The placebo effect is completely relevant. Just as the medical industry knows that a high proportion of patients will report they are better when (unknown to them) they got an inert pill, so home constructors and modders will sincerely believe some circuit detail made the sound better when in fact it could not possibly do so.
From time to time in the history of audio reproduction, researchers that command considerable respect, eg Olson at RCA, Er Wigan and his team in the BBC Reasearch Dept, and folk at Philips have conducted trials to thow light on just what forms of distortion the human ear can detect, tolerate, and in what amounts. The papers written by these folk are noteworthy for teh considerable effort they went to to avoid placebo effects, non-audio cues, etc.
And then some people actually like small amounts of certain types of distortion...
So be it. I post in good faith, and I generally back up claims or statements I make with technical explanation - unlike some others who just make the claims.
If you find my posts usefull, that's good. If you don't find them usefull, you are free to ignore them.
As I said in an earlier post, no, I haven't tried CCS biasing, simply because I am not in the habit of wasting good time and money on things that have no useful benefit. As I and other have explained in detail, CCS cathode biasing makes amplifier distortion worse.
As far as other sorts of biasing are concerned, I have worked on commercial designs that employ simple common cathode resistor biasing, separate resistors for each cathode, etc.
For my own personal equipment, I have implemented the separately biased AC cross-coupled method I have already described in this and other threads. Carefull comprehensive measurement shows distortion under steady state is least, overdrive recovery is best, and theory shows why.
Last edited:
I use FETs in many places in a tube amp. Plate loads are always MOSFET gyrators.
On occasion, I didn't have sufficient voltage FETs at hand. I had a B+ of 650V on a LTP with CCS tail, and MOSFET plate loads.
The plate load FETs I had at that time were 500V types. I thought, why not try it out anyway, they're cheap parts. My local shop has 500 and 600V FETs for 1.20 € or so. Much cheaper in bulk.
Worked perfectly for a few months in daily use, until I disassembled that experiment.
Now I normally work on the rule that FETs should withstand full B+ even though they might not ever actually see it in their position, and I would advise anybody reading this to do so as well.
But there really are no weird voltage spikes or such. FETs are within their specs super reliable even at high voltages.
In my experience, heat is the biggest problem. And construction errors while building. Correctly applied they work just fine always.
On occasion, I didn't have sufficient voltage FETs at hand. I had a B+ of 650V on a LTP with CCS tail, and MOSFET plate loads.
The plate load FETs I had at that time were 500V types. I thought, why not try it out anyway, they're cheap parts. My local shop has 500 and 600V FETs for 1.20 € or so. Much cheaper in bulk.
Worked perfectly for a few months in daily use, until I disassembled that experiment.
Now I normally work on the rule that FETs should withstand full B+ even though they might not ever actually see it in their position, and I would advise anybody reading this to do so as well.
But there really are no weird voltage spikes or such. FETs are within their specs super reliable even at high voltages.
In my experience, heat is the biggest problem. And construction errors while building. Correctly applied they work just fine always.
Before I respond to this, just to make sure: You are including shared cathode CCS with no bypass (LTP) in this as well, right?
Keit seems to have the impression that MOSFETS are the work of the devil himself!
However--Those that use them Including myself, find them extremely Reliable when implemented correctly with Zener from S-G--as is good practice, and appropriately voltage specced for the job they are doing.
NOTHING wrong with MOSFETS in Tube Circuits at all.....
MOSFETS find use in many sorta 'Tube' equipment commercially--such as Plasma TV sets, where they see hideous Capacitive and Inductive loads--particularly in the ERC circuits and Sustain drives....
Thing that eventually (after say 20K hours ave.) that kills them--is Heat,--Keep 'em cool and they'll easily double and triple this MTBF, but TV set makers are Cheapskates don't fit appropriate heatsinking or cooling measures.......
However--Those that use them Including myself, find them extremely Reliable when implemented correctly with Zener from S-G--as is good practice, and appropriately voltage specced for the job they are doing.
NOTHING wrong with MOSFETS in Tube Circuits at all.....
MOSFETS find use in many sorta 'Tube' equipment commercially--such as Plasma TV sets, where they see hideous Capacitive and Inductive loads--particularly in the ERC circuits and Sustain drives....
Thing that eventually (after say 20K hours ave.) that kills them--is Heat,--Keep 'em cool and they'll easily double and triple this MTBF, but TV set makers are Cheapskates don't fit appropriate heatsinking or cooling measures.......
The quote button is at the bottom of the box. You are clicking on the quote button of the post above. How you're getting Rod's name in there is anyone's guess.
That would be difficult for EL84 output tubes - you would need to find tubes for the previous stages that have heaters rated at around 9 V 50 mA, AND meet the other requirements for the stage.
For large power tubes, the ratings required become eaier, but.....
Tube heaters and filaments have a resistance roughly proprtional to temperature, so there is a degree of constant current effect - but not anywhere near as constant as can be had with even the simplest bipolar of FET constant current source.
Since const currenmt cathode baising increases distortion over that of simple resistor biasing (see early posts #12 by me and #28 by DF96 for an explanation), expect distortion somewhere between the two if using the heaters of previous stages.
Since with previous stage tube heaters used as a CCS, the regulation effect is there but far from perfect, the energisation of the previous stages would fall off as the output tube ages. Under-running heaters is not good for them.
It aslo delays the warm up of previous stages.
Much easier and simpler to just use a bias resistor - it does the job better, is cheap, simple, and reliable.
There have been a few designs using the bypassed cathode of the ouput tube to get pure DC for input stage heaters, to secure very low hum and immunity for noise comming in via teh power transformer. There's some validity in that thinking, but it was rarely if ever used commercially for the reasons I gave above.
Last edited:
Keit, please respond:
Please clarify, is the LTP type shared cathode one tail CCS unbypassed included in your comments regarding CCS use?
Please clarify, is the LTP type shared cathode one tail CCS unbypassed included in your comments regarding CCS use?
Very good question. I may have missed covering this correctly earlier.
Since there is no bypass capacitor, overdrive recovery cannot be adversly affected.
Each tube will be overbiased with large signals and will create more even order distortion. However, the effect on output distortion will be not great as even order harmonics will cancel in the output transformer, to the extent that the two tubes are matched. The net result will be a small reduction in power output - about 10% or so.
I still would not recommend CCS cathode biasing as there is no advantage - its adding complexity, cost, and failure modes for no reason.
Last edited:
This EFB mod may interest you and has been reliable and well liked as it improves sonics and output tube life by using a negative regulator which floats with the B+ to give fixed bias for low $.
Dave's Lab
Dynaco SCA-35 -- the EFB mod - AudioKarma.org Home Audio Stereo Discussion Forums
It has been used on other amps like Fisher on that forum in AK It will also work with other easy to drive tubes like the 7591 types.
Dave's Lab
Dynaco SCA-35 -- the EFB mod - AudioKarma.org Home Audio Stereo Discussion Forums
It has been used on other amps like Fisher on that forum in AK It will also work with other easy to drive tubes like the 7591 types.
- Status
- Not open for further replies.