P channel or N channel fets better for audio?

[GregH2]

Hi all,

I'm just interested to know if N or P channel fets are more desirable for audio (cost and availability aside). Is one type generally more linear or less noisy or otherwise better? Or are they much the same?

I ask this because I am designing an output stage at the moment that could potentially use either device and am trying to weigh up the pros and cons of each.

Your advice is appreciated!

Greg.

[lineup]

They are the same mostly.
One advantage using N-Channel is that there are more of them to chose from.
More N in the shop.

[Gyuri]

And usually they are cheaper.
Hej, bratyó, jól vagy?

[GregH2]

Guess you guys missed this part of my post:

Originally Posted by swordfishy

Hi all,

... (cost and availability aside).

Glad to hear that's the only real advantage though.

This design will use plain old irfp9240s, so no problems getting parts

[jcx]

N channel has higher electron mobility, bigger gm, smaller die = less gate, parasitic C for the same current rating, gain

[lineup]

IRFP9240
is to prefer to the older IRF9240

[godfrey]

From John Curl's thread:

Originally Posted by john curl

Well SY, I think that any P channel input is a poor choice. I said the same thing to Scott Wurcer, a few years ago, because he made the same 'mistake'.
Now what is the problem? In this case, ALL ELSE BEING EQUAL, P channel devices have LESS Gm and higher 1/F noise...
[big snip]

He's talking about device choice for input stages, but maybe you'd be interested in the discussion there anyway.

Edit: Oops, I just realized you're talking about power mosfets - that's different. Some P-channel mosfets show slightly weird frequency-dependent behavior. Nelson Pass (and others) have written about this. IIRC, it's the IRF ones.

Re-edit: Ah, found links:

Originally Posted by Nelson Pass

This is not correct. The IRF P channel parts have an odd
characteristic which shelves the transconductance in the mid
audio band. As such, it is somewhat flawed for use in
Common-Source applications, but it is acceptable in Common-
Drain use.

IR P channel parts made by other manufacturers, such as
Harris or Fairchild do not exhibit this.

In the case of X amps, it is not an issue.

Originally Posted by Nelson Pass

You can check your Mosfets easily enough, the IRF parts have a
shelf in the transconductance vs frequency curves. If you need some
hints you can look at the Mosfet testing article at

Pass Laboratories

Of course it makes a difference, but it is minimized in follower applications.

Originally Posted by Nelson Pass

As I have posted elsewhere, parts from Fairchild, in particular
the 9610, should be tested, as some of them now exhibit the
"P channel anomaly". This is apparently due to Fairchild acquiring
the Samsung production.

[Jay]

You're not going to build one transistor amp, are you? For two-stage all-mosfet amplifier, if you use N-channel at the output you will use P-channel at the input. I prefer N-channel at the input because the input stage is more important than the output stage and it is easier to find good quality N-channel transistor (actually, not many).

If the amp is using the (MOS)FET only in the output, and if you are after a quality amp, I think better to use N-channel because there are not many good mosfet around. Why not start with choosing the part first. After all, what matters is the specifications not P or N.

[MiiB]

You can make GND at either rail....there will be no difference.. if you choose the upper rail...the amp swings negative.. if you choose the lower the amps swings positive... after all electrons are running the other way...we point current down.. when in fact its running up...

[GregH2]

Thank you all for your comments. There has been some good advice here - pretty much going to avoid p channel devices now. The circuit I am talking about can be seen here:

My take on the F4 - advice needed

I have already built it with J202 jfets on the front end and it sounds quite good. I just want to finalise a circuit so I can have some PCBs made now.

Does anyone know if the same problem occurs with p channel jfets?