Shef,
I have had good results with 'fixed' (adjustable) bias; self bias, and battery bias.
Each has its advantages, and each has its disadvantages.
I know several of the disadvantages of battery bias.
However, I would appreciate you telling us what you mean by 'weak design' of battery bias.
I have had good results with 'fixed' (adjustable) bias; self bias, and battery bias.
Each has its advantages, and each has its disadvantages.
I know several of the disadvantages of battery bias.
However, I would appreciate you telling us what you mean by 'weak design' of battery bias.
Thanks 6A3sUMMER! My query actually was about equalizing both tubes' currents. I guess it only relies on matched tubes.
My concern about battery bias and the pot is that there will be significant current flowing when the batteries are connected, alkaline batteries are not known for their flat discharge curves and I expect you will be adjusting the bias frequently. A23 in particular are great when you are talking a few uA, the picture I think would change significantly with as little as 1mA of drain during operation, a string of 9V batteries would provide better performance longer term.
When I suggested fixed bias I was thinking that a small transformer might be a possibility, but perhaps not.
When I suggested fixed bias I was thinking that a small transformer might be a possibility, but perhaps not.
That's why the battery would be disconnected from the pot through a relay while the amp is powered off. In reality I was just wondering from the standpoint of a proof of concept. As I said right now I use -27V from ground to grid leak resistor, and a CCS in the cathode. I didn't realise 0.0005a was a lot of current for an A23.
An isolation tranny with a V doubler? I think you could just parallel off of one leg of that PT secondary with a half-wave rectifier then filter and divide that down to get your bias supply. Just like any other PT with a dedicated bias tap.
0.5mA is pretty significant for an A23 battery which typically has something like a 55mAH - 70mAH rating, the voltage will start to drop almost immediately - end of life would be defined by the point at which you can no longer set the bias voltage to the value desired.
Note that I was aware of the fact that you intended to disconnect the batteries when not in use.
I attached a datasheet indicating expected performance of an Energizer A23. It clearly illustrates the issue I was discussing.
As an aside I hope that the B+ is derived from the secondary of an isolation transformer for safety reasons.
Good point. Any suggestions for a HV negative regulator?
Thanks for the datasheet. And yes it's running through a 300VA 62-0-62 with the centre tap disconnected.
Looks like I'll keep those for my twin bias setup.
Last edited:
MagicBus,
The old Dynaco Stereo 70 had only one fixed bias potentiometer per each push pull pair of tubes. That required well matched tubes, and if the tubes did not age together, well, they needed to be replaced. Initially, or after aging, if they were unbalanced, you got early saturation of the output transformer at bass frequencies (and you also got intermodulation distortion on the midrange and highs, whenever bass frequencies were present). This was not a particularly good idea, because there was no easy way to check if the current was unbalanced (initially, and later with aging). It would have been possible to replace the 15.6 Ohm sense resistor that was on 2 EL34 cathodes that were tied together. An individual resistor for each cathode, 1 Ohm for example, would have made it easy for a modern DMM to check the current of each individual EL34 (4 each 1 Ohm resistors). But in order to adjust each cathode current, 2 more potentiometers would have to be wired in, so each tube could have its own.
I only ever used fixed (adjustable) bias and bias potentiometers in 2 kinds of amplifiers:
Dynaco Stereo 70
Single Ended, so only one tube was used per channel, and each had its individual potentiometer.
I do not think that battery bias is a good thing for push pull amplifiers, unless you also have mixed bias, for example some self bias too (like 15V fixed bias, and 15V self bias). But that gets rid of most of the advantages of battery bias. The tubes still must be fairly well matched, or there may be enough difference in plate current to cause early saturation of the output transformer. Full self bias does not require such closely matched tubes.
For battery bias, some will use a potentiometer to adjust the bias, and will use a switch that disconnects the potentiometer from the battery when the amplifier is turned off.
But that means there may still be B+, when the bias pot is disconnected at turn off. That would leave the control grid floating, which might result in extremely high plate current until the B+ discharged.
For the above reasons, the only way I used battery bias was for single ended amplifiers, not push pull, and I also did not use a bias adjustment potentiometer. I only adjusted the B+ to get the current I wanted, and I used matched tubes (these were mono blocks), so the amplifier quiescent current and voltage, and the dynamic signal conditions matched from one mono block to the other mono block. The method I used that allowed me to use the same battery for years, and is described in my post # 17 in this thread.
I hope that helps.
The old Dynaco Stereo 70 had only one fixed bias potentiometer per each push pull pair of tubes. That required well matched tubes, and if the tubes did not age together, well, they needed to be replaced. Initially, or after aging, if they were unbalanced, you got early saturation of the output transformer at bass frequencies (and you also got intermodulation distortion on the midrange and highs, whenever bass frequencies were present). This was not a particularly good idea, because there was no easy way to check if the current was unbalanced (initially, and later with aging). It would have been possible to replace the 15.6 Ohm sense resistor that was on 2 EL34 cathodes that were tied together. An individual resistor for each cathode, 1 Ohm for example, would have made it easy for a modern DMM to check the current of each individual EL34 (4 each 1 Ohm resistors). But in order to adjust each cathode current, 2 more potentiometers would have to be wired in, so each tube could have its own.
I only ever used fixed (adjustable) bias and bias potentiometers in 2 kinds of amplifiers:
Dynaco Stereo 70
Single Ended, so only one tube was used per channel, and each had its individual potentiometer.
I do not think that battery bias is a good thing for push pull amplifiers, unless you also have mixed bias, for example some self bias too (like 15V fixed bias, and 15V self bias). But that gets rid of most of the advantages of battery bias. The tubes still must be fairly well matched, or there may be enough difference in plate current to cause early saturation of the output transformer. Full self bias does not require such closely matched tubes.
For battery bias, some will use a potentiometer to adjust the bias, and will use a switch that disconnects the potentiometer from the battery when the amplifier is turned off.
But that means there may still be B+, when the bias pot is disconnected at turn off. That would leave the control grid floating, which might result in extremely high plate current until the B+ discharged.
For the above reasons, the only way I used battery bias was for single ended amplifiers, not push pull, and I also did not use a bias adjustment potentiometer. I only adjusted the B+ to get the current I wanted, and I used matched tubes (these were mono blocks), so the amplifier quiescent current and voltage, and the dynamic signal conditions matched from one mono block to the other mono block. The method I used that allowed me to use the same battery for years, and is described in my post # 17 in this thread.
I hope that helps.
Last edited:
It may be just easier to use a fixed cathode bias, even though that reduces B+ somewhat. Any related drop in output power rating is probably in the mind of the beholder, as it is likely less than a dB or so reduction.
Fixed cathode bias using a zener, or parallel R-C-Z is pretty simple, and a rotary switch can easily be added to add/subtract standard diodes in series to provide easy manual adjustment of bias.
Fixed cathode bias using a zener, or parallel R-C-Z is pretty simple, and a rotary switch can easily be added to add/subtract standard diodes in series to provide easy manual adjustment of bias.
@ 6A3sUMMER, all clear, thanks! For output stages I use fixed bias with individual pots from a transformer supply. No batteries here! I was thinking to use the idea with the batteries on an input differential stage where the design calls for direct coupling with the next stage. A fixed bias together with the cathodes CCS would help to get balanced DC levels at the output. It would only take 10-20mV of negative voltage at the grids to fix this but it has to be adjustable so, again a transformer supply it has to be...
@ Kodabmx, I think regulated fixed bias also calls for regulated B+.
@ Kodabmx, I think regulated fixed bias also calls for regulated B+.
You might try this method:
Using a differential stage that has precision matched plate load resistors, connect:
Ground to the bottom of the current source.
The top of the current source to the Wiper of a 25 Ohm pot.
One end of the pot to one cathode.
The other end of the pot to the other cathode.
Adjust the pot until the plate voltages match.
If adjusting the pot does not match the plate voltages, you will either have to try another
tube (of the same tube type), or change the pot to a 50 Ohm pot.
I recommend finding a tube that will work with the 25 Ohm pot, because you want DC balance, but you also want gain balance (depends on the impedance of the current source).
Using a differential stage that has precision matched plate load resistors, connect:
Ground to the bottom of the current source.
The top of the current source to the Wiper of a 25 Ohm pot.
One end of the pot to one cathode.
The other end of the pot to the other cathode.
Adjust the pot until the plate voltages match.
If adjusting the pot does not match the plate voltages, you will either have to try another
tube (of the same tube type), or change the pot to a 50 Ohm pot.
I recommend finding a tube that will work with the 25 Ohm pot, because you want DC balance, but you also want gain balance (depends on the impedance of the current source).
I keep this as a final solution. It's CCS impedance and tubes gm that makes me skeptical about inserting resistanse between the cathodes.
As long as the CCS impedance is high enough, all the current change from one cathode passes on to the other cathode. And if the tubes are triodes, and if there is no grid current, then the current change in the matched plate loads are identical, but out of phase.
So G1 = -G2.
Nothing is perfect, but may be pretty close.
So G1 = -G2.
Nothing is perfect, but may be pretty close.
When bias cells were used in early radio sets or later in some laboratory test equipment the circuits were allways designed for ZERO current draw from the cell (direct grid bias). In such circuits the cell had a very long life (actually the shelf life of the cell, up to many years) and no switching was required. ANY load on the bias cell (or battery), even a high value pot, will considerably shorten its life and a monitoring circuit for the bias voltage cell is mandatory. I would NEVER use a battery circuit to bias the output stage of a "fixed bias" power amp, very risky and poor engineering practice. I can't see any advantage in this scheme, only needless complexity and compromised reliability.
That's the crux of it. A grid bias battery that draws no current is the way to go. It 1) lasts for years, especially if it is a modern alkaline or Li ion battery; 2) can be permanently wired into circuit and does not need a switch; 3) provides much more stable and noise-free bias compared to any other arrangement; 4) when used with grid choke or interstage transformer, provides lowest resistance path from grid to ground, which is very beneficial for tube viability and longevity. Done this way, battery is the best way of applying fixed bias.
I do not recommend battery bias for a differential pair that is going to be DC coupled to the next stage.
I do not recommend battery bias for Any stage that DC couples to the next stage.
I only use battery bias where the battery current is essentially zero.
I use it signals that are small enough to not clip, triodes that have extremely low leakage grid current, and low enough miller capacitance so that high frequencies do not draw significant current (which is AC, and charges/discharges with extremely low currents).
And at power off, my configuration has battery current that is only due to GigOhms of resistive load (pretty close to zero current).
The configuration does not use a pot, and does not use a switch.
I have used battery bias for 300B tubes, but care in design is paramount, as is using very good tubes that meet all their specifications, and then using the tube at voltages, currents, and power ratings well below their ratings.
Whenever I have enough B+, I do not use battery bias, instead I use Cathode Self Bias
(or for DHTs, Self Bias applied to the 'center' of the Filament).
A fixed 'adjustable' bias supply has one of the same potential (pun Not intended) problems as the battery bias supply. Neither one prevents thermal tube runaway.
Self Bias may help to prevent thermal tube runaway.
Some new (and maybe some new old stock tubes) do not meet their specifications for maximum grid 'leak' resistor. In that case, even if you use less than the maximum rated grid leak resistor, the tube plate current may very well run away.
(I learned this the hard way).
I do not recommend battery bias for Any stage that DC couples to the next stage.
I only use battery bias where the battery current is essentially zero.
I use it signals that are small enough to not clip, triodes that have extremely low leakage grid current, and low enough miller capacitance so that high frequencies do not draw significant current (which is AC, and charges/discharges with extremely low currents).
And at power off, my configuration has battery current that is only due to GigOhms of resistive load (pretty close to zero current).
The configuration does not use a pot, and does not use a switch.
I have used battery bias for 300B tubes, but care in design is paramount, as is using very good tubes that meet all their specifications, and then using the tube at voltages, currents, and power ratings well below their ratings.
Whenever I have enough B+, I do not use battery bias, instead I use Cathode Self Bias
(or for DHTs, Self Bias applied to the 'center' of the Filament).
A fixed 'adjustable' bias supply has one of the same potential (pun Not intended) problems as the battery bias supply. Neither one prevents thermal tube runaway.
Self Bias may help to prevent thermal tube runaway.
Some new (and maybe some new old stock tubes) do not meet their specifications for maximum grid 'leak' resistor. In that case, even if you use less than the maximum rated grid leak resistor, the tube plate current may very well run away.
(I learned this the hard way).
So far I've only used battery bias with no pot or switch, in conjunction with CCS cathode bias. The trick is ringing another 18V out of the B+ while being able to use cheaper lower voltage caps on the cathode (It's a complex biasing set up) I've been running like this for a good month so far without a change in voltage. I'm obviously still testing longevity. I use carbon-zinc batteries, and they are connected from ground through a 100k resistor to each grid stopper, after the coupling cap. Of course if it drifts or fails completely the CCS takes up the slack, keeping the tube from exceeding it's limits.
EDIT: I've used this with KT88, 5881, and 6V6 with good results so far.
EDIT: I've used this with KT88, 5881, and 6V6 with good results so far.
Last edited:
- Status
- Not open for further replies.