Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: john larkin <jl@glen--canyon.com>
Newsgroups: sci.electronics.design
Subject: Re: LTSpice model for a SiC MOSFET
Date: Sun, 18 May 2025 20:38:28 -0700
Organization: A noiseless patient Spider
Lines: 76
Message-ID: <9n8l2k9pc9jq3iichdq1baq04uqqk13ksg@4ax.com>
References: <100c4og$t4lo$1@dont-email.me> <n7qj2kpht38kdk82o1hn10sss3ie2vnbrf@4ax.com> <100e4nq$1dhk3$1@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Injection-Date: Mon, 19 May 2025 05:38:31 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="1a16d2651680f2075d7a23bb8ddfb2d5";
	logging-data="1518869"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+HhsoGS+EcWc4VdhnlDg9m"
User-Agent: ForteAgent/8.00.32.1272
Cancel-Lock: sha1:nTbplaTJwEZXGMZ+jA7MxwfjqG4=

On Mon, 19 May 2025 12:23:54 +1000, Bill Sloman <bill.sloman@ieee.org>
wrote:

>On 19/05/2025 12:15 am, john larkin wrote:
>> On Sun, 18 May 2025 18:11:58 +1000, Bill Sloman <bill.sloman@ieee.org>
>> wrote:
>> 
>>> I'm looking at a problem where somebody wants to step down a 1kV low
>>> current source to 3.3V.
>>>
>>> The Baxandall class-D oscillator could do it, but it needs a pair 1.7kV
>>> MOSFETs for the job. The Infineon SiC IMH170R450M1 would do it - though
>>> it's a much higher current part (10A) than the job needs (about 1mA).
>>>
>>> I've dived into the Infineon rabbit-hole which promises LTSpice models,
>>> but wasn't able to find one.
>>>
>>> Does anybody know of a similar - ideally cheaper and smaller - part for
>>> which there is an LTSpice model?
>> 
>> I use a Cree/Wolfspeed 1200v part, C2M0280120D, in my Pockels Cell
>> driver.
>> 
>> https://www.dropbox.com/scl/fi/5arhyamrp0bl3tgb2fasn/DSC02771.JPG?rlkey=3ttcc2yt6s9nrtdouuv3aneol&raw=1
>> 
>> They do have an LT Spice model library that works.
>> 
>> Gate drive for SiC parts is a bear. I did it myself, but I think there
>> are chips for that now.
>> 
>> There are multi-kilovolt silicon mosfets too.
>> 
>> Baxandal looks to be inefficient and expensive as a low power
>> converter. The drain swing is 2x the supply voltage, and it needs two
>> fets and a difficult custom transformer.
>
>It isn't going to be inefficient. That configuration is famous efficient.
>
>The drain swing is actually 1.67 times the supply voltage, but it does 
>need two switching devices and a specially wound transformer (and we 
>know how reluctant you are to design them or get them made).
>
>It is probably going to be too expensive for the application, and we'd 
>be grateful for your insights into a cheaper alternative. I can't think 
>of one.

I'm not sure what the specs are, but I have a few ideas.

One could make a flyback converter with a high-ratio transformer.
Coilcraft makes some, capacitor charging transformers and CCFLs. There
must be crazy cheap Indian or Chinese CCFL transformers.

ST makes a 1400v NPN transistor for under a dollar.

It would be cool to put two drum core inductors next to one another,
or on opposite sides of the board, to make a loosely coupled
transformer, exactly what a forward converter needs.

I've finally finished by dummy load board... Gerber day is tomorrow.
The paired Murata drum cores are spaced to tune the coupling factor to
K=0.6

https://www.dropbox.com/scl/fi/57jecrzc894uvktv72wrg/P978_A18.jpg?rlkey=4095oct5enxqp556xf44oy491&raw=1


So, how to get the low duty cycle pulsed base drive? I'm thinking
maybe an RC off the HV supply and a diac, a relaxation oscillator.

If the rig is a forward converter, we could make a non-saturating
blocking oscillator, and share the transformer secondary to drive the
base and rectify to 3.3v.

$6 or $7 might be a reasonable parts cost target in modest volume.
Needs Spicing.