F6 Amplifier

[Zen Mod]

so - having degeneration because of tempco , while not sacrificing xconductance

ZAP per personal choice

[buzzforb]

Nelson, any experience with the Cree parts?

[Nelson Pass]

Originally Posted by lhquam

Rs=0R47 sounds about right for thermal stability.
Any reason not to use one or more "Zen pots"?

1.0 ohms would be even better, unless you use some temperature
compensation. The technique of F5 would be an example.

No reason not to use "Zen pots" - on one extreme, AC degeneration, on the
other extreme, my suggestion above.

Of course some laterals will solve the temperature coefficient issue, but will
have a lot lower transconductance. Still, they'll work fine - as I said earlier,
we have open loop gain to burn.

[lhquam]

EUVL and others have implemented F5s without thermistors with Rs as low as 0R22 and claim that they are thermally stable. I suppose it could also depend on the choice of MOSFET.

Originally Posted by Nelson Pass

1.0 ohms would be even better, unless you use some temperature
compensation. The technique of F5 would be an example.

No reason not to use "Zen pots" - on one extreme, AC degeneration, on the
other extreme, my suggestion above.

Of course some laterals will solve the temperature coefficient issue, but will
have a lot lower transconductance. Still, they'll work fine - as I said earlier,
we have open loop gain to burn.

[Nelson Pass]

The IRFP240's have a lower thermal coefficient than the Fairchild parts in the
F5, so that is plausible.

[labjr]

How come all the semiconductors we want to use have been discontinued but the IRFP240s are still being made?

[lhquam]

I think EUVL's F5X uses Toshiba 2SK1530/2SJ201 MOSFETs.

I am considering Fairchild FQA28N15 MOSFETs in the F6. Using the Fairchild PSPICE models I see some rather strange results for the 3rd harmonic shown below. I do not understand why H3 is so flat over frequency. Until I measure in an actual circuit I cannot trust the FQA28N15 SPICE models. I have been able to get pretty good correspondence between SPICE and measurements for the SemiSouth R100s.

Originally Posted by Nelson Pass

The IRFP240's have a lower thermal coefficient than the Fairchild parts in the
F5, so that is plausible.

[buzzforb]

Ilquam,
I might have some of those as well as the FQA 32n20c, if you want them. I looked at them as alternatives to the 19n20c in the GB.

[lhquam]

Thanks for the offer, but I purchased 50 each FQA28N15 and FQA36P15 a couple of years ago and measured them for Vgs and Gm at 1.3A. If you need some matched pairs, I have them.

BTW: The SPICE sim show a transconductance of 6.6S at 1.4A for the FQA28N15 vs. 7.8S for the SemiSouth R100s. Pretty good!

Originally Posted by buzzforb

Ilquam,
I might have some of those as well as the FQA 32n20c, if you want them. I looked at them as alternatives to the 19n20c in the GB.

[buzzforb]

Reason i mention FQA32n20c is that i believe it is higher.