Tube biasing methods

I'm talking about low level circuits here, such as head amplifiers in a condenser mic... just studying all those tons of schematics from neumann, schoeps, akg, gefell etc. There are issues of basically the same mic but with a different biasing method in later versions. to be more precise, most very early circuits use a fixed bias for the grid, while later ones use self-biasing (meaning, grid resistor to ground instead of to a different point, plus cathode resistor with bypass cap to groubd).

i wonder why they turned to self biasing later on? improving of specs regarding noise? reducing unit-to-unit variations of tubes (since self-bias is the method less affected by it)?
also, do different biasing methods affect the dynamic behaviour of a circuit, i.e. the sound?

thanks for any thoughts, cheers,
volker

[quote author="volki"]I'm talking about low level circuits here, such as head amplifiers in a condenser mic... just studying all those tons of schematics from neumann, schoeps, akg, gefell etc.
also, do different biasing methods affect the dynamic behaviour of a circuit, i.e. the sound?
[/quote]
Yes, every device plugged into circuit modify sound.
Biasing elements are often resistors, some rare current sources, like
transistor or OTA output... .
Resistors are basically noisy elements. Every resistor generates noise.
There are several rare self - biasing circuits Diamond buffer
(OPA660 or discrete) can be example. But these are rare examples
of really electronically modern circuits. Some light from the future in
the electronic stagnation.
Mean, that in the future all circuits will be self - biasing and no resistors
(so unusual devices - only generates noise and heat) will be needed.

And in the tube? Biasing? Only chokes.
xvlk

What early neumann tube microphone does not use fixed bias? I don't know of a cathode biased one.

i have several schematics of different versions of mic's like m49/50, km5x, etc. - obviously, the respective a and b versions use fixed bias, the later c versions use self bias. km54a/b being an example of the former, km54c of the latter. not sure if all those a/b/c-versions are online somewhere, but just as an example:

SM23 - fixed bias with grid resistor referred to voltage divider in cathode circuit

http://www.silentarts.de/lab/Mikrophone/neumann/SM23_Schematic.pdf

SM69 - self bias with grid resistor to ground plus ac-bypassed cathode resistor

http://www.silentarts.de/lab/Mikrophone/neumann/SM69_Schematic.pdf

these are two entirely different mic's, but like i mentioned, they also did this with different versions of the same mic's, keeping the overall circuit and only changing the biasing arrangement. so what i'm wondering about is
- why this was done at all
- what benefits and drawbacks each biasing method has in this very application (impedance conversion with necessarily large grid resistor (to maintain LF response) and low noise figures).
- if the different methods have a noticable impact on the dynamic behaviour, or in other words - if the changeover from fixed to self bias is likely to have resulted in a change of sonics

i have read that fixed bias was the only way to bias directly heated tubes, so maybe a historical issue in some part, which they got rid of later on?
also, one method is supposed to enable higher input impedance than the other...
and/or maybe a headroom thing? the fixed bias in the sm23 circuit seems to include some modest cathode feedback for audio frequencies, looking at the high cathode resistance R13/14 paralleled by the relatively small C10...?

questions, questions, questions.... :roll:

[quote author="volki"]fixed bias with grid resistor referred to voltage divider in cathode circuit[/quote]

Just from the description, that's still sounds like a self-bias setup, but with current feedback?

Peace,
Al.

> i wonder why they turned to self biasing later on?

Because they could get good cathode bypass caps.

Film (oil/paper, poly) caps are impractical for cathode bypass: either not enough uFd for full bass response or WAY too big/costly. So tubes were usually grounded cathode with grid voltage.

Electrolytics made cathode bypassing practical.

"Good" depends what you are doing (and where you are). Electrolytics appeared in the USA around 1930, but early ones had large leakage and short life. OK for power bypass in consumer and light commercial gear where leakage could be tolerated and repairs were expected (and sold new gear).

But note that well into the transistor era, early 1960s, Langevin and others boasted of no-electrolytic design, because in large broadcast systems even if electros lasted 5 or 10 years the replacement down-time was a problem.

Inside a microphone, the cathode voltage is low (won't keep an electro well-formed) and available current may be very low (similar to electro leakage current). And you also have delicate parts and hi-Z nodes in there: you don't want clumsy/rushed technicians clobbering caps and mangling other parts inside.

BTW: a large plate/screen resistor is just as effective self-bias as a cathode resistor, if you can accept the loss of efficiency. Assuming amplification factor of 20, you can use a 1K cathode resistor or a 20K plate resistor and get about the same stability with variation in tube or supply voltage. And many of these mikes do use large plate resistors. If the tube gets hungry, plate voltage drops, tube eats less. In a pentode, a large screen resistor has the same effect. And the plate/screen voltages are high enough to keep old electrolytics well formed, or the effective impedance (about 20 times higher than cathode impedance) may allow using a film-cap. You need about 20 times less uFd but about 20+ times higher voltage rating: the rating of the thinnest paper is 100V-200V which is a good fit to plate voltages.

My first impression is that both circuits are cathode biased - in my simple understanding of guitar amps, I've alwaysunderstood that fixed bias involved a grounded cathode and a fixed negative voltage applied to the grid. What am I missing here? :?

PRR, thanks, veeery interesting. :grin:

But note that well into the transistor era, early 1960s, Langevin and others boasted of no-electrolytic design, because in large broadcast systems even if electros lasted 5 or 10 years the replacement down-time was a problem.

Click to expand...

that seems to correspond well with all those neumann tube mics whose circuits obviously were revised during that time, changing over from fixed to self bias. so up to that time there actually were no good quality 'lytics around which were suitable for very low power app's... wouldn't have thought.

so i guess the major problem with cathode bypassing using worse quality 'lytics would have been increase of noise due to the "irregular" leakage?

lovekraft, be careful with those mouse buttons :wink:
fixed bias doesn't necessarily involve the cathode being grounded. it does mean a fixed negative potential between grid and cathode, either by a) grounded cathode with the grid resistor referred to a negative voltage or b) positive cathode voltage by means of two resistors to ground, to the middle of which the grid resistor is referred. you adjust the bias by the resistance ratio of the resulting voltage divider.

AWGH. so it wasn't your mouse button... :shock: :roll: :?
obviously a system defect rather. which made multiples of my previous post appear minutes after i had unsuccesfully tried to submit it. anyway, deleted the multiples.
cheers!

PRR, one thing though... can you imagine what exactly allows for lower noise figure in a cathode/sel-biased circuit as opposed to a fixed one? for mic's, at least. because when you e.g. look at the data-sheets of different versions of neumann tube mic's, the later (self bias) versions are less noisy indeed.