OK, we all know about the post-PI cathode follower for increased drive current.....

[redelephant]

Obviously I'm bored at work today...usually dont have time for this.

Now ur last post is more agreeable I dont particularly like the CF either and seldom use it. (The only place I use it is for driving grids)
...and that brings up some disagreement...
I still believe u are erronious regarding the drive capabilities of CF'ers. I use 6DJ8s as CF and it can give about 20mA to the load while idling at a low 2-3mA.

Your method of driving the input with a low resistance value for less noise is interesting, tho I am not quite convinced even the best low noise tubes used as CF are more silent than a 1meg ohm metal film...Even if it is, to me even the 2megohms I use at the inputs are no real problem. The guitar cable and piskups far outgun the resistors for noise. Besides it will not shunt any external noise to ground at all, if it did, it would also shunt the guitar signal to ground.

You are also wrong about the bandwidth of a CF. It has wider ditto compared to a common cathode gain stage which has large input capacitance due to gain and the opposite swing at the output.

The only reason to bootstrap a CF would be to get real high input impedance, such as for condenser mic capsules, but not necessary for guitar. Bootstrapping doesn't provide more output juice.

Hehe, thread really took off here, and to try to get it back: Paralleling two triodes per phase in the PI will only double the 'driveability' which isn't enough to make a difference. Perhaps using parallel 6SN7s and running them at their limit will make a noticable difference compared to a 12AX7 running at 1mA. But u need more than 10mA to drive grids positive enough to make a difference. I think...havent tried.
Use an 4:1 interstage tranny driven by 6SN7 or 6DJ8 and u may actually put a watt or two into the grids.

[Amp Kat]

Obviously I'm bored at work today...usually dont have time for this.

Now ur last post is more agreeable I dont particularly like the CF either and seldom use it. (The only place I use it is for driving grids)
...and that brings up some disagreement...
I still believe u are erronious regarding the drive capabilities of CF'ers. I use 6DJ8s as CF and it can give about 20mA to the load while idling at a low 2-3mA.

Your method of driving the input with a low resistance value for less noise is interesting, tho I am not quite convinced even the best low noise tubes used as CF are more silent than a 1meg ohm metal film...Even if it is, to me even the 2megohms I use at the inputs are no real problem. The guitar cable and piskups far outgun the resistors for noise. Besides it will not shunt any external noise to ground at all, if it did, it would also shunt the guitar signal to ground.

You are also wrong about the bandwidth of a CF. It has wider ditto compared to a common cathode gain stage which has large input capacitance due to gain and the opposite swing at the output.

The only reason to bootstrap a CF would be to get real high input impedance, such as for condenser mic capsules, but not necessary for guitar. Bootstrapping doesn't provide more output juice.

Hehe, thread really took off here, and to try to get it back: Paralleling two triodes per phase in the PI will only double the 'driveability' which isn't enough to make a difference. Perhaps using parallel 6SN7s and running them at their limit will make a noticable difference compared to a 12AX7 running at 1mA. But u need more than 10mA to drive grids positive enough to make a difference. I think...havent tried.
Use an 4:1 interstage tranny driven by 6SN7 or 6DJ8 and u may actually put a watt or two into the grids.

Have to disagree with you on the bandwidth and to each his own but I get much better tone using plate coupled eq. The variable loading on the gain stage produces a wider range. Like I said if you want to use them then fine have at it and I've also used the CF very effectively to shunt noise in the 1st stage and it works great. Now if you can get someone in the crowd where I play at to say hey that guys amp is using a 1 meg instead of a CF follower and is driving those grids into AB2 I'll be much more impressed.

[redelephant]

It appears u are talking in practical terms, as in how it sounds in use. I agree a Cf has little merits for tone, and I also prefer a plate couple tone stack.
But technically, I'm afraid ure wrong, and u cannot say a CF has worse bandwidth b/c u get better tone from a tonestack when coupling to the plate.
...this is rambling on..I'm done with this...I think...

[Steve Conner]

I use 6DJ8s as CF and it can give about 20mA to the load while idling at a low 2-3mA.

It may be able to source 20mA, but obviously it can only sink 2-3mA at most. This is what I meant by large-signal behaviour earlier. Not an issue when driving power tube grids, because they only conduct in one direction. But in the general case, it can catch you out.

[Chuck H]

You could get AB2 with AC coupling by making the drive very powerful and the bias feed resistors really small.

This is a second on something I posted earlier:

or an obscene amount of gain coupled to a very small load resistor

In this case the "load" would be the bias feed (for fixed bias) or 0V reference (for cathode bias). Not an easy order with octal tubes but with el84's which need very little drive voltage it may be entirely possible to over gain a LTP PI and then dumb it down with small bias resistors to get the desired effect. If I weren't mathematically challenged I'd do some formulas on it. But actually trying it would probably tell any of us all we need to know.