augmented cathode follower?

[Zen Mod]

Originally posted by Tony
hi,

what about cathode followers with constant current sinks in place of the cathode resistors?

anybody played with them?

Wright (read - Tectronix) fer instance

[SY]

Tek actually used a three triode stack- a cathode CCS, the CF itself, and the upper bootstrap tube. They did a compensation trick there that I don't fully understand- I'll see if I can get a scan of a schematic posted.

[Tony]

reason i asked is i have these 6336 tubes that i would like to do a CF power output using an MJW18020D as current source, i reckon this trannie has morre than 30watt SOA at Vce of around 150volts.......

actualy weighing in heatsink size, versus a PL504 tube as current sink...

i thought the sandstate sink is cool..

[Fuling]

Seems to me that a high gm pentode would be a perfect candidate for the follower job. Even if we wire it as a triode it should behave like a pentode thanks to the constant voltage on both screen grid and plate.

[tubelab.com]

reason i asked is i have these 6336 tubes

I have experimented with CCS circuits in the cathode of a CF. The CCS will keep the current through the CF constant as long as the AC load is light. It also helps with the asymetrical drive issue that a CF has since the CCS can sink a constant current up to its preset value. When you hang a heavy AC load on the CF (isn't that what we want them for?) the load line is no longer flat. A CCS loaded CF works well for driving an output tube grid where the current demands are also asymetrical.

When I started down this CF output stage path six or eight months ago I new that I was going to use some type of constant voltage circuit. Thus I needed to find a tube that would work well with a fixed low voltage across it. I also wanted a tube that could handle a lot of current (hundreds of mA) so that I could use a low impedance OPT (600 ohms) and eventually do an SE OTL. I set up a test fixture and ran through a bunch of different tubes, captured data, and plotted the linearity VS current for a fixed voltage. The clear winner for this application was the 6336A. In second place a triode wired 6LW6. I did not test the Russian 6C33C since I do not have suitable sockets.

If you want to experiment with CF output stages using 6336's stack your follower on top of the output CF and do an "augmented design". If you are not in a big hurry, I will post my schematics after I get it all working correctly. There will be a version with 6336's, but I will probably use a mosfet for the top device due to SOA failures with BJT's. I plan to run the total B+ in the 350 to 400 volt range to achieve 20 to 30 watts.

I am enclosing the ACF-2 circuit from the Macdonald publication. I don't have much time since I am leaving town on the 26th but I have built one of these with the following changes. I substituted a CCS IC for V5, V1 - V2 is a 6SN7, and V3 - V4 is a 6AS7. It works, but I have not had time to do much testing. I could not get it to work with a 6SL7 for V1 - V2.

[Shoog]

what about cathode followers with constant current sinks in place of the cathode resistors?

anybody played with them?

I built a preamp based on Allen Wrights FVP5. It uses a simple triode as a primitive CCS, and a mosfet as the Constant voltage element on top. Its really quite simple to implement and I noticed improvements on each stage of its evolution from a simple CF buffer stage. Allen Wrights offerings are highly respected, and my version was very well received at a recent meet.
Allen Wright claims that its real sonic benefit is in removing the feedback element of the cathode resistor in a normal CF. Don't know myself

There is a brief mention of the circuit over on the Tubcad archive, he implements a very interesting low voltage version using ECC86's, I think the B+ is only 24V. This removes practically all the complexity associated with this circuit approach and because of the extremely linear nature of this circuit - I would feel confident that it would give up little to a higher voltage version.

I intend to do further refinements but am convinced in the approach. It presents an extremely uncoloured but mellow sound - which is exactly what I expect from valves.

Shoog

[David Davenport]

Originally posted by SY
Tek actually used a three triode stack- a cathode CCS, the CF itself, and the upper bootstrap tube. They did a compensation trick there that I don't fully understand- I'll see if I can get a scan of a schematic posted.

Were you able to find that Tek schematic?

Dave

[smoking-amp]

Maybe this?

[SY]

Sorry, Dave, I forgot. I've got it in my hands and will scan and post.

[David Davenport]

That is an interesting circuit, what is the source? However, it isn't the Tek circuit that SY mentioned. From SY's post, it was a lot more complex.

Dave