"Thumper" / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

"Thumper"

first previous 6 7 8 9 next last

Move Thread

LAN_403

Finn Hammer

Tue Jun 09 2009, 08:59PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Just a couple of pretty pictures before I crash.

First the busswork, now all finished, and with copper risers for connection to primary circuit. These risers will go trough the current feedback transformers.

Next, just to show how I intend to wrap it all up:
A nice alluminium box to contain the "Predikter" driver board, just finished soldering the 2nd. version with major bug fixes.
Nice short wires protruding with gate drive signals. They are not there yet, just in case you wonder.

I ordered the box a little small, another one is in the mail.
So it is going to become a nice compact motor for the "MK4"

Cheers, Finn Hammer

Extreme Electronics

Tue Jun 09 2009, 09:51PM

Registered Member #74 Joined: Thu Feb 09 2006, 09:17AM
Location: Nottingham UK
Posts: 99

Fantastic engineering as always Finn. Although it would be a great shame to hide that Bridge in an aluminium case. Someone must make conductive plexy ;)

Derek

Dalus

Tue Jun 09 2009, 10:06PM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

One small thing that you might want to do. If I recall correctly Steve Wards bridge once arced over between two sharp edges on the busbar. You might want to round them just in case. Wouldn't want any destroyed IGBT's

Finn Hammer

Wed Jun 10 2009, 08:48AM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Extreme Electronics wrote ...

Fantastic engineering as always Finn. Although it would be a great shame to hide that Bridge in an aluminium case. Someone must make conductive plexy ;)

Derek

They do, actually. And I have a piece of it somewhere in the pile of junk I call my workshop.
It consists of plexi with a very fine mesh cast into the top 1/10th of a millimeter. The mesh is invisible to the naked eye of us mortals, excluding birds of prey, perhaps.

But with a magnifying glass it becomes visible to the unwashed masses, even. Problem is getting contact with it: Paint the edges with conductive paint, then interface with that: Not to my liking.
Apart from that, pretty as it is, I never was much into plexi. Apo-plexi, perhaps...

Dalus wrote ...

One small thing that you might want to do. If I recall correctly Steve Wards bridge once arced over between two sharp edges on the busbar. You might want to round them just in case. Wouldn't want any destroyed IGBT's

I think Steve was probably having creepage before flashover. Makes a big difference with RF signals of power magnitude.

Cheers, Finn Hammer

Finn Hammer

Thu Jun 11 2009, 07:13PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

What a wonderful day!

Started early after a 24hours shift, downloaded the mazilli schematic from HvWiki, and drew it into eagle.
Got it layed out in Board, printed out the transparancies, and by 11 hours, I had the board, drilled and ready for population.
Stuck the parts in, and tried it out with temporary windings on the pot core transformer: It worked, oscillating at 96kHz.

This is the nice thing about making projects in eagle: They always work 1st. time.
The cuts into the board are to prevent the high side supply from tracking to low side.

Now it is time to work out the magnetics. I had a nice pot e-core I wanted to try out because it looks really pro

and I had to make a custom bobbin for it too, the white one. Separate chambers for each gate supply, some of them bounching up 1200V and back. Well, maby only 1120V

I wound the primary, 2x5turns, with some 0.07mm thin foil I had lying around. And needed to make a really compact primary.

When It was all done, 12 hours had passed, and I was - sort of - satisfied:

No doubt, due to the tight coupling in this transformer, the output is quite stiff, as opposed to when I used a flyback core, with the primary on one leg and the secondary's on the other legs:

So now I have to find a way to stop frying the shunt zeners on the gate driver boards.

What is the most elegant way to limit the current out of the Mazzilli? Chokes on the driver boards or can it be done with another, bigger value series inductor, the one on the Mazzilli driver that goes to center leg of transformer? I am using one 0.2mH on yellow core. (So many things to know)

Ideas are invited here, cause I can't think or even see straight, having worked for 36 hours on end.
(But a great time was had)

Cheers, Finn hammer.

EDIT!

I s'pose it would be double posting, even though I answer the question myself.

The problem was, that the mazilli could put out considerable current, and that the gate driver board had a capacitor input filter, very crudely regulated by shunt Zeners.

These zeners turned hot very fast when I connected the mazilli to them, so some sort of current limitation had to be found.

The solution I came up with, was to change the filter on the gate driver boards to a choke input filter. In this case, a 1mH inductor did the trick. I had to increase the turns count on the mazilli transformer, from 10 to 15 to get enough voltage on the boards with this new filter, but the good thing is this:
I have full gate drive voltage, -10 - +20 V from 160V line, and up to 240V, where the zeners are still cool.

Another cheers from Denmark!

Finn Hammer

Mon Jun 15 2009, 06:47AM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

The board and sch files are here:

All Eagle files, enjoy!

Cheers, Finn Hammer

Z28Fistergod

Tue Jun 16 2009, 12:27AM

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

Do you use just one of your inverter digital phase lag circuits? Or do you use one for each leg of the H-Bridge, which would allow you to set some dead time (or overlap)?

Finn Hammer

Wed Jun 17 2009, 08:56PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Z28Fistergod wrote ...

Do you use just one of your inverter digital phase lag circuits? Or do you use one for each leg of the H-Bridge, which would allow you to set some dead time (or overlap)?

First delay stage arranges ZCS, the next stages set dead time.

Cheers, Finn Hammer

Finn Hammer

Thu Jun 25 2009, 09:00PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

All,

I just assembled the "Thumper" motor:

As you see, it is preset for a 6.01kA current limit.

But under initial testing I saw that all the gates were +20V. What a bummer: during design and debug, I ran them from a sig gen. and did oversee the fact, that when the enable pin of the UCC37322 output chips are low, then the output is also low.
The optocouplers demand a current signal to reach high, which turns the gate drivers off.

At present I canÂ´t see any easy or proper way to get around this problem, so maby this is a hint, that I should go for gate transformer after all.

I am tempted, sure would be easier.....

Those gate drivers were not safe. A missing signal should not cause failure.

EDIT.

Come to think of it, a small bipolar could reverse the drive and make it soso.
Maby I will hear the bridge sing tomorrow after all.

Cheers, Finn Hammer

Mathias

Fri Jun 26 2009, 06:49AM

Registered Member #1381 Joined: Fri Mar 07 2008, 05:24PM
Location: Hungary
Posts: 74

Is it a good idea to mount the control electronics so close, in between the output terminals ?

first previous 6 7 8 9 next last

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