Tube Biasing Questions

[micaddict]

@ granger.frederic.

My MK47 is not noisy.

I agree with your remark regarding dynamic mics BTW.

 

[Matador]

Hi Matador! Thanks! This indeed helps, but I have a few questions. You said "If you keep the B+ the same, this reduces the slope of the line and makes it "flatter".  Quiescent current will drop." Doesn't keeping B+ the same, but reducing the plate resistor increase the negativeness of the slope of the load line, not make it "flatter"? I thought quiescent current will go up, not drop. If I used 50k Rp instead of 150k, my mA at 0V would be 2.4mA instead of 0.8mA. The load line has gotten more negatively vertical, not flatter, and at -1V grid, we're now at a higher current than when we used 150k. Again, I may be misunderstanding how to use/read these graphs.

Sorry yes, you are right.  I was thinking in terms of plate resistance, not 1/plate resistance which is the actual slope.  Reducing the plate resistor indeed tilts the load line "more negative" (e.g., more towards vertical).  Sorry if that was confusing!

 

[Matador]

What resistor do you change to adjust a fixed bias? The plate resistor? What exactly is the purpose? Does the plate output impedance increase as the tube ages?

On adjusting fixed bias:  that's hard to answer without looking at the individual design.

For a U47, the 29ohm resistor can be adjusted slightly to change the generated cathode voltage.  In a C12, the negative voltage can be adjusted by changing the resistance of the ground return resistor (R8 on the original C12 schematic).

 

[Melodeath00]

Hi Matador! Thanks! This indeed helps, but I have a few questions. You said "If you keep the B+ the same, this reduces the slope of the line and makes it "flatter".  Quiescent current will drop." Doesn't keeping B+ the same, but reducing the plate resistor increase the negativeness of the slope of the load line, not make it "flatter"? I thought quiescent current will go up, not drop. If I used 50k Rp instead of 150k, my mA at 0V would be 2.4mA instead of 0.8mA. The load line has gotten more negatively vertical, not flatter, and at -1V grid, we're now at a higher current than when we used 150k. Again, I may be misunderstanding how to use/read these graphs.

Sorry yes, you are right.  I was thinking in terms of plate resistance, not 1/plate resistance which is the actual slope.  Reducing the plate resistor indeed tilts the load line "more negative" (e.g., more towards vertical).  Sorry if that was confusing!

Thank you! So since I can't use the EF80 chart for the EF802... If I want to try to lower the output resistance of the tube, should I just try lowering the 51kohm resistor connected to the plate? Are there any other considerations to be aware of? I have read that the more negatively vertical slope of the load line also results in greater 2nd harmonic distortion. More "tube sound" saturation, I guess, but increased frequency range and transient response from the better impedance match with the transformer?

Do I have any other options besides changing the cathode voltage? Going lower than 1.1V could be "dangerous" if it gets close to 0V grid, and increasing the cathode would seemingly just raise the output impedance.

 

[Matador]

Thank you! So since I can't use the EF80 chart for the EF802... If I want to try to lower the output resistance of the tube, should I just try lowering the 51kohm resistor connected to the plate? Are there any other considerations to be aware of? I have read that the more negatively vertical slope of the load line also results in greater 2nd harmonic distortion. More "tube sound" saturation, I guess, but increased frequency range and transient response from the better impedance match with the transformer?

Do I have any other options besides changing the cathode voltage? Going lower than 1.1V could be "dangerous" if it gets close to 0V grid, and increasing the cathode would seemingly just raise the output impedance.

What is it you are trying to accomplish with the change?  Are you looking for a different sound?  More/less gain?  More or less distortion?

 

[Melodeath00]

Thank you! So since I can't use the EF80 chart for the EF802... If I want to try to lower the output resistance of the tube, should I just try lowering the 51kohm resistor connected to the plate? Are there any other considerations to be aware of? I have read that the more negatively vertical slope of the load line also results in greater 2nd harmonic distortion. More "tube sound" saturation, I guess, but increased frequency range and transient response from the better impedance match with the transformer?

Do I have any other options besides changing the cathode voltage? Going lower than 1.1V could be "dangerous" if it gets close to 0V grid, and increasing the cathode would seemingly just raise the output impedance.

What is it you are trying to accomplish with the change?  Are you looking for a different sound?  More/less gain?  More or less distortion?

My goal is simply to lower the tube output impedance. Visually, it looks like the EF80 (and likely its family of tubes) has a higher plate output impedance than the VF14 based on the slope of the tangents at the operating point. Once I can manage that, then I would probably do some tests and see if it was actually audibly different from how it's set/biased now

 

[Matador]

My goal is simply to lower the tube output impedance. Visually, it looks like the EF80 (and likely its family of tubes) has a higher plate output impedance than the VF14 based on the slope of the tangents at the operating point. Once I can manage that, then I would probably do some tests and see if it was actually audibly different from how it's set/biased now

Ahh ok, good!

The simplest way is to lower the plate resistor.  The output impedance of any common-cathode tube stage is roughly the plate resistor in parallel with the plate impedance (rp).  Lowering the plate resistor will reduce this parallel quantity.

Once you sub in the new resistor (assuming that you also didn't change B+), you would then select a new bias point.  If you want the cathode voltage to stay the same (1.1V I think), you may need to trim some of the other resistances.  Gain will also drop, however bandwidth should increase.  Damping factor will also increase (wrt. the output transformer).

So for a U47 application, you will likely end up fiddling with/lowering the 29 ohm resistor.  Also, keep in mind the passive nature of the U47 supply:  you will be drawing a higher idle voltage, which means the series resistances in the passive supply will be dropping more voltage, which means that your 105V may drop.  You may need to resize those resistors downward to compensate the B+ back up to 105V.

 

[Melodeath00]

My goal is simply to lower the tube output impedance. Visually, it looks like the EF80 (and likely its family of tubes) has a higher plate output impedance than the VF14 based on the slope of the tangents at the operating point. Once I can manage that, then I would probably do some tests and see if it was actually audibly different from how it's set/biased now

Ahh ok, good!

The simplest way is to lower the plate resistor.  The output impedance of any common-cathode tube stage is roughly the plate resistor in parallel with the plate impedance (rp).  Lowering the plate resistor will reduce this parallel quantity.

Once you sub in the new resistor (assuming that you also didn't change B+), you would then select a new bias point.  If you want the cathode voltage to stay the same (1.1V I think), you may need to trim some of the other resistances.  Gain will also drop, however bandwidth should increase.  Damping factor will also increase (wrt. the output transformer).

So for a U47 application, you will likely end up fiddling with/lowering the 29 ohm resistor.  Also, keep in mind the passive nature of the U47 supply:  you will be drawing a higher idle voltage, which means the series resistances in the passive supply will be dropping more voltage, which means that your 105V may drop.  You may need to resize those resistors downward to compensate the B+ back up to 105V.

I'm starting to understand! You said "Once you sub in the new resistor (assuming that you also didn't change B+), you would then select a new bias point." However, the 1.1V on the cathode IS the bias point, right?

I do not mind a gain drop, however, what is the effect of increased damping factor with the transformer?

A few things... This is a U47-esque mic; not exactly the same. It's a little different. It has a separate heater supply, and I'm running at 120V, not 105V. I do have a 29ohm resistor in there, but it's a little different from a real U47 in that it forms a voltage divider with a 100ohm resistor tapped off the heater. This results in 1.1V when the heater is set to 5V, and is attached to the heater. See the attached schematic to get an idea (R8 is 51kohm, not 30kohm, and B+ is 120V, not 105V, so that changes R6 and R7 to 1M each to get 60V polarization on the capsule).

I did not design the PSU, but I was told it's NG, which would mean passive, I think. It certainly has a trimmer for adjusting the heater, and I bet it has a trimmer for adjusting B+. So if I swapped the 150k plate resistance (R4 and R8) to something lower, I might need to trim the PSU back up to 120V?

Thanks so much!

 

[Matador]

Ahh I see.

In reality, the U47 is essentially "fixed bias", since it takes a big current change to make the voltage across the 29 ohm cathode resistor change.  Even if you doubled the current, it's small compared to the heater current already flowing through that resistor.  So to the tube, it looks like it's fixed bias.

So yes, in reality, even if you halved the plate resistor and doubled the plate->cathode current, it probably wouldn't change the 1.1V very much (each additional milliamp makes the cathode voltage increase by 29mV).

Damping factor can be hard to describe audibly with any certainty:  this can be best seen with square wave response, but that's not a real audio signal in any case.  The resonant frequency of the RLC combination doesn't change when changing output impedance (e.g., the low frequency resonance), but the peak magnitude as resonance does change.

In any case, give it a try and let us know what you think.

 

[Melodeath00]

Ahh I see.

In reality, the U47 is essentially "fixed bias", since it takes a big current change to make the voltage across the 29 ohm cathode resistor change.  Even if you doubled the current, it's small compared to the heater current already flowing through that resistor.  So to the tube, it looks like it's fixed bias.

So yes, in reality, even if you halved the plate resistor and doubled the plate->cathode current, it probably wouldn't change the 1.1V very much (each additional milliamp makes the cathode voltage increase by 29mV).

Damping factor can be hard to describe audibly with any certainty:  this can be best seen with square wave response, but that's not a real audio signal in any case.  The resonant frequency of the RLC combination doesn't change when changing output impedance (e.g., the low frequency resonance), but the peak magnitude as resonance does change.

In any case, give it a try and let us know what you think.

Matador, is it the load line and cathode/grid that determine the plate voltage? In other words, if you raise the cathode to 2V (grid becomes -2V), do you follow your load line right and down to  where it intersects with the -2V grid line,  then go straight down, and that is where idle plate voltage will be? I don't know why I find this so confusing/incomprehensible.

 

[Matador]

Matador, is it the load line and cathode/grid that determine the plate voltage? In other words, if you raise the cathode to 2V (grid becomes -2V), do you follow your load line right and down to  where it intersects with the -2V grid line,  then go straight down, and that is where idle plate voltage will be? I don't know why I find this so confusing/incomprehensible.

Yes, exactly!  You can see that as grid gets more negative, tube turns off more, current falls, which means plate voltage rises (since less drop across the plate resistor).

 

[granger.frederic]

as a mic design usualy use an ultra high grid impedance (way higher than the 1 megaohm range of the manufacturer grid resistor test)and a very low plate voltage , sadly, we can't use the slopes in the datasheets...

we can't even use most of the specs in the datasheets

usual tube theory design is not relevant in microphones

the easiest way is to bias the tube and listen...
turn around 1V  (0,8v to 1,6v) and choose the best value
chances are good that you won't hear a big difference...
it's easier to test with loud sources , and with percussive sources as well

damping factor should be negligible as the input impedance of the preamp is almost constant/stable and its input capacitance is small and constant also

 

[Matador]

we can't even use most of the specs in the datasheets

usual tube theory design is not relevant in microphones

This has definitely not been my experience.  I have not seen the traditional gm,mu, and rp triad change much based on the grid impedance.

Noise spec-wise I would agree:  tube types with "leaky grids" (or high inter-electrode cap coupling from other internal components like heater coils, etc) tend to be quite noisy, which means the only way you can know the noise performance is to stick it in and test it.

 

[Gus]

Years ago I posted about a test I did on a few tubes that might be used in microphones.
I tested at typical microphone operating points.
I got some odd reactions so I pulled the post and test schematic and have not shared much else I have tested with microphones

It is funny to read some of the C12 biasing posts I have found on the web. 
People seem to have to label things.  It is partly cathode biased with an added filter to make it work something like fixed bias and there few other things going on if you look close. 
So it should be called C12 biasing if you need to label the biasing

 

[granger.frederic]

All the parameters in the tubes datasheets are given for a typical operating design : usualy around 120V plate , grid resistance around 100K , and rarely in triode mode for pentodes ...
in a mic , things are very different and the tube often works in a non linear part of its slope ....

 

[mhelin]

Hit this thread when trying to figure out what kind of biasing scheme use with my U47ish build. The self biasing appears to be the most recommended way  except with a real VF14 tube, though some designs use the similar cathode biasing as on U47 also with the bias derived from the heater supply - in case of EF80/800 etc. using a 100 ohms resistor to get the required about 38mA current to cause the 1.1V drop over the 29ohms resistor. In this case the anode or cathode current (idle) is targeted to 0.5 mA. This is obviously convenient but requires very clean heater supply. I'm just wondering how would such biasing work in comparison if you used something like 200 ohms cathode resistor bypassed by a capacitor, but adding a resistor from B+ to cathode to help getting the bias to correct voltage level but not entirely fixing it. In this case from 120V B+ the resistor should be about 24kohms to cathode to make the current such that the voltage drop would be close to 1.1V (1.09V) when the 0.5mA is added to that current from B+ resistor (200 ohms*(120V/24200ohms + 0.0005A)). Would that kind of biasing which is somewhere (TBD) between full self bias and U47 almost fixed bias be good for anything? I mean it would be stiffer as self bias but less noisy than fixed bias, right?

Might actually use 15k B+ to cathode and 130 ohms cathode resistors for increased current of 8 mA (or even 12k and 110 ohms for 10mA). Only 1/5 of what's used on U47 though. Full "fixed bias" using the 29 ohms resistor would require a  3100 ohms @120V (2700 ohms @105V) B+ to cathode resistor but that would then consume about 4.5 (4 @105V) watts of power.

 

[mhelin]

Ended up using diode biasing (for now) because the (T.Bone/Alctron) PSU couldn't supply the power needed for the biasing I was going to attempt first (^).  I use two 1N4448 diodes in series with a 220k resistor from cathode to B+ for about 0.5 mA current to  bias the diodes to make sure they keep the voltage around 1.2V (+/- 0.1 mA or so depending on input voltage, tube gm and such things).  You have to check the diode datasheet's "Forward Current vs. Forward Voltage" table (IV curves so to say) to find the current vs. voltage you want to get. So no, even the diode biasing can't give perfectly fixed bias unless you find a diode or diodes which are linear or can be made using some circuit around the diodes to keep the current constant (like in CCDA where perfect diode bias is easy to implement).

 

[Gus]

In an old post I showed adding current from the filament supply to a LED in the cathode to ground section, this was shown in the square microphone schematic. I also bypassed the diode with a good low ESR cap.
I used a LED because I wanted to be able to change the heater voltage and keep the same cathode voltage if I did not want this calculating  fixed resistors in a circuit like is shown in the Oliver circuit is what I would do using a filament supply.

IMO you need extra current in the diode. I think the ones with just the diode are a mistake.

However you do this, from B+ or the filament supply make sure the supply is low noise.

Pay attention to what ever diode you use voltage drop change with forward current, some are more constant than others. You can look up specs or do a simple test with a voltmeter, resistor and a variable voltage supply.

 

[Delta Sigma]

...The method used in the U47A avoids the large capacitor but minimises the negative feedback.

The M49b is quite very similar but the current due to the heater voltage is much smaller and the grid voltage is tapped off the divider rather than the cathode voltage being tapped off. To be honest, the values for R6 and R7 don't make sense to me....

M49b cathode to ground is around 40ohms. I'm not sure what the U47 is 29ohms || VF14M heater but it looks like they're fairly close. Aren't R6/R7  too big to have an affect on the negative feedback?  They do increase the grid resistor to 165k though and are small enough to measure the grid bias with a voltmeter.

 

[Delta Sigma]

The best way is when the cathod is grounded. Its most noiseless way. For example: U67.

What is the purpose of C7? If they wanted a 400M grid resistor value, could they not have skipped C7 and used a smaller R8?