Gate driver question / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Gate driver question

Move Thread

LAN_403

Dinges

Tue Feb 02 2010, 11:09PM

Registered Member #2511 Joined: Mon Dec 07 2009, 02:46AM
Location:
Posts: 36

I was reading this article on gate drivers (http://www.powersystemsdesign.com/design_tips_dec06.pdf), where it stated, without adding much explanation as to why, that for a half-bridge inverter two separate gate drive transformers are recommended; not one GDT with two secondaries.

This is the scheme the author recommended:

1265152138 2511 FT0 Gate Driver Illustration

I'm wondering what the reasoning is behind using two separate transformers?

I've used one transformer in my current project and it seems to work ok. Most amateur projects I've seen use one transformer too, sometimes with 4 secondaries even, for a full bridge.

Peter.

Henry H

Wed Feb 03 2010, 04:55AM

Registered Member #2298 Joined: Sat Aug 15 2009, 08:16PM
Location: ex UK, now Santa Cruz, CA
Posts: 35

It does allow you to adjust the dead time (between one transistor turning off and the other turning on) more flexibly than with a single transformer. Each GDT can be smaller than one GDT driving two outputs. Tradeoff is complexity and the need to either have two well matched GDTs, or fine-tune the timing after it's all built.

Avi

Wed Feb 03 2010, 12:26PM

Registered Member #580 Joined: Mon Mar 12 2007, 03:17PM
Location: Melbourne, Australia
Posts: 410

if you put deadtime into a GDT you should get deadtime on all 4 outputs too shouldn't you?

GeordieBoy

Wed Feb 03 2010, 12:36PM

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

Peter, that dual GDT arrangement provides more flexibility in the drive waveforms than using a single GDT. Specifically it allows independent control of the duty ratio for the top and bottom switches which is hard to achieve with a single GDT.

For example, imagine if you wanted to repeatedly turn on the top MOSFET in the bridge-leg for a relatively high duty-ratio, but did not want the bottom device to turn on at all during this time. Two seperate GDTs give you complete flexibility to do things like this when driving motors etc.

Feeding back-to-back drive waveforms for both bridge-leg MOSFETs through a single GDT only really works when the duty ratios for the two devices are close. It is common to see only one GDT used in things like half-bridge DC-DC converters and full-bridge DC-DC converters because the on-times for both devices are usually variable but EQUAL. A single GDT is the cheapest solution in this case.

In this case a single GDT also "reverse biases" the gate of the MOSFET that is meant to be off, just as the other MOSFET's gate is being "forward biased" to turn that device on. This gives excellent immunity to Miller induced turn-on of MOSFETs and IGBTs in bridge applications.

-Richie,

Dr. Dark Current

Wed Feb 03 2010, 04:09PM

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

I see one problem with these dual GDT's - with 50% duty cycle the gate voltage will be e.g. +-15V, but approaching 0% it will be +30V -0V, overvolting the gate. Also quickly changing duty cycles would lead to gdt core saturation.

Dinges

Thu Feb 04 2010, 07:41PM

Registered Member #2511 Joined: Mon Dec 07 2009, 02:46AM
Location:
Posts: 36

Thanks for your comments, all. Can't say I completely understand it though:

Richie wrote ...
In this case a single GDT also "reverse biases" the gate of the MOSFET that is meant to be off, just as the other MOSFET's gate is being "forward biased" to turn that device on. This gives excellent immunity to Miller induced turn-on of MOSFETs and IGBTs in bridge applications.

Yes, but couldn't the same (driving the 'off' FET with a negative gate voltage) be achieved using separate GDTs? It's more a matter of what drive signal you put *into* the GDT than of whether you use one or two GDTs, I think?

Dr.Kilovolt wrote ...
I see one problem with these dual GDT's - with 50% duty cycle the gate voltage will be e.g. +-15V, but approaching 0% it will be +30V -0V, overvolting the gate. Also quickly changing duty cycles would lead to gdt core saturation

Yes, that's what's also shown in the article - but wouldn't that just as much be a problem with dual GDTs as it would be with single ones?

The only reason for separate GDTs for each FET that I could think of was to achieve more/better isolation between the FETs, to prevent signals from working their way back from one FET to the other via the GDT.

Peter.

Dr. Dark Current

Thu Feb 04 2010, 08:54PM

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

Dinges wrote ...

Yes, that's what's also shown in the article - but wouldn't that just as much be a problem with dual GDTs as it would be with single ones?

It won't as long as the ON time for both transistors is the same. Remember that in this case the core transfers a "tri-state" waveform which always has zero dc component, independent of the gate ON voltage and dead time.

Dinges

Thu Feb 04 2010, 11:56PM

Registered Member #2511 Joined: Mon Dec 07 2009, 02:46AM
Location:
Posts: 36

Dr.Kilovolt wrote ...
Remember that in this case the core transfers a "tri-state" waveform which always has zero dc component, independent of the gate ON voltage and dead time.

Yes, that's the waveshape I had in mind: tri-state, equal positive and negative periods, with some deadtime (though all FETs seeing the same deadtime). I should have been more clear about this from the beginning of what I was having in mind:

I will shortly begin construction of the gatedriver board for the next induction heater (3.5kW full-bridge), so hence the question. The gatedriver will be closely based on this circuit, which I've used before with good results:

1265327718 2511 FT83642 Induction Heater Gate Driver

With the difference that there will be two GDTs, each driven by its own BD139/140 full-bridge. From the previous responses I understand though that it would be overkill to use two GDTs. But keep in mind that each FET has 10nF gate capacitance, and there are 8 of them (2 each parallel in a full-bridge configuration), so about 80nF, at a switching frequency of perhaps up to 100kHz.

I like this drive circuit, it's working very well for me and the FETs remain cool. No nasty surprizes so far. Also it's easy to implement user-settable deadtimes (the 1k8 and 100p components at the left). No gatedriver ICs I've seen allow for deadtime or, when they do, have a minimum deadtime of 500ns (and can sometimes be user-set up to 7000ns). I have decided not to go with dedicated gatedriver ICs (TC4420 etc.) because of the bad things I've heard about them. Also because I'd have to order those ICs, whereas the discrete components for the above circuit I have in the parts bin - with plenty of spares too in case magic smoke escapes

.

Anyway, I have trouble believing a gatedriver IC in a DIL-8 case can compete with 4 BD139/140s in a TO139 case, mounted on a heatsink. And after reading some horror stories about these ICs, I think I'll pass on them.

Peter.

Steve Conner

Fri Feb 05 2010, 10:12AM

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

Dinges wrote ...

Anyway, I have trouble believing a gatedriver IC in a DIL-8 case can compete with 4 BD139/140s in a TO139 case, mounted on a heatsink. And after reading some horror stories about these ICs, I think I'll pass on them.

The gate drive IC might not be able to take as much abuse, but it's much faster than discrete transistors and can deliver more current. Look at the switching times on the BD13x datasheet, and compare to the UCC3732x.

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