blocking capacitor and the GDT magnetising inductance is well below the swithing frequency so is of little interest.

When operating in pulsed mode it is possible for this Low frequency series resonance to get excited by the transient nature of the drive bursts. This can cause the burst to have a low-frequency decaying oscillatory transient added-in to the wanted "binary" signal. This can largely be tamed by oytting a damping resistor across the DC blocking capacitor.

Remember the DC blocking capacitor's main aim is to stop heavy DC currents flowing due to small net DC voltages appearing across the GDT. Therefore it's value should be large, and it is okay to bypass it with a damping resistor in the range of 10 to 100 ohms too.

I hope this helps,

-Richie,

Steve Conner

Tue Apr 28 2009, 09:43PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

On my last few SSTCs, I used 4x 1uF plastic film caps in parallel for the DC block cap. With CW drive of four IRFP460s at 160kHz, the caps got warm, and the gate drive ICs got almost too hot to touch.

I also used a nice 470uF low ESR electrolytic across the gate drive chips' supply rails.

I like to use the 10 ohm damping resistor across the cap in DRSSTCs. I've had my driver circuit malfunction, latch up and burn this resistor out, though.

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