[Roger A. Modjeski]

I was looking for threads concerning the sonic pros and cons of autobias and came upon the Prima Luna page.. http://www.primaluna-usa.com/index.php?option=com_content&view=article&id=118&fb_comment_id=fbc_287873517993116_2045910_394667337313733#f22d93128

I posted the following.

This information is purposely misleading with several untrue and self conflicting points. On the one hand the adjustment is "adapting instantly" and on the other hand it is "steering 5 octaves below the audible region". A well designed amplifier will hold bias for a year at at time not as stated above "re-biasing is needed every few months".

I am looking for threads concerning the sound of auto-bias (cathode bias) vs user set bias (fixed). Adaptive bias like the Prima Luna is yet another class of bias.

[FullRangeMan]

I like self-adjusting Bias or manual Bias.
I voted: I prefer an amplifier with bias adjustments and a built-in meter but just remembr the meter will be connected in the circuit all the time. This is the same prob with the AutoBias, that yet keep Biasing the tube all the time.

So I change my vote to: I prefer an amplifier with bias adjustments and easy access test points
Good survey.

[Roger A. Modjeski]

I like self-adjusting Bias or manual Bias.
I voted: I prefer an amplifier with bias adjustments and a built-in meter but just remembr the meter will be connected in the circuit all the time. This is the same prob with the AutoBias, that yet keep Biasing the tube all the time.

So I change my vote to: I prefer an amplifier with bias adjustments and easy access test points
Good survey.

I would have an off position for the meter so it would also not be damaged if a tube shorts. However the meter, switch and associated wiring likely adds $300-500 to a $2000 amplifier depending on how fancy the meter is.

[Ericus Rex]

Is Adaptive Bias a marketing gimmick?  Or is an accepted, though rarely used, form of bias?

[SteveFord]

I was under the impression that if a tube goes on a self-biasing amp it can take out all of the others.
Is this correct?

[FullRangeMan]

I would have an off position for the meter so it would also not be damaged if a tube shorts. However the meter, switch and associated wiring likely adds $300-500 to a $2000 amplifier depending on how fancy the meter is.

Nice solution, very smart.
As I like to adjust the Bias to know the tube situation and in view of the cost I think it not worth the additional cash.
Some tube owners already have a standalone hand meter.

But audiophiles that dont love tubes may like this built-in meter.

[Roger A. Modjeski]

Is Adaptive Bias a marketing gimmick?  Or is an accepted, though rarely used, form of bias?

Hi Ericus.

I found your post from a few years ago on bias right on. Adaptive bias has no standards so its up to the designer how to go about it. The main problem is how to measure the bias in presence of signal because the signal adds to the bias above a few watts. Most of these circuits I have seen or imagined would employ a simple IC servo to adjust the negative grid voltage to maintain a reference current in the tube. The servo would have a long time constant, however when the amp is played above a few watts the tube current increases and the servo would then make the grid more negative  to reduce the signal plus bias to the bias level. When the signal is reduced (soft passage) the output tubes would likely be cut-off.

What would work, and perhaps someone is doing this is to lock out the adjustment when signal is present and hold the level. This would require some sort of simple dedicated hardware and software. Could be done. Does Prima Luna do this. They don't say as far as I can tell.

As far as Prima Luna's claim of reducing distortion, again not enough is said to know if they are comparing their amps bias to an amp where the bias is way off.

[pehare]

Roger - I'd keep your amps exactly as they are - leave well enough alone.  Biasing is so simple on yours & rarely needed why complicate matters.  I taught a friend how to bias his Cary amps & now he is more confident w/ tubes in general.

cheers, patrick

[Roger A. Modjeski]

Roger - I'd keep your amps exactly as they are - leave well enough alone.  Biasing is so simple on yours & rarely needed why complicate matters.  I taught a friend how to bias his Cary amps & now he is more confident w/ tubes in general.

cheers, patrick

Thanks Patrick,

This poll and post are for a future integrated amp for people new to tubes who might not want to even get close to a meter or any adjustment. The existing amps are unchanged. I have added a cathode biased version of the RM-10 that has almost the same power (25 watts) and nothing to adjust. One further advantage is that the tubes stay well balanced throughout their life. There is one LED per output tube to show that the tube is functioning properly.

[FullRangeMan]

If the owner manual teaching how to adjust the Bias the amp may have manual Bias, which will result in better sound and less cost.

[Roger A. Modjeski]

If the owner manual teaching how to adjust the Bias the amp may have manual Bias, which will result in better sound and less cost.

The cost of manual vs. cathode bias is virtually the same. With cathode bias tube matching is less critical and tube aging is more forgiving.

Although I like manual bias as it gives me the most RMS power (and that's about all it does), I cannot say that it influences the sound at low and moderate power levels below clipping. Manual biasing will result in better sound if the amp is playing full out, however this is such a rare situation. It is clear to me in conversations with users the many have much more power than they are using.

Unfortunately the FTC mandates concerning power measurement forces my hand as to bias methods. Cathode bias, which I do like, performs poorly at full power and I have done research that proves to me that most cathode bias amp designs (those whose parameters are given in the tube manuals) suffer due to this measurement requirement. The much loved Dynaco ST-35, Fishers, Scotts and others using EL-84s and cathode bias fall into this measurement quagmire. In other words, in order to get a cathode biased amp to measure decently at full power the circuit has to be detuned from what it would be if we designed it for music.

 Does anyone have links to the topic of fixed vs cathode vs automatic bias schemes. I am interested in knowing what is being said in the forums.

[FullRangeMan]

Cathode Bias can operate in manual Bias?? or it is just suited to auto bias??

I dont read this link yet, but it seems good stuff:
http://www.diyaudio.com/forums/tubes-valves/90671-self-bias-fixed-bias-big-tubes.html

[opnly bafld]

Does anyone have links to the topic of fixed vs cathode vs automatic bias schemes. I am interested in knowing what is being said in the forums.

Didn't find much with a quick search:

http://www.audiokarma.org/forums/showthread.php?t=434801

http://www.diyaudio.com/forums/tubes-valves/74371-fixed-bias-vs-auto-bias.html

[Carl V]

conrad-johnson amps have what______________?
seems simple to me...easily accessible, stays put a long time.
I've owned others where I had to use a meter fomr underneath...hassle.
As I understand it self biasing has some limitations, if they are audible I don't
know.

[Roger A. Modjeski]

Cathode Bias can operate in manual Bias?? or it is just suited to auto bias??

I dont read this link yet, but it seems good stuff:
http://www.diyaudio.com/forums/tubes-valves/90671-self-bias-fixed-bias-big-tubes.html

Thanks for the link. I am looking for new commentary as I was told some existed.  The tubes noted at the beginning of the link like the 845 and have such high bias voltages (greater than -100V) that developing that bias over a cathode resistor would produce a lot of heat.

Cathode bias is the technical (better) term for autobias, a term which is confusing as it implies something automatic going on. However in Europe it appears they commonly call cathode bias auto bias. For our discussion here I would rather not use that term. I suggest we use the proper terms:

1. Fixed Bias for amps with bias pots that put a negative voltage on the grid and the cathode is very close to ground. In these we measure the "bias" as a small voltage drop across a resistor typically 1 ohm which gives 1 mV per mA. However saying our "bias is 50 mV" as people do is incorrect. The correct statement is that our idle current is 50 mA which we measure for convenience across a 1 ohm resistor. (other resistor values may be used but they are generally 10 ohms or less).  The correct use of the term bias in a tube amp is to say " I have adjusted the negative grid voltage (the bias) on the tube so that the idle current is 50mA and for this tube in this circuit that bias voltage is -35 volts (typical for a EL-34).  If we put a KT-88/6550 in that socket to achieve the same idle current we will have to increase the negative bias to -45 volts (typically) and if we don't the tube will draw excessive current and run red hot. Keep in mind that a given tube within the same type can easily be 10 volts more or less negative so our bias pot allows us to adjust for that also.

My RM-9 allows the use of a wide range of rather different tubes for no other reason than I have a large range on the bias pot that encompasses all. Why other manufacturers don't to this I have no idea. Although David Manley admitted that he didn't want people putting any tubes other than the ones he wanted into his amplifiers. Such was the topic of the Manley-Modjeski letters in Stereophile a long time ago.

Sorry for the long definition but its all in the language. If you start speaking this correctly it will likely enhance your knowledge of what is going on.

2. Cathode Bias (sometimes called auto bias, especially in Europe where tubes are called valves) This method uses a large, often high power (10-20 watt) resistor in between the cathode and ground. If the tube is an EL-34 it is chosen to develop 35 volts across it at 50 mA. If the tube is a KT-88 the resistor has to be larger in value and size to develop 45 volts for 50 mA.  This is why there are some amps that have a switch in the back to choose the other resistor when you choose these different tube types.  It's all terribly simple and one does not then have to provide pots and jacks and a negative bias power supply. The amp runs a bit hotter for the extra watts (only 5 watts per tube) and takes care of itself. The way it does this is rather interesting. On the one hand you can simply say that the cathode resistor degenerates the DC gain of the tube. (There is almost always a capacitor bypassing this resistor to restore the AC gain.) By doing so as different tubes are substituted in the amplifier their differences will be degenerated (made smaller) to the point that they all run close enough to the same current (10% is good enough) to not degrade the performance of the amplifier. Matched tubes are still a good idea but the matching is less critical. The down side of cathode bias is that at full power the cathode voltage rises, reducing the current in the tube, reducing power and increasing distortion. However if we are playing the amplifier at low to moderate levels we do not reach this point. Note that all cathode bias amps are not equal. The Dyna ST-35 and many of that period used a single cathode resistor for all four output tubes. This is a horrible idea and requires very closely matched tubes. I am aware that some mods correct this design flaw that was caused by economy if nothing else.

3. Servo or computer driven bias (for want of a better term). Here an additional solid state circuit that is not in the signal path will monitor the tube current and hold it steady. However this circuit has some time constant and can be confused by the music. There are likely some schemes where a computer of sorts measures and adjusts the bias when the signal is not present or very low.

Servo bias compares the cathode current to a reference value and holds the tube there by adjusting the negative voltage on the grid. Although simple to make, this circuit will make the grid more negative during loud passages and will need time to recover on soft passages. I don't like the idea of my bias moving around all the time. A servo is no better than cathode bias, though it does eliminate the cathode resistor and can have a longer time constant though not long enough to pass John Atkinsons (or any other) amplifier tests. How well these work is totally up to the designer but like any computer if something goes wrong you have no control over it. Personally I am not interested in either of these schemes.

There are the three. Any others?

[Ericus Rex]

1. Fixed Bias for amps with bias pots that put a negative voltage on the grid and the cathode is very close to ground. In these we measure the "bias" as a small voltage drop across a resistor typically 1 ohm which gives 1 mV per mA. However saying our "bias is 50 mV" as people do is incorrect. The correct statement is that our idle current is 50 mA which we measure for convenience across a 1 ohm resistor. (other resistor values may be used but they are generally 10 ohms or less).  The correct use of the term bias in a tube amp is to say " I have adjusted the negative grid voltage (the bias) on the tube so that the idle current is 50mA and for this tube in this circuit that bias voltage is -35 volts (typical for a EL-34).  If we put a KT-88/6550 in that socket to achieve the same idle current we will have to increase the negative bias to -45 volts (typically) and if we don't the tube will draw excessive current and run red hot. Keep in mind that a given tube within the same type can easily be 10 volts more or less negative so our bias pot allows us to adjust for that also.

Roger,
I really appreciate you taking the time to thoroughly explain these bias schemes.  I can't tell you how many times I've heard people hear and elsewhere talk about adjusting their fixed-bias amp to such-and-such mV without understanding that the current is what they should be monitoring.  I've always thought of that approach as being similar to the speedometer in your car displaying gallons-per-hour consumed instead of speed.  But, you try and explain that to them and they say their manuals say to do that.  In regards to this biasing to voltage instead of current, isn't it also true that plate resistance can vary from tube to tube yielding a different plate current for each tube though they may all be 'biased' to the same mV reading?

Just to be clear, (almost) all voltage-gain/driver tubes are biased with cathode bias, correct?  Is there any application where using fixed bias for them would be beneficial?

[Roger A. Modjeski]

Roger,
  In regards to this biasing to voltage instead of current, isn't it also true that plate resistance can vary from tube to tube yielding a different plate current for each tube though they may all be 'biased' to the same mV reading?

Just to be clear, (almost) all voltage-gain/driver tubes are biased with cathode bias, correct?  Is there any application where using fixed bias for them would be beneficial?

The plate resistance of a pentode is so large that it doesn't matter and even a triode with low plate resistance it matters not. When we set the idle current (mV across our cathode resistor) in a tube we are done. All done and nothing else matters. By the way, we bias the tube to reduce the cathode current from its maximum value which occurs at zero grid volts which typically occurs at full output. By contrast bipolar transistors and MOSFETS are always off and we have to bias them on to their idle value.

Your second point is correct, most voltage gain stages are cathode biased because they run very class A where the idle current is so large compared to the signal that the signal can't influence it. There are some DIYers using LEDs in the cathode of a 12AX7 etc for bias. This idea demonstrates a lack of understanding on their part and it is a very bad idea. Perhaps they do it so they can see the tube is on. Regardless, this practice or any other method of fixing the bias on such a tube removes its ability to adjust itself as it will in cathode bias making the amp very tube fussy.

[G Georgopoulos]

By the way, we bias the tube to reduce the cathode current from its maximum value which occurs at zero grid volts which typically occurs at full output. By contrast bipolar transistors and MOSFETS are always off and we have to bias them on to their idle value.

similar to jfet 

[Roger A. Modjeski]

similar to jfet 

Actually most JFETS are depletion mode devices like tubes that need to be biased off. It's only the MOSFETS and a few JFETS that are enhancement mode which means they have to be biased on. Also note that power MOSFETS have a very abrupt turn-on making it difficult to set and hold a steady idle current.

[G Georgopoulos]

Actually most JFETS are depletion mode devices like tubes that need to be biased off. It's only the MOSFETS and a few JFETS that are enhancement mode which means they have to be biased on. Also note that power MOSFETS have a very abrupt turn-on making it difficult to set and hold a steady idle current.

Hi Roger
actually
jfets similar to tubes
mosfets enhancement mode similar to bjt
mosfets depletion mode similar to jfets and tubes

kind regards