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?

previous 1 2 3 4 next

Move Thread

LAN_403

Dr. Dark Current

Thu Oct 04 2012, 09:44PM

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

Patrick, if you want to calculate the dissipation in the transistors then you need to measure the output current waveform of the bridge and then calculate its RMS value. P=I^2*Rdson. Increase of Rdson with temperature must be taken into account, with ~100-120Â°C die temperature, the Rdson is usually as high as 200% of the original value.

Just as an extreme example, take the bridge driving an inductor (a leakage-limited transformer into a short-circuit, for example), if you measured the average input current to the bridge it would be close to 0 (actually covering just the losses) but the bridge output current would be large.

Alex M

Thu Oct 18 2012, 01:02AM

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

Back now finally, college got in the way.

I ended up building another one based around a large heatsink I had.

I have got mixed results from it. On one hand I seem to require less power for the same size arc (although this is a plasma speaker so arc length is not too important beyond an inch or two).

Thing is that heatsink still gets pretty warm after a few minutes

The only thing that seemed to help was putting a capacitor and resistor in parallel with the primary coil. I used a 270nF and a 560 ohm resistor, not sure if that is too big or too small but it also seemed to make the audio in the arc slightly clearer.

I also tried putting a DC blocking capacitor in series (1uF) with the primary coil but that stops any output from the transformer. When I viewed the gate to source waveforms on my scope they appeared to be slower than when I had it made on my breadboard! odd...

Gate drive transformer is type 77 materiel from easternvoltageresearch and I have 3x 13 turns trifilar using bell wire.

Driver runs on 15v and the bridge is 24-36v depending on how much power I want to give it, ill add a few more turns to the primary coil at some point so I can use 36v and less current for the same power (again I seem to be getting bigger arcs at lower powers than my last setup)

Also can anyone give me any tips/pointers as to the layout of my power section? I mean should I add more lengths of bare copper and add lots of solder or will that increase stray inductance? (I ran out of solder so that is why the joints around the FET's are poor).

Should I add a capacitor directly across the top FET's drain and bottom FET's source pin (Power and ground)? As my main 36v rail cap is quite far away since I ran out of space but I can easily add an MKP cap on the underside of the board.

Primary coil connects here via a 10cm long tightly twisted pair.

Finally, should I connect my power and driver grounds together? As I had an idea of connecting the LOPTs/flybacks return pin to the power section ground along with the heatsink and actually using the heatsink as a return electrode.

Its probably a stupid idea I know but it seems a bit safer having at-least one end of the HV connected to circuit ground. The flyback transformer is actually an AC one now, well it started life as a DC one but the internal diodes have failed from over voltage and I can measure the internal secondary winding resistance via my multimeter in both directions.

Thanks again.

Shrad

Thu Oct 18 2012, 07:03AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

you should have a single ground plane, a central power ground point where all and every power ground joins, and then you have to connect the power ground point to the ground plane, close to the filtering caps

Alex M

Tue Oct 23 2012, 03:54PM

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

@Shrad Don't you need a PCB to do that?

I have remade my power sections layout and taken more waveforms.

Is this better? (I know its ugly and there is flux everywhere!).

1351005316 3943 FT145001 Vgs Unloadedjpg

Gate drive waveform with no voltage on the bridge
5v/div 200ns/div

1351005652 3943 FT145001 Vgs 22v 150rloadjpg

Gate drive waveform with 22v on the bridge and a 150R power resistor as a load (that was using some used 9v battery's in series). So I would imagine with an inductive load drawing a few amps it will be even worse.
5v/div 200ns/div

This also appears when there no load on the bridge, just voltage being applied to the bridge. Could this be the MOSFET's miller effect?

1351006484 3943 FT145001 Vds 3x9v 150r Loadjpg

Vds with the 3 9v battery's as the power source, 150r resistor load.
5v/div 10us/div

Now I am probably wrong here, but that looks like it happens at turn off indicating the body diode of the FET is ringing? Again that is using a 150r wire-wound resistor as the load.

Same as above but with a 470nF film capacitor across the upper FET's drain and lower FET's source pins (ie supply rails).
5v/div 10us/div

Does this look ok and what other improvements can I make to the power sections layout? Should I make the tracks even wider? I don't really know what makes for a "good" low inductance layout besides having things as physically close as possible.

Also I realise with an inductive load on the bridge there will be more ringing from the windings leakage inductance, so if I was to use RC snubbers across the DS of each FET what value should the capacitor and resistor be for 36v operation? or will just using an RC snubber across the primary coil be enough?

My frequency is 50khz btw.

Thanks.

Daedronus

Tue Oct 23 2012, 04:22PM

Registered Member #2329 Joined: Tue Sept 01 2009, 08:25AM
Location:
Posts: 370

you can try to slow down the turn on, while keeping the turn off, with R in parallel with a diode.

This will reduce a bit the turn on spike.

IMHO the waveforms look quite good.

Shrad

Tue Oct 23 2012, 07:58PM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

sorry for the ground planes, you can just assume the same concept with grounded zones interconnects I think

wouldn't the wirewound resistor have inductive kickback at turn off? after all it's an inductive resistor...

Alex M

Wed Oct 24 2012, 01:43AM

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

Daedronus wrote ...

you can try to slow down the turn on, while keeping the turn off, with R in parallel with a diode.

This will reduce a bit the turn on spike.

IMHO the waveforms look quite good.

Hmm well I guess I will have to try putting a heavier load on the bridge and checking again, since that was only with a very small load and half dead 9v battery's.

I thought that it was at the FET turn off when the spikes occur, since the voltage across the FET only appears when the FET is turned off.

I got some more waveforms when powering a 12v PC fan drawing 200mA, 36v(ish) on the bridge. That was giving me about 18v when I rectified the bridges output and added a capacitor. Then I put a 12v regulator between the fan and cap.

So that is around 38v on the bridge since my battery's are fully charged drawing a very low current (60-ish milli-amps).

Top FET Vds
20v/div 5us/div

1351042646 3943 FT145001 Bottomfet36vjpeg

Bottom FET Vds
20v/div 5us/div
I also had my probe set to 10x here, but am not sure if I really needed to.

I am curious, is this difference in ringing amplitude becuase of the miller effect or is this also layout related?

Perhaps the top FET is effected by miller capacitance more or some other reason?

@Shrad What do you mean by "grounded zones"?

Thanks.

Dr. Dark Current

Wed Oct 24 2012, 03:04PM

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

The difference in ringing might as well be some interference your scope picks up.

Alex M

Wed Oct 31 2012, 06:31AM

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

Dr. Dark Current wrote ...

The difference in ringing might as well be some interference your scope picks up.

So how would I go about making sure its not just the scope probes picking up noise?

Daedronus

Wed Oct 31 2012, 09:26AM

Registered Member #2329 Joined: Tue Sept 01 2009, 08:25AM
Location:
Posts: 370

Disconnect the ground lead and clamping tip on your probe, and directly touch the probe to what you are probing.

previous 1 2 3 4 next

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