I have idea - ask AW for copy of his Tube Preamp CookBook ;
after reading that , you'll know much more ..... (I didn't , but I spent many hours looking on his schematics ...... and that certainly not mean that one day I'll not buy his book )
so decision when to use tube and when SS part will be much easier .
as Revintage said - with few SS parts it's much easier to make decent CCS .
if not - then you must use pentode , not triode .
in Allen's book you'll find one jewel - SLCF ..... which is probably explained in details and why is superior to plain cathode follower
after reading that , you'll know much more ..... (I didn't , but I spent many hours looking on his schematics ...... and that certainly not mean that one day I'll not buy his book )
so decision when to use tube and when SS part will be much easier .
as Revintage said - with few SS parts it's much easier to make decent CCS .
if not - then you must use pentode , not triode .
in Allen's book you'll find one jewel - SLCF ..... which is probably explained in details and why is superior to plain cathode follower
Am I right when I assume you want to run your GC inverted and have the CF to isolate the input from some lowZ filters at the inverting input?
What is full tube sound? If it is a positive remark it will still not happen when adding the CF to your GC. When done right it should be transparent😉.
What is full tube sound? If it is a positive remark it will still not happen when adding the CF to your GC. When done right it should be transparent😉.
A small step for you is a big step for me so full tube sound is 100% tubes in front of the LM3886 
My design idea is to develop Joe Rasmusens
JLTi MK2
and make it adopted for the forum users but also to give some poor members their pcb's which they have paid for. See my sig about group buy round 2.
My design idea is to develop Joe Rasmusens
JLTi MK2
and make it adopted for the forum users but also to give some poor members their pcb's which they have paid for. See my sig about group buy round 2.
I use DC on my twin triode heaters and there are zero hum issues or any need to elevate the heater windings.
Triodes can make acceptable current sources if you bias then as you would a BJT or an enhancement mode MOSFET for a CCS; with a grid voltage reference and cathode degeneration. The plate impedance is roughly multiplied by the number of Vg-k 's dropped across the cathode degeneration.
For a flexible power-on /heater warm up / de-thump timer you can’t beat a small PIC uC or similar.
Triodes can make acceptable current sources if you bias then as you would a BJT or an enhancement mode MOSFET for a CCS; with a grid voltage reference and cathode degeneration. The plate impedance is roughly multiplied by the number of Vg-k 's dropped across the cathode degeneration.
For a flexible power-on /heater warm up / de-thump timer you can’t beat a small PIC uC or similar.
in any case - to each his own ;
if we share what we know , there is always possibility to try new ( or just forgotten ? ) things ..... then decide what's your cup of Tea .
my fave circ for these de-anythings is plain darlington with RC in base ;
but - that's just mine fave
if we share what we know , there is always possibility to try new ( or just forgotten ? ) things ..... then decide what's your cup of Tea .
my fave circ for these de-anythings is plain darlington with RC in base ;
but - that's just mine fave
GK,
You wrote: "I use DC on my twin triode heaters...or any need to elevate the heater windings"
Why do you say that? I know of no "need" to elevate the heaters - except that 30 years of experience has proven to me that preamp stages simply sound better when the heaters are well positive to their respective cathodes.
Don't "believe or not believe" me, try it!
Regards, Allen
You wrote: "I use DC on my twin triode heaters...or any need to elevate the heater windings"
Why do you say that? I know of no "need" to elevate the heaters - except that 30 years of experience has proven to me that preamp stages simply sound better when the heaters are well positive to their respective cathodes.
Don't "believe or not believe" me, try it!
Regards, Allen
Yeah I know, and in all circumstance series regulators are a waste of time a money and LM394's sound so grainy one's missus can hear the difference from the other room.
PA,
The correct formula for calculating the impedance of a triode CCS is
Ri´=Ri+Rk(mu+1). So in your case, if we assume Ua=35V, Ri=6k, mu=30, Ug=-1V and Ia=3,5mA after consulting the E88CC specifications, it will be:
Ri´=6+(1/3,5)(30+1)~15k.
If you use a cathode resistor instead of the tube CCS it will be in the ballpark of 10k. Big difference..........
If you go for Ua=90V, Ri´ will be almost doubled.
Another solution could be adding positive gridbias as a higher value Rk would be needed. +5V would give something like 60k. This is in theory, not sure if it will work IRL.
If you want to do something extra, use DC-coupling with a servo around the CF and move the couplingcap to the input where it can be made much smaller. Also go for a sand CCS. To get some ideas, why not check SY´s "Heretical Preamp" minus input transformer.
The correct formula for calculating the impedance of a triode CCS is
Ri´=Ri+Rk(mu+1). So in your case, if we assume Ua=35V, Ri=6k, mu=30, Ug=-1V and Ia=3,5mA after consulting the E88CC specifications, it will be:
Ri´=6+(1/3,5)(30+1)~15k.
If you use a cathode resistor instead of the tube CCS it will be in the ballpark of 10k. Big difference..........
If you go for Ua=90V, Ri´ will be almost doubled.
Another solution could be adding positive gridbias as a higher value Rk would be needed. +5V would give something like 60k. This is in theory, not sure if it will work IRL.
If you want to do something extra, use DC-coupling with a servo around the CF and move the couplingcap to the input where it can be made much smaller. Also go for a sand CCS. To get some ideas, why not check SY´s "Heretical Preamp" minus input transformer.
Heater elevation has a definite function, once you stop paying attention to the idealized models and start looking at what actual hunks of metal and carbon do.
The heater-to-cathode impedance is not infinite and especially isn't so when the tube warms up or as it ages. Leakage currents can definitely affect the performance.
Common mode noise is easily coupled to the cathode if the circuit topology does not have the cathode pinned firmly to ground (e.g., the upper cathode in Allen's cascodes). And regulating the heater DC does not eliminate the common mode noise unless you use common-mode regulators (rare). Even then, the HF components will break on through to the other side because of the necessary 6dB/octave reduction in open loop gain of the error amp.
Having a heater voltage derived from a winding on the same transformer as the HV winding, very common practice, will almost ensure that there will be noise coupled to the heater supply.
Nope, there's a reason that heater elevation has been standard engineering practice since the 1930s, even by hard-nosed engineers. Even Australian engineers (RDH4!). Allen's observation is certainly plausible and should not be dismissed out of hand.
The heater-to-cathode impedance is not infinite and especially isn't so when the tube warms up or as it ages. Leakage currents can definitely affect the performance.
Common mode noise is easily coupled to the cathode if the circuit topology does not have the cathode pinned firmly to ground (e.g., the upper cathode in Allen's cascodes). And regulating the heater DC does not eliminate the common mode noise unless you use common-mode regulators (rare). Even then, the HF components will break on through to the other side because of the necessary 6dB/octave reduction in open loop gain of the error amp.
Having a heater voltage derived from a winding on the same transformer as the HV winding, very common practice, will almost ensure that there will be noise coupled to the heater supply.
Nope, there's a reason that heater elevation has been standard engineering practice since the 1930s, even by hard-nosed engineers. Even Australian engineers (RDH4!). Allen's observation is certainly plausible and should not be dismissed out of hand.
Exactly!
From another angle - when I experiemented with SRPP circuits, I used the same tube top & bottom. It sounded good for about 3 months, then rapidly got much worse. A new tube solved the sound...for another 3 months.
I can only put that down to heater cathode leakage. But if the upper tube was actually another tube, with it's own heater circuit, and lifted around +30 to 50V higher than the cathode V, I doubt if this 3 month tube failure would occur. I never did try that and went the SLCF route.
Unfortunately, this heater complexity make the very simple SRPP circuit far more complex.
Regards, Allen
From another angle - when I experiemented with SRPP circuits, I used the same tube top & bottom. It sounded good for about 3 months, then rapidly got much worse. A new tube solved the sound...for another 3 months.
I can only put that down to heater cathode leakage. But if the upper tube was actually another tube, with it's own heater circuit, and lifted around +30 to 50V higher than the cathode V, I doubt if this 3 month tube failure would occur. I never did try that and went the SLCF route.
Unfortunately, this heater complexity make the very simple SRPP circuit far more complex.
Regards, Allen
In my experience I must agree with Allen Wright and SY about the lift.
When I employ the SRPP I always make a stereo version using the two halves (one for the right and the other for the left) as lower tube and the two halves of another tube as upper tube. Each tube has its own filament supply elevated respect to its cathode. In this way there are zero problems and IMHO is the only way to get the real performance of the SRPP.
The filament supply is a black beast in SRPP circuits.
Often I read about people having more distortion from a SRPP than its single ended version. Then looking at the actual implementation they do not care about the filaments! They use just one winding!! ( lifted halfway between the cathodes in best cases).
When this is not possible I parallel the halves and use four tubes! Fortunately in those circumstances it always happen to be no-compromise stuff.....
Cheers,
45
When I employ the SRPP I always make a stereo version using the two halves (one for the right and the other for the left) as lower tube and the two halves of another tube as upper tube. Each tube has its own filament supply elevated respect to its cathode. In this way there are zero problems and IMHO is the only way to get the real performance of the SRPP.
The filament supply is a black beast in SRPP circuits.
Often I read about people having more distortion from a SRPP than its single ended version. Then looking at the actual implementation they do not care about the filaments! They use just one winding!! ( lifted halfway between the cathodes in best cases).
When this is not possible I parallel the halves and use four tubes! Fortunately in those circumstances it always happen to be no-compromise stuff.....
Cheers,
45
Thanks, but I did not say that doesn't have a definite function, I simply put forward an instance where I think it is unnecessary.
With my heaters (12AX7, 12AU7, 12AT7, 12AY7, 6922, etc ) operating on DC heater supplies referenced to ground (or a respective negative supply rail – no “common mode” noise problems here that would be any different if there was a biasing voltage between the heater supply and the rail used as a reference) there simply isn’t any coupling of hum that I can detect.
As for leakage currents – got any measurements? I’d be especially interested in just how much cathode-to-heater leakage could be induced by a cathode-to-heater potential of a couple of volts.
If somebody is going make the blanket claim that any circuit made this way is going to "sound better" if done with elevated heaters, then that’s fine, and you can call me dogmatic for being dismissive on that one for all I care.
Last edited:
This is what I was getting at, and it does work - have simmed it, built it and measured it and, shock! horror!, performs just as calculated and as confimed in LTspice
Both a decent plate supply and a grid biasing potential are required here. I generally like to cathode degenerate about at least ten times Vg-k.
I have two choices:
1 tube one for the cathode follower with it's own heater power supply, lifted 30 V, decoupled to ground. Tube 2 as a constant current source, also with it's own heater power supply, lifted 30V above -50V decoupled to ground. Each tube handles two channels
2 Only one tube for the cathode follower, lifted heater PS and semiconductor CCS.
I will have separate transformers for +-50V and the heaters. I want also a not too big pcb.
1 tube one for the cathode follower with it's own heater power supply, lifted 30 V, decoupled to ground. Tube 2 as a constant current source, also with it's own heater power supply, lifted 30V above -50V decoupled to ground. Each tube handles two channels
2 Only one tube for the cathode follower, lifted heater PS and semiconductor CCS.
I will have separate transformers for +-50V and the heaters. I want also a not too big pcb.
If memory serves, there's discussion and results in Morgan Jones's book. There's also considerable discussion in Tomer. I haven't measured it because I design my circuits to avoid the issue in the first place.
If you reread my post, you'll see that hum was not the central thrust- it's HF noise that's the issue.
OK, no measurements then, fine. As for HF noise, well I simply don't have enough of that making it past my heater supply regulators and output bypassing and if I did the interelectrode (and circuit) capacitances would be a much more effective coupler of it into the audio signal than any pA (or whatever) leakage path between the heater and cathode.
- Status
- Not open for further replies.