Hi,
I read from Pg 91 Audio Reality by Bruce Rozenblit
"Here goes another myth down the drain, There is no precise bias current setting for optimum operation. There are only range of operation. Generally speaking, the heavier the bias current, the warmer the tube will sound. Some people think warmer is better. I definitely do not. My design are based on maximizing accuracy. What is critical here is not the absolute level of current , but the balance of current. Output transformer cannot tolerate much of DC current imbalance. Too much imbalance will cause premature saturation & severe distortion."
....
He then went on describing how he adjust the bias voltage of the tube to obtain current balance in both sections of the pp output stage.
As an engineer (not EE but ME) I can understand the science behind the PP transformer but really don't know what to believe any more regarding the part in the bias point now!
I raise this subject here to hear what the great minds here has to say and hopefully, will help clear my thoughts.
Cheers
Ken
I read from Pg 91 Audio Reality by Bruce Rozenblit
"Here goes another myth down the drain, There is no precise bias current setting for optimum operation. There are only range of operation. Generally speaking, the heavier the bias current, the warmer the tube will sound. Some people think warmer is better. I definitely do not. My design are based on maximizing accuracy. What is critical here is not the absolute level of current , but the balance of current. Output transformer cannot tolerate much of DC current imbalance. Too much imbalance will cause premature saturation & severe distortion."
....
He then went on describing how he adjust the bias voltage of the tube to obtain current balance in both sections of the pp output stage.
As an engineer (not EE but ME) I can understand the science behind the PP transformer but really don't know what to believe any more regarding the part in the bias point now!
I raise this subject here to hear what the great minds here has to say and hopefully, will help clear my thoughts.
Cheers
Ken
Certain transformers require perfect current balance or they saturate as described. These are ones without airgaps such as toroidals. These tend to have low hysteresis which is good when the current direction changes because micro detail is preserved at these very low signal levels.
In the case of airgapped transformers hysteresis will be much higher and will tend to produce loss of micro detail at low signal levels. This is why SE transformers can be better at micro detail. However some people prefer to run their output transformers with a bit of DC imbalance as this minimises hysteresis effects at low signal level.
Personally I opt for the none airgapped transformers run at zero offset. My prefered transformer is the toroid. The advantage are
1-wider bandwidth
2-less hysteresis induced blurring at crossover - better micro detail.
3-almost total cancellation of residual power supply hum from the finals. In theory if the current balance were perfect you could run with no power supply capacitance and still have zero hum !!
Shoog
In the case of airgapped transformers hysteresis will be much higher and will tend to produce loss of micro detail at low signal levels. This is why SE transformers can be better at micro detail. However some people prefer to run their output transformers with a bit of DC imbalance as this minimises hysteresis effects at low signal level.
Personally I opt for the none airgapped transformers run at zero offset. My prefered transformer is the toroid. The advantage are
1-wider bandwidth
2-less hysteresis induced blurring at crossover - better micro detail.
3-almost total cancellation of residual power supply hum from the finals. In theory if the current balance were perfect you could run with no power supply capacitance and still have zero hum !!
Shoog
In push-pull operation is hard to avoid imbalance especially as tubes are aging and this may require often rebiasing to get back to the sound you were used to; sometimes you may even have to change one or both tubes though they are still good but impossible to balance. I prefer to use balanced pp schematics where non or very little DC is flowing through the primary winding like the circlotron or bridge arangements ; these arengements require more components ( tubes or power supplys etc. ) but allow for not expensive single end transformers without a gap, often power toroidal transformers without compromising to much on the sound.
They are, and have rather large hysteresis. However, a conventional SE amplifier operates on a linear part of the BH curve. Current never changes direction, there is only varying amount of current flowing in one direction. So hysteresis doesn't really matter for convention SE transformer.
For a push pull transformer, current does change directions so hysteresis becomes more important. This makes air gap undesirable in push pull transformers. But this also means that it is important to keep the current matched fairly close between each half of the winding. An equal current flowing through each half magnetizes the core in opposite directions which cancels each other out. So the transformer sees no DC when each side is matched.
Parallel feed (parafeed) single ended transformers are like push pull transformers, but require a coupling cap to keep DC out. Since there is no opposing current source in SE to cancel DC effects on the transformer.
Like Wavebourn said. As soon as you have signal the transformer is unbalanced, hence the intermodulation. It's easy to see here (compare post 92 and 91) http://www.diyaudio.com/forums/showthread.php?postid=801247#post801247
Sheldon
Sheldon
Audio_idiot said:Hi,
I read from Pg 91 Audio Reality by Bruce Rozenblit
"Here goes another myth down the drain, There is no precise bias current setting for optimum operation. There are only range of operation. (So far, so good.) Generally speaking, the heavier the bias current, the warmer the tube will sound. (No idea what this is supposed to mean, unless, since the tubes will run hotter they sound "warmer" because they are warmer. Running a higher Q-Point current will move PP finals more towards Class A operation, and reduce X-over distortion. Other than that,) Some people think warmer is better. I definitely do not. My design are based on maximizing accuracy. What is critical here is not the absolute level of current , but the balance of current. Output transformer cannot tolerate much of DC current imbalance. Too much imbalance will cause premature saturation & severe distortion."
....
He then went on describing how he adjust the bias voltage of the tube to obtain current balance in both sections of the pp output stage.
As an engineer (not EE but ME) I can understand the science behind the PP transformer but really don't know what to believe any more regarding the part in the bias point now! (Don't listen to audiophool "folk wisdom". Believe what your loadlines and ears are telling you.)
I raise this subject here to hear what the great minds here has to say and hopefully, will help clear my thoughts.
Cheers
Ken
There is an awful lot of audiophoolery concerning bias for amps, most of it is contradictory, and will probably cause you to ruin your finals. Stick with what your loadlines tell you, what the manufacturer's suggestions are, and forget about the nonsense.
As for current balance, this is easily done with fixed bias. You can also include a balance pot with cathode bias, or use a DC servo to maintain balance.
What I have read, and seen on O'scope pictures, is that the type of distortion is affected by the bias point. Bias point changes which half of the waveform starts to flatten out first, top or bottom, and that will affect the sound. We want distortion in guitar amps, so we will hear the different tone of the different bias.
Hi guys,
Thanks for the great insight. I've been reading the reply and took me some time to digest. Please pardon by poor English.
It does bother me to think that only the current is critical. What about when the 2 winding are imbalance, eg, DCR & L of the primary coil differ by +/- 10% from each half? I know we can add small resistor to equal the two sides but is it neccessary to add some extra windings for the primary to balance the L?
Shoog,
Is the current balance the reason/basis of your good performance of your use of power toroidal transformer used as OPT?
Wavebourn,
I could not understand "hum in pauses" in your reply... Pardon my english.
Wavebourn, Sheldon,
I get the IMD part, are you to say that there are no advantage in PP to simplify or more tolerance to PS design. i.e we still need to reduce Vripple to 5% of B+ in question? or maybe we can let it go to 20% of B+ and still get away with things. I know SE will hum happily when that happen.
Flysig,
This is great info, so does the Q point falls on a sweet spot of over a wide range?
Cheers
Ken
Thanks for the great insight. I've been reading the reply and took me some time to digest. Please pardon by poor English.
It does bother me to think that only the current is critical. What about when the 2 winding are imbalance, eg, DCR & L of the primary coil differ by +/- 10% from each half? I know we can add small resistor to equal the two sides but is it neccessary to add some extra windings for the primary to balance the L?
Shoog,
Is the current balance the reason/basis of your good performance of your use of power toroidal transformer used as OPT?
Wavebourn,
I could not understand "hum in pauses" in your reply... Pardon my english.
Wavebourn, Sheldon,
I get the IMD part, are you to say that there are no advantage in PP to simplify or more tolerance to PS design. i.e we still need to reduce Vripple to 5% of B+ in question? or maybe we can let it go to 20% of B+ and still get away with things. I know SE will hum happily when that happen.
Flysig,
This is great info, so does the Q point falls on a sweet spot of over a wide range?
Cheers
Ken
Audio_idiot said:
I don't know at what level this is audible, and I would guess that sensitivity to it depends somewhat on the individual and the music they listen to. From my perspective, it's not that difficult or expensive to design a power supply with low ripple, so I shoot for that. I would not accept 20% ripple, or anything close to it. However, this is DIY, and we are free to experiment to suit our own taste. By the way, the design part is easy - PSUD (free software) is your friend.
Sheldon
Its a chicken and egg thing really. I started out designing around the DC intolerance of toroids. Since this allows me to implement a differential output stage I have stuck with it. If I transfered the same techniques to normal OT's hysteresis would degrade the result and bandwidth would be limited as well. So having tasted the benefits of toroids I am stuck with using them and cannot tolerate any DC inbalance in the output stage. The only improvement would be to use toroids specifically designed for output duty. I would be very reluctant to try normal EI's in their place. That day will come.
Shoog
Yes, PSUD is our best friend! make my DIY life much bearable. Really, to get below 5% is easy, but just want to know what is the possibility of not having a reservoir cap in a PP amp, or better, in a LTP, I may try it out in my next test rig.Sheldon said:
Shoog said:
I see you are addicted to power toroid tranny... what do you mean by special toroids? Is winding technique you are refering to?
Cheers
Ken
No I mean the commercially available output toroidals.
its interesting to note that my current amp has about 1mA of mismatch on the outputs on one side and this is more than enough to produce an audible amount of hum from that side. I have minimal power supply filtering caps. 4uf 5H 4uf 110R 100uf. I can hear no artifacts in the music, but I need to get that balance down to less than 1mA - just shows how critical it is. there is also the issue of premature saturation to think about. New CCS are in order I think.
Shoog
its interesting to note that my current amp has about 1mA of mismatch on the outputs on one side and this is more than enough to produce an audible amount of hum from that side. I have minimal power supply filtering caps. 4uf 5H 4uf 110R 100uf. I can hear no artifacts in the music, but I need to get that balance down to less than 1mA - just shows how critical it is. there is also the issue of premature saturation to think about. New CCS are in order I think.
Shoog
Audio_idiot said:
I am far from an expert, but my understanding is that a slight imbalance does not require any correction. A transformer is a current device. The current flowing through the primary causes the magnetic field, which then induces a current in the secondary. The voltage is not what is doing the work. The current is doing the work and the voltage is created by the transformer.
Output tubes do not increase the voltage much from the input. The gain is fairly small. But the current is multiplied many many times. So the output section is a current multiplier. Or, we could say that the output is a power multiplier. It is a power amplifier. The preamp is a voltage amplifier, with very little current gain.
As the output tube operates, the transformer DC resistance is very low. When you see a primary impedance specified as 5000 ohms, for example, that is the reflected impedance of the speaker impedance during AC signals. The 8 ohm speaker on the secondary looks like 5000 ohms on the primary side during AC signals. But for DC the primary is very low resistance.
Now here is where the experts can jump in and correct me if I am wrong:
During AC signals, the output tube generates or provides a certain amount of power. The transformer will soak up that power into a magnetic field and then transfer the power to the secondary and then deliver it to the speaker load. Any slight differences in the primary will result in the same power being transferred, but the current and voltage may vary slightly. Power is the product of Voltage x Current. That power stays the same. If the current is a little bit lower, the voltage will be a little bit higher, and vice versa.
Audio_idiot said:
For the minimum distortion you want to center bias the Q point on the DC load line. But.... it depends also on the slope of the load line and the grid characteristic curves. The grid curves are not spaced evenly on the graph, so the distortion near cutoff may happen at a different rate than the distortion near saturation. If your load line is too flat or too steep you might encounter more cutoff or more saturation, not at the same time, even though your bias point is centered.
The goal is to pick the bias point on the DC load line so that the waveform is symmetrical. This gives you the maximum input swing before distortion. If your input is less than maximum, you can move the bias point without causing distortion. The harder you drive the tubes, the more precise your bias point has to be.
Most output transformers are designed to tolerate a small amount of DC imbalance. Some are not. Put DC into one that is not and it will saturate and nearly all the inductance will bleed away - result much distortion. A none airgapped transformer will only tolerate a few mA of current imbalance.
The upside is that such a transformer will have no problems crossing over from one current direction to the other, that is it will not store energy and release it in such a way as to blur the current trying to move in the opposite direction. It is generally reported that PP transformers are not as good as SE transformer in dealing with very small signals and this magnetic memory is the reason why. Vacuum State electronics go so far as to introduce a bit of high frequency bias signal to make certain that when a small signal is been delivered, the transformer is current biased such that this magnetic memory is shifted out of the sensitive low signal area of its range.
If you can deliver perfect current balance to a none airgapped and histeresis free transformer (such as a toroidal) the signal will almost perfectly transition from one current direction to the other with no blurring. So it has the potential to deliver SE quality low level detail but with better high and low frequency extension, because it doesn't have to cope with a large DC bias current. Those who have got this arrangement to work have consistently reported superior results to conventional transformers.
Hope that clarifies why perfect DC balance in the right type of PP transformers is important. In the wrong type of PP transformer it will tend to worsen the low level signal smearing.
Shoog
The upside is that such a transformer will have no problems crossing over from one current direction to the other, that is it will not store energy and release it in such a way as to blur the current trying to move in the opposite direction. It is generally reported that PP transformers are not as good as SE transformer in dealing with very small signals and this magnetic memory is the reason why. Vacuum State electronics go so far as to introduce a bit of high frequency bias signal to make certain that when a small signal is been delivered, the transformer is current biased such that this magnetic memory is shifted out of the sensitive low signal area of its range.
If you can deliver perfect current balance to a none airgapped and histeresis free transformer (such as a toroidal) the signal will almost perfectly transition from one current direction to the other with no blurring. So it has the potential to deliver SE quality low level detail but with better high and low frequency extension, because it doesn't have to cope with a large DC bias current. Those who have got this arrangement to work have consistently reported superior results to conventional transformers.
Hope that clarifies why perfect DC balance in the right type of PP transformers is important. In the wrong type of PP transformer it will tend to worsen the low level signal smearing.
Shoog
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.