I should not reply to your arrogant posts. However just to show what you understand, or better what you do not understand, here is one comment to one of your gems of knowledge.
You wrote:
Distortion WILL be amplified as it is in the signal. Where do you think cancellation happens?
To make it work one makes a prelimary assessment, often based on previous experience, then trial and error with small ajustments of working conditions to get the best result. Something you don't seem to be capable of....
Last edited:
According to Klaus' testing of the 6N8S which is an analog of the 6SN7 no grid current flows until 0V
Tube Tester Files - 6N8S - 6H8C
Tube Tester Files - 6N8S - 6H8C
Nice article, thanks.
It is visible from the picture that the test limit was set to 287uA and appears in red at 4mA bias at 10Vrms grid signal at +0.5V of positive bias.
I am sure there is absolutely no problem driving at 8mA bias a 2Vrms signal in the grid at -0.5V.
Just imagine, the grid swinging from +8V to -7V at +0.5 bias with 287uA grid current, this is pretty impressive, BTW a 25K volume pot would be required instead of the typical 100K input load.
He has marked the curves is red because his equipment starts to have problems and the limit is close. The grid current at 0V-0.5V is already too high considering the resistors used around the grid of a signal tube. Can read the value in the left side table.
Generally speaking grid current starts before the grid reaches zero volt because the tube will never be perfect with perfect vacuum. As a consequence of this the point where it starts to kick in changes from tube to tube and it changes for the very same trube during its life. It is good practice not to go beyond -0.5V for high gain tubes and -1V for lower gain tubes, unless properly addressed.
An alternative way to assess it is to measure distortion! When it starts to kick in one will very likely see the 2nd harmonic suddenly going down but then, almost immediately, odd harmonics (3rd and 5th especially) will rocket if the driver can't cope with it. Better to stay away in the latter case.
Last edited:
You don't have to imagine that. If you have an amp with a driver biased -0.5v to -0.75v that's going to drive the amp to full volume as designed. The sensitivity will be published as .5v RMS for full power.
I was talking about the tests in kodabmx link. They show a very manageable grid current in the worst condition and very nice curves with positive bias.
Actually the gain of the triode is higher in this region, the transition linear, grid current negligible can be biased with power supply, driver impedance does need to be alright, nothing special there either. I don't know what the fuss is about.
Actually the gain of the triode is higher in this region, the transition linear, grid current negligible can be biased with power supply, driver impedance does need to be alright, nothing special there either. I don't know what the fuss is about.
Last edited:
Why is cathode resistors bias 'not work' in simple theory class AB? I am again confused... I have no clue why you say this and talk about dynamic range.
The grid current does not depend on the circuit, but just the valve voltages. Most valves will start to have some grid current from around -1V upwards, although this is sample and life dependent.gabdx said:
It will be small. Whether it affects anything depends on the signal source impedance. You appeared to describe a reduction in second order distortion, which is a classic sign of distortion cancellation due to grid current.
That is a common mistake.
Class AB means that the current drawn varies with signal level. That means that the cathode resistor bias varies with signal level. Provided that the bypass cap is big enough and the music has sufficiently high dynamic range then the occasional loud parts will not cause too much bias shift. Continuous loud music (or sine wave testing) will shift the bias too far and cause crossover distortion.
You are confusing grid current due to gas with grid current due to a hot cathode. It is the hot cathode which causes grid current to start while the grid is still a bit negative. Gas ions cause the opposite grid current, which can lead to thermal runaway. In either case the result is signal distortion, unless the source has low impedance. Of course, if it has a low impedance then there is no second-order cancellation.45 said:
No I was referring to hot cathode. I just used an improper way to make it simple. "Vacuum" was referred to the presence of the grid that cannot disappear....
Class AB means that the current drawn varies with signal level. That means that the cathode resistor bias varies with signal level. Provided that the bypass cap is big enough and the music has sufficiently high dynamic range then the occasional loud parts will not cause too much bias shift. Continuous loud music (or sine wave testing) will shift the bias too far and cause crossover distortion.
Entering Class AB a cathode bias will go down as the Vka rises forcing the operation in further Class AB minimizing the transition region and time.
This is especially true that you can bias higher for cathode bias, fixed bias introduces crossover distortion and are harder to find a nice transition point.
This would be sad that people believe cathode bias in not friendly of class AB while it is the opposite.
Entering Class AB a cathode bias will go down as the Vka rises forcing the operation in further Class AB minimizing the transition region and time.
This is especially true that you can bias higher for cathode bias, fixed bias introduces crossover distortion and are harder to find a nice transition point.
This would be sad that people believe cathode bias in not friendly of class AB while it is the opposite.
> no way two triodes with different signal amplitudes and distortion characteristics can cancel anything
Clear your mind.
Run a 1st stage with large *current* swing. Say idle at 1mA, swing zero to 2mA. The output will be inverted, large, and bent one way.
Feed that to a 2nd stage. Again, large current swing, 40mA idle zero to 80mA swing. The output will be inverted, large, and bent the *other* way.
The Gm/Ip curves of tubes tend to reflect Child's Law. Practical tubes are not ideal geometry and are often jiggered for "better specs". However they all curve, the same way, and roughly similar amounts.
It IS the case that careful selection of a hard-worked driver with a power stage will show much lower 2nd harmonic than a "good" driver pushing a typical power stage.
Take a warped wood stud. Take another warped wood stud. Lay them with bends opposing. Bind them together tight. The bends tend to cancel-out. Any carpenter knows the bend will NOT go-away completely. But often better than one bent stud alone. I just stuck four home-store economy studs together to make a very straight column for my cellar. 1/4" kink here, 5/16" there, 13/64" on another, the result is as straight as my level. (Straighter than the house.)
This is an old technique, 1930s I think. It works less-well with less-old power stages, because they do not need huge drive and the driver rarely works large current swing.
And the real quicksand is that the 2nd does not "cancel", it re-appears as the 3rd. There's no exact conversion because 2nd rises as square of signal and 3rd rises with cube of signal. In practice, at low levels the THD may be quite small, but at high levels the sour 3rd may be quite strong. And high-order appears. If I squint real close, I'd probably see my warp-cancelled column really has about 5 or 7 micro-bends. Too small to matter for Euler, but maybe "kinky" for audio.
Same general thing for push-pull, except now we do have "same" conditions in both tubes so the cancellation is routine and often good.
Clear your mind.
Run a 1st stage with large *current* swing. Say idle at 1mA, swing zero to 2mA. The output will be inverted, large, and bent one way.
Feed that to a 2nd stage. Again, large current swing, 40mA idle zero to 80mA swing. The output will be inverted, large, and bent the *other* way.
The Gm/Ip curves of tubes tend to reflect Child's Law. Practical tubes are not ideal geometry and are often jiggered for "better specs". However they all curve, the same way, and roughly similar amounts.
It IS the case that careful selection of a hard-worked driver with a power stage will show much lower 2nd harmonic than a "good" driver pushing a typical power stage.
Take a warped wood stud. Take another warped wood stud. Lay them with bends opposing. Bind them together tight. The bends tend to cancel-out. Any carpenter knows the bend will NOT go-away completely. But often better than one bent stud alone. I just stuck four home-store economy studs together to make a very straight column for my cellar. 1/4" kink here, 5/16" there, 13/64" on another, the result is as straight as my level. (Straighter than the house.)
This is an old technique, 1930s I think. It works less-well with less-old power stages, because they do not need huge drive and the driver rarely works large current swing.
And the real quicksand is that the 2nd does not "cancel", it re-appears as the 3rd. There's no exact conversion because 2nd rises as square of signal and 3rd rises with cube of signal. In practice, at low levels the THD may be quite small, but at high levels the sour 3rd may be quite strong. And high-order appears. If I squint real close, I'd probably see my warp-cancelled column really has about 5 or 7 micro-bends. Too small to matter for Euler, but maybe "kinky" for audio.
Same general thing for push-pull, except now we do have "same" conditions in both tubes so the cancellation is routine and often good.
Well said PRR.
It is important to stress that in practice one has to find a balance such that 2nd harmonic is effectively canceled but there is no significant increase in higher harmonics. This might not result in massive cancellation but still significant and will depend on the specific case.
I am pointing this out because in recent years some people with no real grasp (practice mostly) on the subject just took two working conditions they "liked" (basically one where the driver has minimum THD and one where the driver will cancel most of the 2nd without caring about the rest) to prove that it is bad and published it....that's exactly how it should NOT be done!
It is important to stress that in practice one has to find a balance such that 2nd harmonic is effectively canceled but there is no significant increase in higher harmonics. This might not result in massive cancellation but still significant and will depend on the specific case.
I am pointing this out because in recent years some people with no real grasp (practice mostly) on the subject just took two working conditions they "liked" (basically one where the driver has minimum THD and one where the driver will cancel most of the 2nd without caring about the rest) to prove that it is bad and published it....that's exactly how it should NOT be done!
Once again, you are offering us the opposite of the truth. Cathode bias requires you to bias at a higher current, so that there is room to reduce it when signal arrives. Fixed bias does not need this.gabdx said:
As I said, it can be made to work for real music but it is not ideal. The bias shift will alter forward gain, which will either affect gain (if no feedback is used) or alter bandwidth (if feedback is used). Neither effect is desirable in a hi-fi amplifier, but fortunately hi-fi is generally used with real music so the effect will not be too great.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.