Most linear region of Mosfets ?

[Bernhard]

Where is the most linear region of Mosfets ?

When looking at transfer curves of power mosfets it seems they start to be linear at a few A, is that right ?

[traderbam]

When you don't apply electricity to them.

[traderbam]

Seriously.
Can you find the most linear region of either of these?

No semiconductors are particularly linear. The mfrs sort of craft their characteristics to suit different applications. I think the most linear thing you may find is the beta of some BJTs over certain Ic and Vce ranges. BJT Gm is exponentialish. MOSFET Gm is...square-law, cubed law, a bit of both, depends on the specific device.

If you are designing a Gm buffer (source follower driven by a voltage source) I think your best bet is to try to linearize a MOSFET either with local NFB or source degeneration resistors. Eg: if you use a 200m-ohm source resistor the MOSFET Gm needs to be higher than 5 at the bias current (very roughly), and then you have to look at the Gm vs Ic graph to find the lowest bias current. BJTs are much better in this respect because their Gm is way higher for a given Ic...in this example you might need 2A bias for a FET but only 150mA for a BJT.

Brian.

[kf_tam]

The lateral MOSFET seems to be better. This may justify their higher price.

[Bernhard]

Originally posted by kf_tam
The lateral MOSFET seems to be better. This may justify their higher price.

Yes, the lines are almost straight from 200mA on, and 2SK216 looks better.

[Bernhard]

Originally posted by traderbam
Seriously.
Can you find the most linear region of either of these?

For the ZVN I would try to stay around 80ms.
The IRF scale is not linear, so it is useless.

Here is another IRF250.
Starts to look acceptable at 5A and good at 7,5 A

Originally posted by traderbam
in this example you might need 2A bias for a FET but only 150mA for a BJT.

My thought for classA was to use only one single IRF that conducts all the highest possible current.
And use a cascode to keep the dissipation down to 25W.

Or better toast some 2SK175.

[Bernhard]

It looks like the curve of a bipolar transistor goes into the other direction, 2SC3281 for example, is it possible to combine both for something better ?

[Bernhard]

Originally posted by traderbam

No semiconductors are particularly linear.

That's why I have no semiconductors in the path between DAC chips and power amplifier...

[Nelson Pass]

Originally posted by Bernhard
Where is the most linear region of Mosfets ?

When looking at transfer curves of power mosfets it seems they start to be linear at a few A, is that right ?

The most linear region depends on the exact context of the load line,
and this varies from application to application. The curves cited here
are only part of the information needed.

[Bernhard]

What about a circlotron with a Mosfet on one side and a bipolar transistor on the other ?