Thermal dependence of gate voltage to temperature

[mohamis288]

Hi,
I have uploaded two file as attachment. one is the datasheet for MRF9080 which is a RF power amplifier and another is thermal compensator for gate biasing of this IC. in the latter one, it is claimed that the MRF9080 is the gate voltage has a -2.8mV/Celsius temperature coefficient. but when I closely see the MRF9080 datasheet, I can find out a graph or table showing this thermal coefficient. do you know according to what information, -2.8mV/Celsius temperature coefficient is expressed? do you think it is as per measurements on the real IC?
any help will be appreciated.
thank you

 

[crutschow]

All MOSFETs have a Vgs(ths) temperature variation which is not often documented in their data sheets.
Here's a paper discussing that dependence.

 

[d123]

Hi,

As your question is very specific, and presumably for a professional application, you should have no problems contacting Renesas (Intersil owner now) and querying the -2.8mV/C value. More efficient (one would hope/expect) would be to contact NXP, who make the transistor, and ask them what Vgs temperature drop per °C is for the part - that's what Support and I'd suppose also your friendly neighbourhood Field Support Engineer are there for. Support departments are usually helpful people, even with lowly hobbyists who have serious (not frivolous or simpleton) questions.

I searched a bit (so I can pretend I'm an expert on LDMOS now as well ) and LDMOS stuff repeatedly mentioned - 2.5mV/C, but serious design based on 'some figure in an app note I can't qualify and stuff some idiotic guy on a forum said' is not good enough for a professional design, you should check with the part manufacturer.

 

[crutschow]

Any design using that device would need to have the bias compensated (typically with negative DC feedback) so the circuit will operate properly over the desired temperature range.

 

[danadakk]

Of course one is trying to prevent thermal runaway/HV, so a design
would want to look at those parameters, and correct the bias based
on that. That would include load parameters as well. In that scheme of
things Vgs drift might become irrelevant as one thinks of measuring SWR,
Power, peak V, current and using that to control bias.

YO4HFU Website

But if you are not going to do all that then

https://www.ee.co.za/wp-content/uploads/legacy/26-29.pdf

Regards, Dana.,