PlasmaSonic II based SSTC, with poor performance / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

PlasmaSonic II based SSTC, with poor performance

first previous 3 4 5 6 next last

Move Thread

LAN_403

HV Enthusiast

Sat Oct 04 2008, 11:27PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Can you post a photo of the core you wound?

It appears you may not have wound it properly (assuming you do have the correct core)

It looks like excessive leakage inductance.

Cotlone

Sat Oct 04 2008, 11:47PM

Registered Member #1383 Joined: Mon Mar 10 2008, 11:09AM
Location: Wollongong, NSW
Posts: 27

Here are my latest windings of the GDTs.

HV Enthusiast

Sun Oct 05 2008, 01:38AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Cotlone wrote ...

Here are my latest windings of the GDTs.

I don't like how the wiring is exiting the core where you have one half the windings on one side, and the other on the opposite side. You want the windings to exit at the same point and then twist together before connecting to the board. With the arrangement you have, you have some seriously large loop area which just leads to more leakage inductance.

Your gate waveforms as measured with the MOSFETs connected should look like the waveforms you measured no load.

Herr Zapp

Sun Oct 05 2008, 02:58AM

Registered Member #480 Joined: Thu Jul 06 2006, 07:08PM
Location: North America
Posts: 644

Dan

I'm not sure I understand your critique of Cotlone's GDT construction.

He's got the core wound with three conductors that are rather loosely twisted together. The windings are fairly equally distributed on the core, and look like they cover about 330 degrees of the core's circumference.

Given the locations of the pads on his PCBA where the windings are connected, what can he change that would reduce leakage inductance?

Regards,
Herr Zapp

Cotlone

Sun Oct 05 2008, 05:24AM

Registered Member #1383 Joined: Mon Mar 10 2008, 11:09AM
Location: Wollongong, NSW
Posts: 27

Do you mean that all six wires should exit the same side of the toroid or do you mean that the two groups of three wires should each exit separate sides and be twisted together?

I've wound a new core, which of these configurations do you mean to be used (if any)?

1.

cheers

HV Enthusiast

Sun Oct 05 2008, 01:34PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Herr Zapp wrote ...

Dan

I'm not sure I understand your critique of Cotlone's GDT construction.

He's got the core wound with three conductors that are rather loosely twisted together. The windings are fairly equally distributed on the core, and look like they cover about 330 degrees of the core's circumference.

Given the locations of the pads on his PCBA where the windings are connected, what can he change that would reduce leakage inductance?

Regards,
Herr Zapp

If i look at his photo, he has the primary winding on one side of the core, and then the other end of the primary exiting and connecting to the board on the opposite of the perhaps at least 1-2 inches away. For this winding alone, you have a big loop area which increases leakage inductance and reduces performance of the transformer.

You really want the winding to connect to the board as close to one another has possible and as close to the device its connecting to. In instances where the transformer isn't close to the device its driving, you would use either a coupled plan on the PCB board, or at least have traces that are on top of one another to maximize coupling.

Remember, inductance is directly related to how much "loop area" you have. Loop area is minimized by twisting wires together, proper routing, using a ground plane, or having outgoing and return traces of a circuit directly on top of one another.

With his transformer photo, i see (3) giant loops for each winding.

Assuming the core is correct, its these loop areas which are killing performance.

Cotlone

Sun Oct 05 2008, 02:32PM

Registered Member #1383 Joined: Mon Mar 10 2008, 11:09AM
Location: Wollongong, NSW
Posts: 27

I attached the GDT with the wires twisted, I raised it off the board. The waveform is still pretty bad, I guess the PCB layout is the problem then, is it? The terminals for the GDT coils are too far apart from each other. Is there any other possible explanation, assuming correct core material?
Thanks for your help,

HV Enthusiast

Sun Oct 05 2008, 04:49PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

You can try reducing number of turns.

Try 10T:10T:10T instead of 15 turns.

Cotlone

Sun Oct 05 2008, 11:50PM

Registered Member #1383 Joined: Mon Mar 10 2008, 11:09AM
Location: Wollongong, NSW
Posts: 27

I have tried both 10 and 8 turns with no great improvement, it may have helped a little but nothing too drastically. So by having my PCB layout (seen in my first post on this thread) the way it is, my inductance areas are too great? Without proper redesign on the PCB I won't be able to improve performance?

HV Enthusiast

Mon Oct 06 2008, 12:44AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Okay, here is your homework.

1. Take your transformer. Dress all your windings so they exit the core TIGHTLY twisted together (looks like pigtails).

2. With all windings open, measure the inductance of your primary winding. Record as Lo.

3. Now short both secondary windings to each other respectively. Remeasure inductance across primary. Record this as Ls.

4. Now as an approximation, solve (1-k^2) = Ls/Lo (Solve for k)

5. Ls should approximate leakage inductance and k = coupling factor. If the core is right and wound correctly, k should be > 0.990

Next Step

1. Disconnect secondaries of transformer from board. Dress them with twisted pigtails. Take your MOSFETs and wire both directly to the secondaries of the transformer. The MOSFETs will not be connected to anything else. Connect each secondary across gate to source. Gate resistor not really needed here.

2. Look at waveforms again.

3. If your waveforms still look crappy, its your transformer.

4. If the waveforms look excellent, its either a board layout problem, or another component on the board that is wrong.

first previous 3 4 5 6 next last

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