• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Long-Tailed Pair, pse. share your views...


I'm in the process of building a Stereo EL34PP amp.
Actually the very first tube-amp I'm putting together and if it sounds well enough, it will also be the last :)
That's why I try to do 'do everything right' which of course won't be possible.
There seems to be as many 'right' solutions as there are constructors :D
Anyway, I've started on firm ground buying Lundahl Transformers and Winged =C= EL34's. That should set me off in the right direction I suppose.
I've also invested in a couple of Tungsol EF806SG which may turn-out more dubious. At least they SHOULD be less noisy that regular EF86's and of course especially run as triodes but I really can't say yet.
What I'm here to ask is more your opinions on the Long-Tailed Pair. I've come across a few NOS ECC808's that I'll try to use here. Since I understand a high AC-impedance of the Tail-resistor is required to balance the output reasonably I've decided to go for a Constant-Current-Source, see the attached pdf.
BUT, IF it's such a great idea to use a CCS, why isn't it done more widely?
Theoretically the AC-impedance of the CCS will be something close to:
R-Emitter x hFE x 2 -> 1000 x 160 x 2 = 320K-Ohm.
First of all; is this equation correct?
If it is is running the ECC808's at 1.5mA CCS reasonable?
What's your opinion on the construction in general?

Thanks in advance,




  • EF806_ECC808.pdf
    12.9 KB · Views: 80
  • ZTX652_ZTX653.pdf
    61.6 KB · Views: 65
Last edited:


diyAudio Moderator Emeritus
2002-10-24 10:19 pm
Since I understand a high AC-impedance of the Tail-resistor is required to balance the output reasonably I've decided to go for a Constant-Current-Source, see the attached pdf.
BUT, IF it's such a great idea to use a CCS, why isn't it done more widely?

Inertia. Incompetence. Superstition. A CCS is absolutely the right way to go for an LTP.

Noise is generally not an issue for power amp input tubes.
Thanks, it's what I feared ;)
Don't rock the boat, stay on the straight and narrow... etc.
I really couldn't see any reason why NOT to use the CCS.
Probably there are more views on the matter.
What about 'musicality', 'fidelity', 'tube-sound' and the rest of the boss-words?

A CCS is absolutely the right way to go for an LTP.

I built my Baby Huey with a resistor in the tail as I didn't have all the parts for Gingertubes ccs. When the ccs was finally added it made imho a noticeable increase in the background detail. Of course it may only have straightened out some non-linear aspect of my personal frequency range - but what do I care!

My opinion on LTP biasing is ccs usually is the best solution. On some circuits you don't want fully symmetric output, and in those cases it may be better with a simple resistor. For instance on totempole output following a PI you want one phase a little smaller than the other, and a resistor will be better than a CCS.
As it turns out, when the resistance of the tail goes over a certain value, a CCS really doesn't provide much benefit. I aint smart enough to figure it out in my head right now without looking up in the books first, but with a 6922 LTP a tail over 10kohm, or for petimeters say 20kohms, a CCS won't do much other than complicating your effort.
Of course, if you have limited voltage for the tail, a CCS is always the best way.
Just wanted to put my pennies in because it's not always lazyness or incompetence that lays behind the reasons for not using CCS.

Would not change the LTP/CCS for anything in that design. It is a very musical/dynamic amp (Are those the words you are looking for?)

WoW :hypno2: , it took a while to read through all those postings.
VERY nice work. It shows how new parts enhance the vintage ones nicely.
Perhaps the EL34PP won't be my last amp after all... :D
Another aspect of building a nice amplifier is of course the mechanical details.
In your case they're superb! Nice and classy!
My EL will go on a laser cut steel-chassis painted into very deep hi-gloss red, almost black with hi-gloss black wooden side panels. Cover-shield for the transformers in steel painted into the same hi-gloss black. Tubes standing free.
Well, at least it's my INTENTION to have it like that and I know a lot of people in the right places to make it happen :D
It will also be nice to be able to document the working of the amplifier, but I just have basic tools like a scope, a generator and a dummy-load. So far no distortion-analyzer as they don't come cheap around here; working or not.


Not to offend LTP fans but, sound quality wise how does it compare to another popular splitter the (concertina, cathodyne, split phase)? Sometimes called a Simplified Williamson. One member said he prefered it for more it's detailed sound. Maybe there are disadvantages that could be fixed with some sort of CCS?

A third type that seems to be liked is the Floating Paraphase. How too does it compare?

Both seem to be popular on low power amps.

A CCS is ideal, but no real CCS is actually ideal. A resistor is simpler and cheaper. Whether it is good enough depends on its value, and which valve you use. The aim is to get the common-mode gain as low as possible, when compared with the differential-mode gain. Diff-mode gain is a bit less than mu. Common-mode gain is about RL/(2Rk), and is likely to be somewhere in the region of 1 if a tail resistor is used, or 0.1(?) with a CCS. If you are happy with a few percent error (should be good enough - some people seem to deliberately design-in up to 10% error using other phase splitters), then you need mu bigger than around 30-40 for a resistor or 3-4 for CCS.

So if you want to use a resistor, use 12AT7 or 12AX7 as LTP. If you want to use 6SN7 or other low-mu triodes then you need a CCS, unless you like second-order distortion. 6DJ8 is right on the boundary: resistor OK, CCS slightly better.