Tube amps: Self Biasing vs. Manual?
- Thread starter StevenZ
- Start date
Shadowdog
Super Member
Cathode bias or self adjusting bias may not work well with 8417 tubes just like it doesn't work well with 7591 types (Big power loss etc.). Cathode bias prefers much higher primary impedance on the output trans. also according to the tube data sheets.
Individual adjustment pots could be added for the cost of the pots,needed resistors, wire and maybe test points.
Randy
Individual adjustment pots could be added for the cost of the pots,needed resistors, wire and maybe test points.
Randy
shelly_d
Not An Audiophool
Matching may or may not be needed depending upon the circuit. Some circuits will tie the output cathodes together and sent the current of both tubes through the cathode resistor. Matching would be required for this situation because with unmatched tubes, one could "hog" the current over the other. If each tube had its own cathode resistor, each tube will find its own current level. It may or may not affect the sound quality but nonmatched tubes will not effect reliability.
Shelly_D
bhamham
Super Member
How is that? 7591/7868 was used in a few cathode bias schemes by Fisher e.g., X-100-B as Dave mentioned in post #8.
.
Shadowdog
Super Member
Re cathode bias for 7591 types: I think it was in Gordon W's thread where someone asked about converting the Heathkit AA-100 in a thread on rebuilding them , and he said you would loose a lot of power. He was talking about the new production tubes which wouldn't work as well apparently. Even on the old TS tube data sheets the wattage is way down to 28W running the tubes hard. They also show upping the primary of the output trans. to 9K from 6.6K. You only get 26W in UL , but you can have a 6.6K primary.
Randy
Randy
All else being equal, an output stage designed for cathode bias operation will always produce less power output than an output stage designed for fixed bias, due to two basic issues: the presence of the cathode resistor itself, and the load impedance required for optimum operation.
Starting with the cathode bias resistor, anything that reduces peak current flow in the output stage circuit will reduce power output. Since the cathode bias resistor, output tubes and the power supply all effectively form a series circuit, then the cathode resistor must be limiting peak current flow in that circuit by definition. With a fixed bias design, you effectively only have the power supply and the output stage in series with each other (amounting to being placed in parallel), allowing more current to flow, and greater output to be produced.
From this understanding then, it can be shown that tubes requiring a higher bias voltage for proper quiescent current flow will suffer the most in a cathode bias setting, simply because the cathode bias resistor has to be relatively larger to create the larger bias voltage required for such tubes. The greater the cathode bias resistor is, the more peak current is restricted, producing --relative to other tubes requiring smaller value cathode resistors -- a greater loss of power output.
This is (in part) why the last "new" batch of audio power tubes produced were all high Gm tubes, where relatively much smaller grid voltages were needed to control a given plate current. Such designs allow for smaller cathode resistors to be used, allowing the development of more power from the most economical circuits, benefiting both consumer and manufacturer.
The other part of the equation is the required load impedance that a cathode bias stage must operate into versus that of a fixed bias output stage. It can be shown that maximum power transfer in any circuit is had when the driving impedance (in this case that of the output tubes) is equal to the load impedance (the speaker in this case, whose impedance is transformed into a higher impedance being reflected back to the output tube plates by the OPT). As a basic point, as this relationship deviates from one of equality, power output is reduced, and distortion is increased.
Since the cathode bias resistor inherently reduces current flow in the output tube circuit, then the impedance of the circuit is automatically raised (restricting current flow) by such a configuration. Within the tubes, it is the screen grid that determines the peak amount of current that the tubes themselves can flow (within ratings), so to achieved the lowest distortion, the operating parameters of these grids are also adjusted (typically downwards) so that what the tubes can flow, matches what the total circuit will flow.
Since the the cathode resistor has raised the effective impedance of the circuit, and the tubes have been adjusted to properly operate at the elevated circuit impedance, then the load that the tubes operate into must also be raised to maintain an equality of circuit elements, to allow production of the most power at the least distortion that the configuration can support.
UL operation throws a third monkey wrench into the works which I won't go into any further, other than to say that because the AC signal applied to the screen grid affects the ultimate current flow capability of the tube, it is also therefore affecting the effective impedance that the tube represents, and so has its own impact on optimum loading requirements.
As can be seen, regardless of the bias scheme used, it is all basically a juggling act, where the circuit impedance of each element is aligned with that of the others to produce optimum performance. When they get out of lockstep with each other, performance suffers resulting in elevated distortion and reduced power output.
Understanding these things then, casually converting from one bias scheme to another is not always a good move to make, unless the other circuit elements are known or adjusted accordingly. In many cases, compromises were made in the interest of economy, so that a change in bias configuration can in fact produce significant performance improvements.
For example, this was exactly the case at hand with the installation of the EFB(tm) modification into the small Dynaco amplifiers. This modification (in part) transformed those amplifiers from cathode bias to fixed bias operation, and in the process, correctly aligned the impedance of all the circuit elements to maximize the performance of those designs. For those who have heard or measured the results of this modification to those units, they are not insignificant.
I hope this helps!
Dave
Starting with the cathode bias resistor, anything that reduces peak current flow in the output stage circuit will reduce power output. Since the cathode bias resistor, output tubes and the power supply all effectively form a series circuit, then the cathode resistor must be limiting peak current flow in that circuit by definition. With a fixed bias design, you effectively only have the power supply and the output stage in series with each other (amounting to being placed in parallel), allowing more current to flow, and greater output to be produced.
From this understanding then, it can be shown that tubes requiring a higher bias voltage for proper quiescent current flow will suffer the most in a cathode bias setting, simply because the cathode bias resistor has to be relatively larger to create the larger bias voltage required for such tubes. The greater the cathode bias resistor is, the more peak current is restricted, producing --relative to other tubes requiring smaller value cathode resistors -- a greater loss of power output.
This is (in part) why the last "new" batch of audio power tubes produced were all high Gm tubes, where relatively much smaller grid voltages were needed to control a given plate current. Such designs allow for smaller cathode resistors to be used, allowing the development of more power from the most economical circuits, benefiting both consumer and manufacturer.
The other part of the equation is the required load impedance that a cathode bias stage must operate into versus that of a fixed bias output stage. It can be shown that maximum power transfer in any circuit is had when the driving impedance (in this case that of the output tubes) is equal to the load impedance (the speaker in this case, whose impedance is transformed into a higher impedance being reflected back to the output tube plates by the OPT). As a basic point, as this relationship deviates from one of equality, power output is reduced, and distortion is increased.
Since the cathode bias resistor inherently reduces current flow in the output tube circuit, then the impedance of the circuit is automatically raised (restricting current flow) by such a configuration. Within the tubes, it is the screen grid that determines the peak amount of current that the tubes themselves can flow (within ratings), so to achieved the lowest distortion, the operating parameters of these grids are also adjusted (typically downwards) so that what the tubes can flow, matches what the total circuit will flow.
Since the the cathode resistor has raised the effective impedance of the circuit, and the tubes have been adjusted to properly operate at the elevated circuit impedance, then the load that the tubes operate into must also be raised to maintain an equality of circuit elements, to allow production of the most power at the least distortion that the configuration can support.
UL operation throws a third monkey wrench into the works which I won't go into any further, other than to say that because the AC signal applied to the screen grid affects the ultimate current flow capability of the tube, it is also therefore affecting the effective impedance that the tube represents, and so has its own impact on optimum loading requirements.
As can be seen, regardless of the bias scheme used, it is all basically a juggling act, where the circuit impedance of each element is aligned with that of the others to produce optimum performance. When they get out of lockstep with each other, performance suffers resulting in elevated distortion and reduced power output.
Understanding these things then, casually converting from one bias scheme to another is not always a good move to make, unless the other circuit elements are known or adjusted accordingly. In many cases, compromises were made in the interest of economy, so that a change in bias configuration can in fact produce significant performance improvements.
For example, this was exactly the case at hand with the installation of the EFB(tm) modification into the small Dynaco amplifiers. This modification (in part) transformed those amplifiers from cathode bias to fixed bias operation, and in the process, correctly aligned the impedance of all the circuit elements to maximize the performance of those designs. For those who have heard or measured the results of this modification to those units, they are not insignificant.
I hope this helps!
Dave
If you notice, the X100, as well as most other cathode-bias 7591 amps (Eico HF40, etc) are significantly less power than their fixed-bias brethren- pretty much a max of about 20 watts per pair of output tubes, while fixed bias versions could make 35-40 watts.
Dave mentioned many of the factors involved with this- but one other, is that the 7591 and 7868 have relatively low plate dissipation, compared to other tubes commonly used in '35 watt class' amps. With only 19 watt plates, they have to be idled at pretty low currents, or be used with low plate voltages, to not exceed rated plate dissipation- which under cathode bias, really limits power output. OTOH, with a fixed-bias amp, they can be run "hard Class AB"- much closer to Class B, with very low idle currents and high B+ voltages- and be able to swing pretty large amounts of current and voltage on peaks- therefore, producing significant power.
Regards,
Gordon.