Ideally, yes... but as Shoog said the dual CCS is a good workaround. And Broskie has written a bunch on the subject, e.g., https://www.tubecad.com/2011/08/blog0210.htm
The pot should have a low value, say 10Ω, which would not degrade the differential function (much). Not sure you really need a bypass cap, as it is only used to set the idle bias.
The two CCS's allow the cathodes to float to whatever voltage allows the passage of the DC currents. This means the cathodes will likely be at different bias values. This is not particularly important since the back to back caps will prevent a DC flow from one cathode to the other. Think of the two halves as been isolated for DC purposes.
The high impedance of the CCS prevents any AC signal passing to earth through them - the only path remaining for the AC signal is the relatively low impedance presented by the back to back capacitors and the plate impedance of the opposite valve (and the very low impedance of the PSU). Since the signal going down from one side can only ever go up the other side the signal must be absolutely equal (but opposite) between the two valves at AC. Its a tug of war with no slack in the rope and the overall signal is a compromise between the signal trying to be amplified by each half of the PP pair.
This is the very definition of a LTP.
There literally cannot be any other behaviour and the effect is to squeeze out of the equation nearly all second harmonic distortion. This amp will not sound like an SE.
Let me just say that this back to back arrangement originates from the venerable old member of the forum Brian Beck and not myself.
Shoog
The high impedance of the CCS prevents any AC signal passing to earth through them - the only path remaining for the AC signal is the relatively low impedance presented by the back to back capacitors and the plate impedance of the opposite valve (and the very low impedance of the PSU). Since the signal going down from one side can only ever go up the other side the signal must be absolutely equal (but opposite) between the two valves at AC. Its a tug of war with no slack in the rope and the overall signal is a compromise between the signal trying to be amplified by each half of the PP pair.
This is the very definition of a LTP.
There literally cannot be any other behaviour and the effect is to squeeze out of the equation nearly all second harmonic distortion. This amp will not sound like an SE.
Let me just say that this back to back arrangement originates from the venerable old member of the forum Brian Beck and not myself.
Shoog
Last edited:
So, the two CCS's are set for equal currents so the OPT sees balanced DC. Clear now! Thanks Shoog!
The problem is that there is never a perfect solution for anything. Bypassing the CCSes, capacitors are introduced back, which Gary spent so much effort to eliminate.
Faith isn't really the issue.
They were an engineering firm and they would always have produced the product that produced the best measurable result. If they had a choice between a choke and a CCS they would have measured the result and chosen the CCS. They didn't have a CCS so they chose a choke over the inferior and wasteful alternative - a resistor.
There is no magic or romance in this and it isn't a ******* match - these are hard engineering decisions and the reason why Gary Pimm spent years perfecting the highest possible performing CCS's for audio use. If he hadn't there would be no Tabor design.
Chuck that away for the romance of Iron if you like.
Shoog
I am not necessarily advocating this approach above a single CCS, but at least you now under4stand why it works. For me the ability to have an adjustment free amp far outweighs any theoretical drawbacks.
Make an informed decision based upon what works for you.
Shoog
idle output tube bias balancing is done via the input long tail pair G2 adjustments, and resistors at the plates of the 1624 are used to monitor plate currents
the reason i thought about separate CCS's is to avoid possible current hogging,
and that random tube selection can be done rather than selecting tubes that are reasonably well matched......i could be wrong...
BTW, the last news I've heard on Gary's progress with this design was that he replaced pentode driver with MOSFETs. I guess perfection will be achieved when output tubes are also replaced with MOSFETs.
Indeed and from what I heard it sounds significantly better than the tube version.He attributes this to a lack of physical resonance in the tubes.
The only issue for me is that those SS pentodes are quite complex to build and need big heatsinks.
Shoog
Yes, it's not a typical SS build, and might be even more expensive than the tube version.
i have made up my mind, i am using a cathode resistor bias on the 1624's....
paralleled 2 x 1k ohm resistors by passed by a 470uf/250v ecaps...
i will mount them on pcb's.....
for the input ltp's i have not decided yet on the 6BC5 or the 6GK5's. why those?
i have lots of them to choose from...
paralleled 2 x 1k ohm resistors by passed by a 470uf/250v ecaps...
i will mount them on pcb's.....
for the input ltp's i have not decided yet on the 6BC5 or the 6GK5's. why those?
i have lots of them to choose from...
- Status
- Not open for further replies.