The sound of bipolar vs FET

[Jimbo]

Hi,

Can anyone comment on the sound of bipolar amp stages versus FET -- specifically JFET -- in a class A circuit?

What are the timbrel differences, if any?

Thanks.

[Matt Syson]

Hi
The differences will be more due to the overal design rather than whether they are bipolar or FET. Competent amplifiers could use either technology, equally, either can be terrible!
Matt S

[Wavebourn]

Jim;

I call it Tango pair. It sounds very detailed though vaporizes screwdrivers and binding posts:

http://wavebourn.com/images/audio/tower-II.gif

(output BJT+FET pair, see upper rail, voltage follower; lower rail is a voltage to current convertor for SE class A output where voltage follower sees always the same current)

[JohnRoberts]

What Matt said...

JR

[Ethan Winer]

Originally Posted by Jimbo

What are the timbrel differences, if any?

If the timbre (frequency response) of two amplifier circuits are not identical, then one or both of them are in error.

--Ethan

[Matt Syson]

Hi
There are so many factors to consider that it is virtually impossible to say what the differences between transistor types would be.
You have to have very similar circuit design and layout, which you can't because the FET has a low current drive requirement compared to bipolar. You also need to see how it responds to 'out of band' signals, both inputted deliberately, and 'interference' of RF and impulses fed in via the input, speaker leads or the mains.
Why do you want to know anyway?
I didn't think that Jfets were available as power devices, you have MOS fets and IGBT fet / bipolar hybrids and plain bipolars.
Matt S

[tINY]

If you overdrive the circuit, the JFET will sound different than the Bi-Polar topology.

This is due to the JFET being a square law device (like a tube). In low-feedback topologies, the even-order distortion will sneak in long before overloading. But, most modern Pre-amp circuits use a lot of negative feed-back and this makes them almost impossible to distinguish from bi-polar circuits until they start distorting.




-tINY

[Matt Syson]

If you are intending to drive it to clipping then all bets are off as you have to consider the possibility (likelihood) of latch up either temporarily (with good design) or even bursting into oscillation.
I had the pleasure of repairing a guitar amp a while back, which broke just about all the 'rules' of amp design as proposed by Mr Doug Self. It hummed, distorted and oscillated BUT for it's intended purpose it 'sounded' great.
Matt S

[Jimbo]

Originally Posted by Matt Syson

Why do you want to know anyway?

Hi Matt. Thanks for your response. There are a few things that brought this question:
1. About a year ago I built a Hamptone HJFP2. The Hamptone box was originally a tube design, and he modified the circuit, somewhat, to replace the tubes with JFETs. It's mentioned somewhere (on the Hamptone site I believe) that JFETs approximate tubes in response.
2. I use a Hafler Transnova to power my monitors. The Transnova series uses MOSFETs. Some reviews of the Transnova series mention that MOSFETs are softer sounding than other solid state amp devices.
3. I recently spoke with Geoff Daking who told me that his preamp design is bi-polar.

So, I was just curious if the transistor type made a difference in the timbre of the amplified signal.

Originally Posted by tINY

If you overdrive the circuit, the JFET will sound different than the Bi-Polar topology.

This is due to the JFET being a square law device (like a tube). In low-feedback topologies, the even-order distortion will sneak in long before overloading. But, most modern Pre-amp circuits use a lot of negative feed-back and this makes them almost impossible to distinguish from bi-polar circuits until they start distorting.

Thanks for this response, tINY. This is very good information. In an email Scott Hampton told me that his circuit has very little (almost none) negative feedback. I didn't know what that meant at the time.

When you say, "the even order distortion will sneak in long before overloading" do you mean that as the gain is increased the even-order distortion will increase up until the amp is overloaded?

Thanks for the responses.

[Matt Syson]

Hi
Thanks for answering.
In a design with 'no' feedback the inherent non linearity of the amplifying device (whatever type it is) will increase and the harmonics emphasised (distorted) will change. When the output voltage gets close to the supply rails then it will distort far more accutely (clipping).
An amplifier with plenty of feedback will tend to have lower distortion (the function of the feedback) until it gets close to the supply rails at which point the amplifier will go into pretty severe clipping and if the design is not 'up to par' can latch up, which would sound terrible.
Matt S

[Jimbo]

Great information Matt and tINY.

I've learned a lot.

Thanks,

- Jim