Mu-follower vs common cathode w/- active load

[jnb]

I have been contemplating some mu-followers and their troublesome associated heater supplies. I have been planning to use a transistor to increase the signal to the upper triode ala Morgan Jones'.

If the upper section is really behaving as a constant current source, I figured I may as well just use a diyAudio CCS on top af a common cathode stage.

My question is, will there be a noticeable difference? In sound for example? (The load will be a simple one, suitable for a mu-follower) The CCS will be working from its own reference as opposed to the mu-follower's trying to oppose current changes in the stage and if I really want to make the ultimate mu-follower, wouldn't it be by using a dedicated CCS?

[ray_moth]

Have you seen Alan Kimmels' article on mu stages? He goes into the use of a MOSFET as an active load.

[jnb]

Yes, thanks, I have. Now that does seem like the ultimate mu-follower. Either way, though, it is still a mu-follower.

I am conflicted about using the CCS though. It is referenced to the B+ rather than tracing the signal which is what makes it different, but at least that way the CCS is independent and can be optimised separately. I doubt PSRR will worsen doing this as the impedances involved are very high.

It also seems to be taking the easy way out? Unless others have had good results with it.

[SY]

With a good, linear triode, CCS loading will show minimal distortion BUT will have a much higher source impedance and be more sensitive to loading than a mu follower. In practice, a CCS-loaded common cathode stage will either drive an impedance that's an order of magnitude higher than the plate resistance or it will be buffered by a follower.

[Gluca]

You can use a CCS (I mean a solid state CCS) and take the output from the upper cathode. You don't have that coupling cap, the lower tube works under the high CCS load and your output is buffered (and low impedance) as you are stepping out from the CCS. I had quite a good success this way with a simple IXYS 10M45S.

It sounds good too.

Ciao
Gianluca

[jnb]

Originally posted by SY
With a good, linear triode, CCS loading will show minimal distortion

Good

BUT will have a much higher source impedance

Aha! There's the catch I was waiting for!

Still though, I think I can spice it and adjust it to suit my needs. I think it'll still be alright.

Originally posted by Gluca
You can use a CCS (I mean a solid state CCS) and take the output from the upper cathode.

This is a lot like what I was already thinking of doing only supersized, I mean it could be. I could build a powerful CCS and connect it within the mu-follower, or directly to B+ as above. Kinky

[EC8010]

A mu-follower is a gain stage and a cathode follower connected in such a way that the cathode follower provides a CCS load to the gain stage, making it much more linear. The downside is that the cathode follower is in series with the gain stage and consumes HT voltage. If you don't combine them in this way, but use a semiconductor CCS load for the gain stage, you get the same distortion etc as the mu-follower but a higher maximum output swing for a given HT voltage.

[jnb]

Originally posted by EC8010
A mu-follower is a gain stage and a cathode follower connected in such a way that the cathode follower provides a CCS load to the gain stage, making it much more linear. The downside is that the cathode follower is in series with the gain stage and consumes HT voltage. If you don't combine them in this way, but use a semiconductor CCS load for the gain stage, you get the same distortion etc as the mu-follower but a higher maximum output swing for a given HT voltage.

Fortunately I have 500V to use up so may even need to burn some. Luckily, without the extra heater supply I have a spare heatsink on my chassis to use with this stages regulator.

So my only choice is N-channel or PNP. Silicon is easy and cheap, I'll have to try both