How to improve gate drive waveform's? / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

How to improve gate drive waveform's?

1 2 3 4 next

Move Thread

LAN_403

Alex M

Fri Sept 28 2012, 06:07PM

Registered Member #3943 Joined: Sun Jun 12 2011, 05:24PM
Location: The Shire, UK
Posts: 552

Well I just recently got a scope and have set about trying to correct the gate drive waveforms on my half-bridge plasma speaker. Below are some waveforms taken with the GDT connected up with their secondary's connected to the gates of some IRFP260 MOSFET's (the scope is connected to one of these MOSFET's G-S).

This is what I was originally getting (forgot to center it).

1348854386 3943 FT0 Halfbridge Witoutuccpng

(5v per division)

Now with a bit of tweaking I have got this

(10v per division)

As you can see they are still not prefect and the rise and fall times (5.us/div) are not very good either (What sort of rise/fall times should I aim for?).

Can anyone tell me what I need to do in order to get squarer waveforms? They look pretty clean when measured straight from the SG3525 chip.

I have tried using damping resistors in series with the GDT primary and that doesn't do much. I got the second waveforms by adding a UCC37321 and UCC37322 buffer stage and also by adding lots of decoupling capacitors onto the board as per the SG3525's datasheet. I also put a 1000uF across the UCC chips power rails (which it shares with the SG3525). The frequency it will be run at is around 50khz.

Any help/tips on improving my gate drive waveforms would be appreciated. Also I am still quite new to oscilloscopes in general so sorry if I have missed anything out.

Thanks.

Dr. Dark Current

Fri Sept 28 2012, 06:12PM

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

The waveforms look OK to me, How is your GDT made?

Experimentonomen

Fri Sept 28 2012, 07:30PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

This may well be imperfections of your scope probes combining with the inductance of the GDT rather than actual rings and sags of the gate voltage.

Patrick

Fri Sept 28 2012, 07:53PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

depending on what you want to do, the rise and fall times look like 500nS, and that looks pretty good to me, if you really want super fast rise and fall Use IC or transistor case type gate driver chips... (i mean instead of GDT)

now noobs and beginners, ive noticed, worry about droop on the negative slope, horizontalish section. If this section is far enough above the threshold then it doesnt mattrer.

a plamsma speaker should work fine with this.

Are you activly having problems? or just worried?

Alex M

Fri Sept 28 2012, 08:22PM

Registered Member #3943 Joined: Sun Jun 12 2011, 05:24PM
Location: The Shire, UK
Posts: 552

Dr. Dark Current wrote ...

The waveforms look OK to me, How is your GDT made?

Thanks,

@Dr. Dark Current
Its made out of Fair-Rite 5977006401 Type 77 Material that EasternVoltageResearch kindly sent me, so I know the core is suitable for 50khz. I have 14 turns per coil on the core wound using tri-filer windings and I am operating this from a 12v lead acid battery. Is this a suitable amount of turns for 12-15v operation?

Using the scope I have made sure each section's waveforms are looking good and have used plenty of decoupling capacitors (which has made a huge difference to the original waveforms, guess I will take that as a lesson

).

Also I am not sure at what point the waveforms become "acceptable". Here are some taken after some more tweaking, probes connected to IRFP260's G-S via a 10 ohm gate resistor and UF4007 fast turn off diode.

There is a 0.82ohm damping resistor in series with the GDT.

1348862075 3943 FT145001 Sg3525 50khz Both Waveforms Gdt 14 Turns Png

That is 10V/div. 5us/div (I must try and work out how to add this data to the images on the scope).

@Experimentonomen I don't think its the scope probes, as I can measure a clean square wave from other sources (including the scopes built in generator). But it could well be the inductance of the GDT.

@Patrick Heating of the MOSFET's has always been the biggest issue.

But as you said I am still a noob with scopes right now so don't really know what a "good" GDT waveform looks like. I just assumed that the squarer it is, the better.

Thanks,

Alex.

Patrick

Fri Sept 28 2012, 08:43PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Alex1M6 wrote ...

@Patrick Heating of the MOSFET's has always been the biggest issue.

But as you said I am still a noob with scopes right now so don't really know what a "good" GDT waveform looks like. I just assumed that the squarer it is, the better.

Thanks,

Alex.

just watch out for the duration that the current and voltage are present at the same time across the junction, then your heating will be minimized for that particular transistor.

Alex M

Fri Sept 28 2012, 10:57PM

Registered Member #3943 Joined: Sun Jun 12 2011, 05:24PM
Location: The Shire, UK
Posts: 552

Patrick wrote ...

just watch out for the duration that the current and voltage are present at the same time across the junction, then your heating will be minimized for that particular transistor.

Could you elaborate a bit more on this? I am not sure I quite get what you are saying.

Do you mean I might have a cross conduction problem (this is a half-bridge I am driving).

Patrick

Sat Sept 29 2012, 12:21AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Alex1M6 wrote ...

Could you elaborate a bit more on this? I am not sure I quite get what you are saying.

Do you mean I might have a cross conduction problem (this is a half-bridge I am driving).

cross conduction is not what i meant...

here, lets step through this lesson logically....

first -- we have two desireable states, and a third undesirable, but neccesary state.

ON, voltage drop from source to drain is near zero, while current is maximum. [ V x I = low heat]
OFF, the voltage is now high from source to drain, and current is of course, infintely small. [ V x I = zero heat]

Transistion, a switch isnt perfect, it takes time to change from one state to another.. . right?
during that transistion the V and I are both high, I x V = P in watts right? therefore the impedance converts the V x I into heat for the duration of that transsistion time...

since transition is where we have both V and I present at the same time [ V x I = super high heat ] , we take great care to reduce the time this high heat state is allowed to exist.

With this knowledge, we can get the most possible out of a small cheap, or a large expensive transistor.

Dr. Dark Current

Sat Sept 29 2012, 07:03AM

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

The MOSFETs have a very slow intrinsic diode, in cases the current is leading the voltage and when switching at higher frequencies, the diode recovery becomes a big issue.

Watch out for transformer core saturation (spikes on the primary current waveform).

Also, some FETs start to de-saturate at higher temperatures with "relatively" low currents. Of course this then results in huge heating.

Steve Conner

Sat Sept 29 2012, 01:16PM

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

Also, bear in mind that the waveforms will change a lot (for the worse :( ) from Miller effect once you apply the main power and put some drain voltage on the FETs. It's under this condition that you want to optimise the waveforms.

1 2 3 4 next

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