diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   temp drift in MOSFET Vgs (http://www.diyaudio.com/forums/solid-state/82314-temp-drift-mosfet-vgs.html)

SemperFi 29th June 2006 07:26 AM

temp drift in MOSFET Vgs
 
What is the voltage drift of Vgs due to temp in MOSFETs? BJTs have a farely constant -2mV/degree drift, but I have never seen such data for MOSFETs.
I am trying to make a stable CCS with a MOSFET that is biased with a zener in series with a standard diode. A zener with voltage over ruffly 6V starts to drift positively while a diode has the constant -2mV/degree drift. So with a 6.2V zener and a diode in series I get close to zero drift. If the MOSFET has a known drift in one direction, I can compensate with appropriate diodes/zeners.
Thanks

AndrewT 29th June 2006 07:42 AM

Hi,
can you calculate it from the datasheet graph?

The problem is that due to the curve in the graph the slope changes as Id changes. But I think you should be able to estimate it over the current range you plan to use.

With BJTs, once past the curvy knee at the bottom of the slope (at turn on) the rest is close to a straight line and so a nearly constant slope equates to a published figure. I think the same cannot be done for a FET due to the curve.

What I cannot tell you is how the spread of production values and licence on the part of the data sheet compiler will affect your answers.

Maybe after you have done this you will be able to come back and give us a tutorial (even a Wiki) on temp compensation.

SemperFi 29th June 2006 08:04 AM

Thanks for quick reply. This is such an active site it's amazing how fast answers come up!
Having my cup of coffee while scetching up the breadboard made me recall MOSFETs dont have a constant drift, just like you mention. But go from a negative Vgs drift to a positive above some Id value. Which is why MOSFETs do not run away as easily as BJTs or tubes. And at the crossover between negative and positive drift there is zero drift:) But at what value of Id?:( And is it the same for all MOSFETs? Doubt it. Probably a function of voltage and current ratings of the device.
So I agree this is interesting enough for a little research and a follow up report. But where is the left over time for yet another project?:( No matter how much coffee I drink time just doesn't go far enough, and the kids get up at 7 no matter how late I go to bed.

morpheus82 29th June 2006 08:10 AM

i know that the threshold of the mosfet has a temp drift for sure, and depends on the mos...this means all the circuit is influenced, vgs included!you just have to calculate it...

ilimzn 29th June 2006 12:03 PM

Quote:

Originally posted by SemperFi
...MOSFETs dont have a constant drift, just like you mention. But go from a negative Vgs drift to a positive above some Id value... And at the crossover between negative and positive drift there is zero drift:) But at what value of Id?:( And is it the same for all MOSFETs? Doubt it. Probably a function of voltage and current ratings of the device.
This is heavily dependant on the actual MOSFET structure and geometry. Datasheets usually contain a graph of Vgs versus Id plotted for several temperatures. Look for the point where all the various temperature graphs cross, this is the turnover point.
For standard VMOS (HEXFET, Trench FET), this happens at a very high walue of Id, usually about half of the maximum rating or higher. Lateral MOSFETs exhibit a turn over point at MUCH lower currents, on the order of 100mA or so.

It is fair to assume you will be using your MOSFET in a current source providing way below half the current rating of the MOSFET, it should be relatively easy to calculate the thermal drift for the particular MOSFET you want to use [b]at your chosen drain current{/b].
Just look at the Vgs versus Id plot, along the line where Id = Isource, and read off the distance between the various temperature plots, then divide by temperature difference. The drift is not linear but this should be enough for a start. The drift itself is also far less dependant on actual MOSFET tolerances than the threshold voltage, so you don't need to select MOSFETS for this. You will also find that for Idmax >> Isource, it will come out between 4 and 6mV/K, and becomes lower as Isource approaches Idmax for a given device.

Onra 29th June 2006 04:43 PM

Hi,
have a look at Design Note 4 published by ZETEX at:
http://www.zetex.com/3.0/3-8-2.asp
"Temperature Effects On Silicon Semiconductor Devices"
Its a very informative summary of Tcs.


Regards
Onra

poobah 29th June 2006 04:50 PM

Why do you wish to use a MOSFET?

:confused: ;)


All times are GMT. The time now is 04:49 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2