DRSSTC on a "chip"!! / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

DRSSTC on a "chip"!!

Move Thread

LAN_403

Terry Fritz

Wed Aug 23 2006, 04:39AM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi,

OMG!! Check out what IR just announced:

20V gate drive, 2.5A, 100kHz++ operation. About $1.00 in 10K quantity. About 25nS times with under voltage cutoff, programmable or no dead time, Schmitt trigger inputs... This removes the gate drive transformers, isolation, gate drive power supplies, cross conduction worries... They have all kinds to for different needs. Just needs a primary current CT, low voltage power supply, IGBTs, buss caps, plus a standard DRSSTC controller and your pretty much done!!

They now have the voltage, current, speed to do our bidding

I hope they hit DigiKey soon!!

Cheers,

Terry

[Edit: Picture width reduced to the required 400 pixels]

Dr. Drone

Wed Aug 23 2006, 06:47AM

Registered Member #290 Joined: Mon Mar 06 2006, 08:24PM
Location:
Posts: 1673

Marko

Wed Aug 23 2006, 11:09AM

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

But isn't that chip a bit too weak (2,5amps) for driving IGBT's (If you have to add gate drive, what's the point?)

I would rather call it SSTC on chip

I didn't look at them all but do they have some kind of full bridge driver, optoisolated maybe?

HV Enthusiast

Wed Aug 23 2006, 11:31AM

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

Interesting find Terry, thanks for posting . . .

Although the chip looks promising from a quick glance, there are several things that make it unsuitable for even a small sized DRSSTC.

1. Shutdown is not synchronized with input drive. Output gate drive signals are shut down immediately upon receiving a shutdown signal, therefore this could lead to considerable voltage spiking across the devices when the device is shut down mid pulse (high peak current through it)

2. Propagation delays in the chip are quite high between input signal and output gate drive. Time between rising edge of input pulse and rising edge of output gate pulse can be a much as almost 1 us and typically 680ns. This isn't good and will be detrimental to achieving soft-switching of the IGBTs.

3. Output gate drive is pretty low and would not be capable of driving the typical IGBTs we use in DRSSTCs.

I would agree with Firkragg that this would likely only be adequate for standard SSTC operation of small coils. (small devices)

teravolt

Thu Aug 24 2006, 03:07AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

fairchild also has one of thease half bridge drivers FAN3782 just look up bridge drivers
----------
the FAN3782 and your IR chip has built in dead time I think

[mod edit] Please don't double post.

Steve Conner

Thu Aug 24 2006, 10:12AM

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

Meh, call me old fashioned, but I'm not sure I trust these new fangled "X on a chip" chips. The problem is that the silicon guts of the chip are exposed to the full DC bus voltage and the inverter output voltage. On whatever kind of SSTC, these voltages tend to be filthy and riddled with spikes, especially if you get a primary-to-secondary flashover. Gate drive transformers don't seem to care about the spikes, and they protect the circuitry behind them too. But I bet the spikes would trash a HVIC like this.

Having said that, if you could get it to work reliably, I'm sure it could start another craze!

Terry Fritz

Thu Aug 24 2006, 05:34PM

Registered Member #393 Joined: Tue Apr 18 2006, 12:30AM
Location:
Posts: 297

Hi,

A lot more is explained here:

The internal charge pumps can run buffer stages directly if you "need" to pump "8" amps "directly" into "each" gate as shown on section 7 and figure 15. The external FETs are tiny 75 cent items. Figures 12a and 12b or doing 600nC modules. They show twisted pairs as if the whole circuit could not fit on top of the brick (all available in surface mount ICs

"I" would check the typical 2.5 amp drive though first to insure that it is "not" enough.

"Ideally" they would be run directly from a small microprocessor that could take care of all shutdown and timing issues directly. Then you could "look ahead" and time the actual transitions as you wish. There would be external 20VDC power, CT feedback, and the controller timing signal. Not much else to it... The controller would need to be programmed and all, but the cost is very low after that. Even the delay line talked about elsewhere might do the look ahead timing. If you want to switch exactly at zero current, you need look ahead logic...

Since the microprocessor could completely control each of four IGBTs directly, the opportunity for spectacular explosions is great

But the entire control circuit might cost less than a single IGBT, and be smaller too...

All "Pie in the Sky" Stuff, but "Fun" Stuff!!

Cheers,

Terry

Marko

Thu Aug 24 2006, 09:13PM

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

But if yu need to add external components again, what's the point?

I would still simply use UCC's and GDT, simple bistable would keep timing in tune. (this scheme also pwns others by cost and availability)

Maybe it's better idea for bricks where you could eliminate isolated supply that way.. (you need a bit bigger mosfets then)

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