HV gate driver

I need to make some high voltage pulses, and found this cool IR2213 gate driver. The datasheet is fairly fuzzy and there are no appnotes or eval boards, so I figured I'd breadboard it just to make sure.

formatting link

formatting link

Seems to work OK, but it really needs a 15 volt supply to work. And then the logic inputs have to be 15 volts, which means I'll be using a gate driver chip to drive a gate driver chip.

The inputs have to be driven out of phase, with anti-shoot-through timing, too. I guess I shouldn't complain.

I'll be needing some high-voltage scope probes. Digikey has them from $270 to $1600. Yikes. So I tried Amazon: $25 each, free shipping with Amazon Prime.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

Why, what for? The thing (like a plethora of other similar ICs from International Rectifier) has 2 separate supplies for a reason. The VDD (logic) side is specified to work from 3 V (up to 20 V) and the VCC and VB (power output) side works from 12 V (also up to 20 V).

The logic side should therefore run at 3.3 V. Even if that turns out to have some gotchas (hey, they're IR, they've grown up in the days when the dinosaurs roamed the wide plains), strapping a 78L05 off the 15 V and running the logic at 5 V should work just fine.

Is there a reason why you needed the logic to run at 15 V as well?

Regards Dimitrij

Reply to
Dimitrij Klingbeil

We have one at work cost ~ $5k and that is a cheap on3, but it does 50kV @200 Mhz. Of course, I also rolls around on a cart with a BNC at the base :)

Jamie

Reply to
M Philbrook

Hey, you're right: it does work with 5 volts on the Vdd logic supply. It needs 15 on the Vcc side (it actually starts up at 13.4. Maybe because I'm using four diodes in series in the bootstrap thing) So I don't need the IXYS pre-drivers. I do still need to do some logic, which is easy at 5 volts.

Thanks. I'm making +-5 from +24 with a cute little CUI dual-output DC/DC converter.

Yeah, IR is one of the few survivors from the toob age. Their data sheets and web site look it. Recently acquired by Infineon.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Den fredag den 9. oktober 2015 kl. 03.06.44 UTC+2 skrev John Larkin:

yeh, it says right on the first page that it is 3.3V logic compatible but it never specifies what the thresholds are for anything but 15V

the under voltage lockout can be is as high as 11.7V, that doesn't leave much head room with 4 diodes

-Lasse

Reply to
Lasse Langwadt Christensen

Right, the "Static Electrical Characteristics" threw me. There are strange graphs of Vl vs Vdd later on in the data sheet.

Fig 23 and such are hilarious. "IR2213S vs. Frequency"

This is an unusually bad data sheet.

I suppose I should bump the Vcc voltage up a little. Nuisance.

Reply to
John Larkin

Den fredag den 9. oktober 2015 kl. 18.42.53 UTC+2 skrev John Larkin:

or only use one diode

do you need the 1200V? there are lots of different parts that can handle 600V

-Lasse

Reply to
Lasse Langwadt Christensen

I'm already using some MMBD5004S (dual 400 volt SOT23) diodes in the C-W multiplier power supply, so I figured I could use two of the same (1600 volts total) in the bootstrap. I guess I could use three sections for 1200 volts, save one diode drop.

The customers are still a little fuzzy about how much voltage they need (of course the delivery date can't change) so I'll design for the max that the IR chip can handle. 1200 volt mosfets are fairly cheap, too.

The real nuisance with high-voltage design is connectors.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Geez, I've been using the IR2113S since 1998! Great little chip.

IR makes a WHOLE series of gate driver chips. many of the newer ones will take lower logic levels while still running off 12 - 15 V supply. I now use the 2184 in some newer designs with 3.3V CLPDs in them, instead of 40xx series logic.

Jon

Reply to
Jon Elson

The 2184 is cute, but only 600 volts.

1998? Wow.
--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

I used the HIP4080 to build a PWM servo amp. But, they kept blowing up. I finally called an app engineer at Harris, and he said "Oh, you've got them working at 59 V? That's really good, none of our other customers got them to work over 50 V." Great news, as the part is advertised as able to run off an 80 V DC supply, with transients up to 96 V. But, it was the PERFECT chip, put everything but the FETs on one chip. So, I redesigned the whole thing, using a raft of parts, as I had to make a triangle wave oscillator, PWM modulator, and the FET drivers. But, never have blown another one of them up, using the IR FET drivers. The original analog-input servo amp was from 1998. In 2002 or so, I started making PWM input servo amps, and have made hundreds with the IR2113S.

Then, I made an AC servo amp using the IR 3 half-bridges in one driver. This had a lot of problems due to really weak drive, so I redid it using 3 of the 2184, and was able to greatly reduce the dead time as the transistors were switched a lot faster. I think I went from 800 ns dead time to 200 ns.

Jon

Reply to
Jon Elson

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.