Feeling at home, new coil / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Feeling at home, new coil

previous 1 2 3 next

Move Thread

LAN_403

Steve Conner

Mon Apr 02 2007, 01:17PM

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

Ye gads! It looks like you have a leet new scope AND a DEIC420, was it your birthday or something?!

ShawnLG

Mon Apr 02 2007, 01:46PM

Registered Member #286 Joined: Mon Mar 06 2006, 04:52AM
Location:
Posts: 399

Oh my! Looks like you can reach 13.56Mhz now.

...

Mon Apr 02 2007, 04:37PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

smokes, they make gate drivers that can do 4nf in <5nano seconds

shikes!

Also, those waveforms look pretty good! Lets see if you can get the coil working now

Marko

Mon Apr 02 2007, 05:31PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Omg matt.. the supreme god of all gate drivers (OK, maybe after DEIC515 ;) ..)

Are you running resonant drive there? AFAIK, that driver should do even better in hard drive..
It's rated for 45Mhz 4nF with 3ns rise and fall times O_o

...

Mon Apr 02 2007, 06:38PM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

I think it might simply be to save the fet, I don't even want to think about what current it would be driving into the gate to swing it +/- 10v in a few ns...

WaveRider

Tue Apr 03 2007, 07:47PM

Registered Member #29 Joined: Fri Feb 03 2006, 09:00AM
Location: Hasselt, Belgium
Posts: 500

Hi Matt,
Excellent, you got some DEIC420 gate drivers. The output of this driver should be fed directly to the gate of your MOSFET... No blocking capacitor is needed. Also, earth those ground tabs right up to the case for low inductance (using the nice coplanar PCB structure. I see you are beginning to discover the virtues of that "straight-line" RF layout topology... First one to 30MHz wins!

Great work!
Bill

Marko

Tue Apr 03 2007, 07:56PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

If it can be kept with low enough inductance a large value DC block cap may just save the driver from exploding if mosfet gate goes short.
I also saw DEI use a small value resistor, few ohms, in series with DC supply but *before* decoupling capacitors, to save the driver from overcurrent.

Some kind of fuse may also be a good idea! If I had DEIC420 blow up on me it would definitely leave great emotional impact.

Steve Conner

Tue Apr 03 2007, 08:13PM

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

Yes, a fuse in the supply to the DEIC420, or a LM317 regulator with its 1.5A current limit, might be a good idea.

The layout with the large through-hole DC blocking capacitor and long connecting wires probably hurts the performance by about an order of magnitude, compared to what you could do with the minimum inductance layout without blocking cap on double-sided PCB.

But what I sense BP is worried about (and I would be too) is if the MOSFET shorts between gate and drain, and feeds the drain supply back into the DEIC420, blowing it up too. This could easily happen if the MOSFET failed short between all three terminals and then the source bond wire blew out.

My approach to solving this would be to sense reverse current out of the DEIC420's positive supply pin (or overvoltage on its supply rail) and use it to crowbar the drain supply with a large SCR.

Marko

Tue Apr 03 2007, 08:26PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

There are SMT capacitors wich could be used for DC block, they have very low inductance but they aren't easy to get a hold off.

I never had the failure mode steve described to happen, although some kind of crowbar or protection device would be a good idea.

DC block cap alone isn't best protection there as it would still pass initial 'pulse'.

ragnar

Wed Apr 04 2007, 07:38AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

The new scope is the best bit, but the DEIC420 is to distract me from designing a proper (discrete) high-power gate driver for a little while.

The leet oscilloscope is a Fluke PM3092 (4ch, 200MHz) I'm still hanging onto my old Labtech (2ch, 20MHz) for audio frequency and/or 'risky' stuff. It wasn't my birthday, I'm just in debt. :P

The DEIC420's current layout actually resonates with the gate C now at about 9MHz; it could well be optimized, but it was a little too late at night when I was designing that circuit board. The waveforms at lower frequencies (e.g. 5MHz) are actually surprisingly ugly; there are huge voltage transients at turnon and turnoff, presumably the leakage inductance coming into play can be blamed.

Ah, yes Firkragg, I noticed the DEIC515! **drool**
At first I thought it was a typo... but nope -- as soon as I get my hands on a DEIC420, they supersede it with a newer, better driver chip, hehe.

I don't think it could swing 4nF at 45MHz; all the test waveforms in the datasheet (also on the EVIC420 evaluation board datasheet) are circuits with a 1nF load, but it's still one crazy driver! O_o

I think the high cost is attributed to the fact it contains actual metal-gate devices in the output stage...??

Even if I could hard-drive it at higher frequencies, I would still add a parallel inductor to smooth the waveforms and minimise the current the gatedriver has to provide.

Bill, I'm using the decoupling capacitor to keep the gate voltages reasonable (he resonates to +/-20V with only a 12V supply), and to perhaps prevent hundreds of volts appearing at the output in the event of a gate-drain short. Even though this board isn't a good example, I am very much appreciating RF design and layout techniques.

My roommate had an UnWired-brand wireless microwave modem that failed recently, and I opened it up to have a look. How beautiful it was! Gold-plated PCBs, meticulous use of groundplanes and short traces and neat component orientations; I gained a great deal just by osmosis from looking at it. I am of course always open to criticism or recommendations for my projects, but I do feel terribly guilty when I receive some wonderful advice, but the project falls away into oblivion or I am overcome with university work.

I will seriously consider 27.12MHz... after I conquer the 13.56MHz milestone. And I need a signal generator that can reach 27MHz for that, first. BUT at least I have an oscilloscope that will trigger! =D

I have simply put a 3.15A fuse before the decoupling caps... and it hasn't blown on startup, yet. I made plans for a crowbar circuit sensing the DEIC420's output, but because I was in a hurry to get my board made and catch a train into sydney (5:45AM), it didn't really happen.

At the moment, the DEIC420 draws almost exactly 2A, and the supply voltage drops to 11.7V under load.

Darn, the only surface-mount capacitors I have are tantalums. I must get my hands on some surface-mount ceramics.

Thanks for your feedback guys. It has encouraged me to use a smaller device and go back to the higher frequency work. I still have a very hardy ~14MHz secondary made from 1.5mm diameter wire. I must put it to some use!

Any suggestions about a discrete gatedriver? I'm thinking about class-E using an IRF510 as the switcher. Big supply voltage, big turns ratio on the transformer, and a capacitor on the secondary to make the gate C appear smaller! :)

previous 1 2 3 next

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