diyAudio (
-   Tubes / Valves (
-   -   FET Source Follower Distortion (

TheGimp 31st December 2012 06:37 PM

FET Source Follower Distortion
2 Attachment(s)
Over the holiday I have taken off on another tangent instead of building an amp.

I was trying to get better measurements of tube distortion so I can sort the used tubes I have on hand. I'm focusing on triode strapped pentodes including 6CB6, 6EJ7, 6J51P, 6JC6 and 6J9P.

I built a test box with :
(1) FET with Zener regulator to help filter the B+ comming in
(2) Switch Selectable tube anode loads of FET/Bipolar current source, Cascode DN2540 current source, and Gyrator.
(3) Fully bypassed adjustable cathode resistor.
(4) A Current Source Loaded FET Source Follower to buffer the output signal.
(5) DC input for cathode heater supply

This combination allows me to get measurements that seem to be 20dB better than I had previously made. The two screen shots show examples of previous and current measurements.

Is the data presented in the second FFT realistic for such measurements?

Is the distortion level in line with what should be expected for a current source loaded triode strapped pentode with fully bypassed cathode resistor?

Is the background noise level what should be expected?

In other words am I getting true representative data or do I still have a way to go to be able to accurately measure tube distortion and S/N ratio?

mogliaa 31st December 2012 07:19 PM

Hi Steven,
I've been using a similar test circuit based on DN2540 cascoded CCS and fixed bias. I had some ground loops issues which blew some of the interface ICs from the pete millett's interface and still got some 100Hz level a bit higher in my perspective. Otherwise the noise floor is below 90dB and generally get 80-90dB SNR. Either way, it's very useful to identify misaligned electrodes in DHTs and refine operating points....

Happy New Year!

TheGimp 31st December 2012 08:08 PM

5 Attachment(s)
The title to this thread is because I found that for every current source I tried, and for every FET I tried, Source Follower induced distortion was inversely related to current.

At 3 mA and below, current source induced FET distortion dominated tube distortion!

At 2mA second harmonic is at about -74dB wrt 2.82Vrms 0dB ref.

At 20mA second harmonic is down below -100dB, a difference of over 25dB!

mogliaa 31st December 2012 08:19 PM

Why not trying a FET (e.g. 2sk170) for lower currents?

TheGimp 31st December 2012 09:29 PM

I tried a ZVN0545 and got the same results at 3mA.

I tried pairing it with a LND150 current source at 1.1mA(max I could get) and got high distortion as well.

TheGimp 31st December 2012 09:35 PM

Happy New Year in 1.5Hrs...!

mogliaa 31st December 2012 09:39 PM

All are mosfets. I was suggesting a jfet of low current, I recently tested a 2sk170 gyrator on a cx301a running at 3ma with superb results...
Have a great new year!!

adason 31st December 2012 09:41 PM

why is this in tubes?

TheGimp 31st December 2012 10:17 PM

Because we use Source Followers made of MosFets with current sources as the driver stage of tubes.

I believe they may be a significant a source of distortion.

mogliaa 31st December 2012 10:24 PM

I'm working on my driver for the 4-65a SE amplifier. It's a 46 (triode-strapped) in filament bias loaded with a single IXYS MOSFET. It should get the best out of the 46. The spice simulation gets 0.15% at 200Vpp. Currently I tested it on the bench with limited tester signal which gives 0.04% at 10Vrms.
Gyrators in mu-follower are fantastic as drivers specially in class A2. Will give the flatest load line, minimum distortion and lowest output impedance...which is a plus for an A2 driver. In this case, the FET distortion is minimum given the circuit used....

All times are GMT. The time now is 02:09 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright 1999-2015 diyAudio

Content Relevant URLs by vBSEO 3.3.2