John Curl's Blowtorch preamplifier

[scott wurcer]

Originally posted by john curl
However, nowhere, no-how, can I find the exact expression for an ideal (or non-ideal) FET differential pair. The analysis just stops in the texts, and I was hoping that someone would have it handy and would put it here, so we could discuss it.

That's probably because the result never interested Barrie Gilbert
also the IEEE seems to not allow search engines to find key words in their archives. I don't see why, they might sell more downloads of the complete papers.

Like TANH the transfer function of a long tailed FET pair HAS to be expressed as A + Bx**3 + Cx**5.... because it is a symetric odd function. The coefficients will be a function of Idss, Vp, and Ibias. This IS math and not graphs, the exact coefficients might be interesting or not. I have found that one cannot come up with any magic degeneration or arrangement that flattens the transfer function around zero like a bipolar. It is possible an exact solution requires some esoteric transformation of variables or maybe Volterra techniques. I do know that no one I have run into ever found the exact answer interesting.

There is no mystery or magic and no non-idealities are necessary. A long tailed pair of anything creates and i1-i2 function that is symetric and odd.

[andy_c]

Originally posted by scott wurcer
That's probably because the result never interested Barrie Gilbert
also the IEEE seems to not allow search engines to find key words in their archives. I don't see why, they might sell more downloads of the complete papers.

I'm part way there. Can somebody check my math? Here is page 1.

[andy_c]

I am still not finished, but it looks like the last equation on page 2 below can be squared on both sides, then solved for Idm in terms of Vdm. Does this look right so far?

[scott wurcer]

You're on the right track, these days I would use a crutch like Mathcad or Matlab to do explicit symbolic answers. They don't miss those factors of two or the square root of PI.

[Nelson Pass]

Let me guess. The answer is 42.

[john curl]

Nelson, we are doing this for you and me. Stick with the program!

[andy_c]

Originally posted by john curl
Nelson, we are doing this for you and me. Stick with the program!

Hi John,

I'll finish this up tomorrow. I got a new pair of speakers yesterday, so I'm doing a marathon listening session at the moment

[PMA]

Originally posted by scott wurcer




There is no mystery or magic and no non-idealities are necessary. A long tailed pair of anything creates and i1-i2 function that is symetric and odd.

Exactly. That is what I had in mind when I spoke about difference between trasnsfer function of a single device and a long tailed pair.

[john curl]

So far, you have said nothing that we do not already know.

[PMA]

I have not expected to say anything new. You were asking about 3rd in a long-tailed pair, and argued with 2nd in a single device.