CF output stage - float heater or tie to cathode?

[geezertron]

Is there any consensus on which approach is better for a cathode-follower output stage: tie heater to cathode or let it float? I've been playing around with this design and (21JZ6 triode-strapped + various transformers) and am impressed by the low distortion. All the rest of it seems hard to make work well though, but that's what makes it a hobby.

I've been floating the heater for now.

Thanks...

 

[SpreadSpectrum]

Personally, I'd use a dedicated winding for each output tube and tie to cathode.

I've been experimenting with a circuit for driving a CF output stage if you are interested: http://www.diyaudio.com/forums/tubes-valves/288372-idea-driver-cf-output-stage.html

I haven't built the HV supply to test it up to its full potential yet but I have tested up to 100Vrms so far with very good results.

 

[Johan Potgieter]

Is there any consensus on which approach is better for a cathode-follower output stage: tie heater to cathode or let it float?

I would say neither.

I would elevate the heater to a suitable dc potential and bypass to earth. One always has to keep heater-cathode potential within limits. If output signal still exceeds that somewhere, next step is to connect to the cathode. But then signal may be driven back into the circuit through the common connection. In that case as SpreadSpectrum said, separate transformer heater windings.

 

[Tubelab_com]

I have built some pretty big cathode follower output stages. 30 WPC SE using 6528's and over 100 WPC using triode wired 6LW6's. In all cases I tied the heater to the cathode to avoid the possibility of HK breakdown when the Big Dumb Blonde One plugs his guitar preamp into the amp and sets it on KILL!
I use an Antek toroid with two 6 volt windings to feed the output tubes. This supply is isolated from the heater supply for the other tubes in the amp, otherwise you will violate their HK ratings.

I have used several driver circuits but the simplest is a 6EM7 / 6EA7 with the small triode as a voltage amp (CCS plate load to maximize gain) and the big triode running into a CCS load (if you have 500+ volts of B+ available) or a choke load on 300 volts or so.

The best driver isn't a driver at all it is a unified design called an "Augmented Cathode Follower" Invented and patented by Ross MacDonald in 1957. The augmented cathode follower design can be set up to have voltage gain. In my simulations the design becomes unstable at gains above 2, but I have seen gains as high as 4 in real world circuits. See these threads for more info.

http://www.diyaudio.com/forums/tube...wer.html?highlight=augmented+cathode+follower

http://www.diyaudio.com/forums/tubes-valves/105859-20w-cathode-follower-amplifier.html

 

[Merlinb]

"Augmented Cathode Follower" Invented and patented by Ross MacDonald in 1957.

I'll post the actual link, shall I?
http://www.pearl-hifi.com/06_Lit_Ar...04/Sec_19/955_Augmented_Cathode_Followers.pdf

 

[DF96]

A heater circuit should never be left to float. It must always have a potential established via a resistance not exceeding the value given in the datasheet. If not specified, assume a few 10's of k as maximum.

 

[Tubelab_com]

I'll post the actual link, shall I?

MacDonald has written many papers since the 1950's in the fields of electronics, physics, mathematics, solid state physics, and chemistry. As of last year he was still at it. Many of the technical electronics papers, especially the current stuff, were hosted on an IEEE site which required membership for access. We had a blanket license for all access when I worked at Motorola, but I no longer work there. Some are linked from MacDonald's web site.

Ross Macdonald | James Ross Macdonald

This is one of the complex audio amp designs from 1957.

http://jrossmacdonald.com/jrm/wp-content/uploads/026MultiloopAmplifier.pdf

Pre - distortion.....commonly used today in linear cell phone RF amplifiers for LTE.

http://jrossmacdonald.com/jrm/wp-content/uploads/052DistortionReduction.pdf



I got a email from him in 2008 when I published the article in Circuit Cellar magazine (no longer on their web site). He was surprised that people are still exploring such stuff. He stated that he still tinkered with electronics, but his interest have been strictly solid state since the 60's.

 

[geezertron]

Right now it takes a ~300V swing to drive 3W into 8 ohms using a 3.5K Edcore SE xfrmr. The tube is spec'ed for +200V/-100V heater/cathode (I think all compactrons share this spec). Elevation won't work, I'm going to tie it to the cathode for testing, since there's just one tube.

I'm thinking about getting a 2.5K or 1.7K xfrmr and see how it works. This will lower the required swing. I played around with some power transformers with the right ratios and the tube seems to drive them OK. Not sure how low I can go with the impedance and keep the low distortion though.

I'm driving it right now with a jury-rigged setup: 2V sig generator -> SS amp (opamp + driving transistors) -> "backwards" output transformer to get a high enough voltage to drive it. I'm going to investigate some of the referenced postings about real drivers.

Thanks!

 

[SpreadSpectrum]

For my Unity-Coupled amp I got four small Hammond filament transformers, one for each tube. I just tied the center tap for each secondary to the corresponding output tube.