1. Home
  2. Electronics Design
  3. Phil : SAV-551 model

Phil : SAV-551 model

Hey, Phil,

I just nabbed your SAV-551 model

formatting link

and noticed that there is no Cds, which I can add, and no lead inductances, which I can add and you don't need for a low-frequency model. Does Spice even allow lead inductance in a mesfet model?

I measured Cds as about 1 pF at zero drain volts, dropping below 0.5 pF above 1 volt. I guess Spice can manage the nonlinearity with some mumble mumble parameters.

Thanks for the model. I'll tweak it a bit for my switching applications.

Any guesses about lead inductances? Maybe 1 nH pin+wire bond? Maybe the chip sits on the source lead frame, so source inductance may be less. It has two pins anyhow.

I need an x-ray machine with wirebond resolution.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin
Loading thread data ...

It's probably a bit optimistic on the drain impedance. The Avago ATF38143 model I started from had a bunch of parameters that LTspice doesn't understand. I didn't either, so I just commented them out. ;)

Besides that, my main contribution was getting the noise a bit closer to right (they were pessimstic by like 14 orders of magnitude iirc).

Avago and Infineon datasheets often have a really detailed subcircuit models. 1 nH seems on the high side. I've seen laser driver circuits that were less than half of that--about 200 pH in the drain circuit and

200 in the source. (This was based on SPICE fitting to measured pulse shapes.) So a SWAG would be 150 pH per lead.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com
Reply to
Phil Hobbs

That's encouraging. Some things that I'd like to do are wrecked in simulation by 1 nH of drain inductance, or 0.5 assuming two source leads. If the source pins are part of the lead frame, that could be pretty low. I guess I could measure between the two source pins, with TDR or something.

Or an RF generator and a scope, maybe resonate it with a cap or two.

Two source leads do make a half-Kelvin connection.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin

The circuit I was working on had one of those flip-chip GaN FETs, an

0402 cap of a few nanofarads, and one of the triple-stacked Osram pulsed lasers. It produced 5-ns pulses of a few amps from a 40V supply. That was about a factor of 2 faster than competing units.

BTW: Anybody interested in a brand new, top-of-the-line Velodyne lidar for fairly cheap? I've had this one kicking around for a year, and I'm getting sick of tripping over the crate. So far I can't get the client to take it back.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com
Reply to
Phil Hobbs

Hard to get the cap small enough that its inductance is negligible. (It would be a fun demo to make a loop out of high-K ceramic disc caps, so that most of the current path is pure displacement current, and show that it has inductance anyway.)

You folks don't do a lot of sine wave stuff, but that's the sort of measurement that network analyzers are good at, especially handled by somebody who knows how to de-embed the device from the fixture. (I'm not one of those people either--the last time I used a network analyzer for anything nontrivial was in about 1982.)

Near DC, at least. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com
Reply to
Phil Hobbs

Suppose I bias up the gate, force some sinewave current into the drain, ground one source, and scope the other?

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin

Might work fine. I don't have a good feel for the coefficient of inductive coupling between the two paths, but it's probably fairly small--good enough for one significant figure maybe.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com
Reply to
Phil Hobbs

We both miss the Avago SOT-89 phemts. The source was a (relatively) giant metal tab that could be soldered to a big ground pour with lots of vias going down. They made great laser drivers. The SAV parts are fine for low-level amplifying and switching, but the package is a thermal and electrical limit.

The thing that most resembles the big Avago parts is a SOT-89 MMIC. They are even worse characterized than the phemts.

I need to learn more about MMICS. I think there are some GaN parts now, which might be really interesting.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin

This looks new:

formatting link

It's electrically similar to one of the SAV parts (low voltage, low current) but in a better package. The data sheet is much worse... no DC curves at all.

They label the gate RF IN and the drain RF OUT so's not to confuse the poor RF boys.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin

Jeez. The package is no bigger than fly poop.

Reply to
John S

That's a big problem with fast stuff. It needs to be tiny and needs a lot of power dissipation, so the thermal situation can be grim.

The EPC GaN fets are really tiny BGAs, even worse. It takes real soldering skills to use these parts, and creates powerful reasons to avoid blowing them up.

The only way to test a circuit with parts like this is to make a multilayer PC board, and have it assembled by experts. Keeps the riffraff away, I guess.

It doesn't help that the data sheets are so bad, and that there are very rarely Spice models for RF discretes and MMICs.

Grrrrr.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin

Well, 'drain' does sound like a gazinta, after all. ;)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com
Reply to
Phil Hobbs

snipped-for-privacy@highlandsniptechnology.com wrote: ,,,>

Put some "Loethonig" (solved flux) on the pads, put the EPC on the pre-soldered pads, best preheat the PCB and than take the hot air gun with low flow. Watch under the scope,. Been there, done that...

--
Uwe Bonnes                bon@elektron.ikp.physik.tu-darmstadt.de 

Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt 
--------- Tel. 06151 1623569 ------- Fax. 06151 1623305 ---------
Reply to
Uwe Bonnes

The Chinese or someone should make a PCB with those tiny chips.

Reply to
LM

It's crazy that EPC doesn't have loaded breakout boards. We had to make our own.

formatting link 

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard
Reply to
jlarkin

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.