desirable characteristics for microphone tubes?

[soapfoot]

Just curious if anyone has any insight into the desirability of certain tube data characteristics in condenser microphone use.

I'm looking for things like this:

If looking at two tube data sheets, would the following factors have any value as predictors of which tube might theoretically/generally work well in a typical plate-follower microphone circuit?

Amplification factor?  Say one tube has a higher mu than another.  Is this theoretically desirable, undesirable, or inconsequential in microphone head-amp design?

Transconductance?  Say one tube has twice the transconductance of another similar tube.  Which would you be more likely to try first?

Internal Resistance  Say one tube has a very high internal resistance and the other a very low internal resistance.  I assume this would affect how well the tube would mate with the chosen transformer, or influence the choice of transformer, correct?  If you had to choose, would you shoot for a tube with a very low Ri, or a higher Ri?

 

[PRR]

Mu, Gm, and Rp are three sides of the same coin.

Rp=Mu/Gm
Gm=Rp/Mu
Mu=Rp*Gm

Mu, Gm, and Rp are NOT "constants". Mu is semi-constant. Gm and Rp vary a LOT with current.

With a free choice of current, we might (and might not) decide to run a huge tube at high current for high Gm and low Rp. Large current often implies large voltage, and thus large power.

Inside a microphone, this is NOT a "free choice". A video-amp or pass-reg tube would burn the paint, limpen the diaphragm, and blister the assistant. Heat inside a microphone must be small.

At lower current, Rp rises and Gm falls.

*Generally*, tubes of similar heater power and price (mass-market or microwave) working at the same current and voltage have similar Gm. Some "improvement" from 1930s types to 1960s types (though "improved" Gm is often less-linear Gm).

Data-sheets often cite a "show-off" condition, rarely what we want. You must read between the lines to estimate tube parameters at smaller currents.

You can rough-assume the Gm falls as square-root of current. That's good enough to extrapolate from 4mA spec to 1mA goal, not from 100mA to 1mA.

Whether Rp or Gm guides design depends on cathode-follower or plate-loaded.

You have not touched on noise (hiss). That's a secondary reason you won't find 100mA tubes in mikes, their grid current is high and that becomes hiss. The ordinary small tubes work OK. When mike companies could special-pick tubes, lower grid current types were used.

Short answer: tube type is NOT a key issue for a budding mike designer. Capsule and transformer make FAR more difference. Plagiarize. Very ordinary tubes such as 6Au6 and 12AT7 have worked well.

 

[abbey road d enfer]

Everything PRR says, plus:
For obvious safety reasons, tubes in microphones are generally run at lowish voltages (50-100V Vak from a 100-200V supply). In order to minimise grid current, they're run at low current (<1mA).
The output impedance of the tube stage will be quite high as a result. Cathode follower offers lower output Z but drive capability is the same as plate load. The use of NFB is also a factor.
Spec sheet don't tell all because the tubes in a mic are used way out of specs. I have never seen a tube spec sheet saying you can use it with a 1Gohm grid resistor, but that's a necessity.
The most revered tube for mic applications, the VF14 (a tube originally designed for low-voltage radios), is used in the U47 at an operating point that's clearly out of the chartered zone, even the heater voltage is different!
Clearly Neumann's designers went out of their head to extract the ultimate drop of performance out of tubes, but the much simpler approach of a large number of mic manufacturers today, which is basically to use a cheap and available tube in a simple basic circuit in combination with a good capsule works well enough for most applications.

 

[soapfoot]

Thanks for your responses and insight.  Thinking about it more, another reason the data sheets won't tell "all" is because when talking about pentode tubes wired for triode operation, the data sheets rarely will give numbers for triode operation.  So would the information measured with the tubes operating as pentodes even be relevant at all?

If I may get more specific for a moment--

"Plagiarism" is indeed what I'm doing... I'm building a 47-inspired microphone probably using an EF-800 tube, a BV-8 style transformer (AMI/TAB), and a nice capsule (probably blue line Thiersch).

However, I've become curious about the EF-184.  I've seen in literature where the EF-184/814 is is the successor to the EF-80/800.  Then I see where Bock uses an EF814, specially selected, in his high-end microphones like the 5-ZERO-7, saying that it's characteristic was very similar to the VF-14, only without the steel tube.

Looking for EF-814, which from everything I can find is most-likely an "improved" version of the EF-184, there's very little to be found.  Hard to find data sheets, and can't find even a single one for sale. I'd imagine that Bock found a large stock for their microphones and to supply for their customers for years to come.

Looking at the data sheets for the "regular" EF-80 and EF-184, the specs are similar.  Pinout is the same, heater current is nearly the same, heater voltage the same.  With a constant plate current of 10mA, transconductance and gain is higher in the EF-184, internal resistance is higher in the EF-80.

Does this give me any meaningful information about whether or not it's "worth it" to try and find a quiet EF-184?  They're only a few bucks each, so I might pick up a couple to try (since they have the same pinout and very similar heater requirements).  But if it's unlikely to perform much better/differently than an EF-800, I might just not bother.

 

[strangeandbouncy]

Hi Soapfoot,



  I think Oliver was suggesting ef42, ef80, and ef802, not ef800 . . .


  I could be wrong . .



  ANdyP

 

[Sredna]

I could be wrong . .

Yes indeed!  

 

[pstamler]

The other important characteristic, on which most datasheets say precisely nothing, is how microphonic the tube is. That's where the selection happens; you get several tubes and use the one which is the least microphonic.

Peace,
Paul

 

[Johan]

6ak5/EF95 is easy to find doesnt cost much and has been used in tube mic's before...google SELA( now long defunked). Swedish made Neumann knockoffs made for movie industry where distant micing makes noise a big concern...

 

[strangeandbouncy]

sredna,


 Very intersting stuff. But i took my info from another forum , Neumann I think. i cant find it now. I remember it well, because I went out and bought some EF802s on this recommendation. I still havent used them. He suggested,iirc,  EF172, EF42, EF80 and EF802 as the following generations.


  ANdyP

 

[strangeandbouncy]

Johan,


 true, but they had to select them, and had loads of rejected tubes(I remember reading). ALso consider the 6au6(EF94). Sounds amazing in my Schoeps(philips)CM61. Having a hard time trying to find a REALLY quiet one tho. In CM61, output transformer is Haufe 14-1, same as most AKG C12. Very reasonably priced, and still being made.


 ANdyP

 

[strangeandbouncy]

found the article, and I quote Oliver on the 27-2-2000


   


I tried nearly all tubes that are out there that are somehow useable for VF14 replacement. Only ones that realy work are the Telefunken made EF14 (UF14), the E(U)F42 and the EF802s. All those tubes have nearly the same system construction



  Kindest regards,


    ANdyP

 

[Sredna]

Interesting,

seems that he changed his mind!  

...or that he has different opinions about VF14 replacements and tubes for U47.  

http://www.tab-funkenwerk.com/id68.html

http://www.tab-funkenwerk.com/id85.html

 

[strangeandbouncy]

  i am no expert, but it seems the 802 is virtually identical to 800 . . .


  anyone . . .

 

[Sredna]

Found this:

The EF802  is similar to EF800, but without the anode shield, thus low output capacitance (about 1.8 pF) and therefore well suited for amplifiers with a maximum of 30 MHz bandwidth. High slope. Through special manufacturing measures the life span compared with normal types is greatly extended, moreover, the tube has less hum and microphonics. Information from "Funktechnik" No 7-1952, page 176: "Telefunken developed a series of special tubes for long life and some special properties for commercial purposes, such as the Federal Post and broadcast companies. It is the EF800, EF802, EF804 and EF804S."

 

[Wagtbtoobz]

I didn't see anything about "microphonics" in this thread  ;-)
The EF184 and all this family is really poor on low signals ... like a singing cowbell ! The subminiature toobz are far better : 5703, 5840 and 5719... a special AF triode , like the AC701!  ;D

 

[soapfoot]

The EF184 and all this family is really poor on low signals ... like a singing cowbell !

Good to know, thanks!  I'll save my time.

I am still interested in the ORIGINAL question, however-- in theory, how does transconductance, mu, and R(i) impact the theoretical choice of tube, assuming all other factors held constant?

 

[abbey road d enfer]

Although it is generally preferrable to have as much gain as possible, Rp/Gm/Mu is only a minor concern compared to noise and linearity.
In all cases a higher Gm gives a lower output Z, so the transformer is better fed.
In a plate-load case, higher gain can marginally improve the overall system noise performance, but also may induce headroom issues.
In a CF, gain is essentially not affected by Rp/Gm/Mu.

 

[mjk]

In my experience, this is a system design problem including the other circuit parameters; there is no one "best" relationship.

Assuming the "typical" circuit e.g. C-12 or M49; common cathode, transformer coupled, no feedback...

For a microphone, you will likely want to operate the tube at less than 1 mA idle to reduce the grid "leakage" current. The data sheet numbers for gm & Rp are usually given for a higher operating current than you would use in a microphone. Rp is much higher at low plate current. I construct a load line and read Rp from the slope of the tangent to the plate curve line closest to the idle point.

There is a design relationship with the Rp and transformer ratio which determines the output impedance of the mic. Output impedance = Rp / (turns ratio ^2) Typical Zout target is on the order of 200 ohms. Transformer ratios typically used in tube mics vary from 5:1 to 14:1, giving an impedance multiplier of somewhere between 25 and 200.

The amplification factor (much less at low current, also can be read off the actual load line) divided by the transformer ratio equals the gain from the capsule to the output. This determines microphone sensitivity. Typically this gain is between 0 and 6 dB. You can calculate the loaded capsule output from the capsule capacitance, capsule sensitivity, and capsule circuit loading. I like to shoot for about 20 mV/Pa sensitivity.

The gm of the tube is the ratio of amplification factor to Rp, and is a sort of "figure of merit" for the desirable combination of high gain and low Rp. gm is of course less at low plate current. Assuming you are starting with the capsule selection, you need enough gm to achieve the desired gain and Zout with a given transformer.

For example, the 6072 as used in the C12 runs at about 40K ohms plate resistance and uses a 14:1 transformer. The amplification factor in circuit is ~20, so there is about 3dB gain and about 200 ohms Zout. (numbers from memory...)

The AC701 as used in the M49 runs at about 10K ohms Rp in circuit and has a gain of about 15. The transformer ratio at 7:1 yields Zout of something like 200 ohms with a gain about 6 dB. (also from memory...)

The AC701 has somewhat higher gm than the 6072, which allows more in-circuit gain at the same Zout, given the same Zout target, but the end impact on mic performance is negligible in the context of the overall circuit design.

Michael

 

[bezen4uk]

Michael, It's a good point about real tube parameters at low plate currents.

re soapfoot:

I'd add some points from my experience:
*input capacitance of tube (including Miller capacitance) dramatically reduces gain. You should consider that real gain will be lower than you calculate from DC curves. Tubes with higher gain and higher Gm have high capacitance and are rarely used in tube mics. More typical are medium-mu triodes with moderate Gm (about 2-4mA/V in working point).

* Noise: since the capsule is high impendance source, general part of noise is produced by input current noise of tube. High Gm tubes have higher current noise. Another part of current noise depends from Up. You can see the majority of Neumann tube mike circuits have Up about 35-45V. Such circuits typically use medium-mu, moderate Gm tubes.
Also, there are a lot of tube construction factors, and sometimes "good looking on paper" tubes really are not suitable for microphone use

*microphonics: usually, pure triodes have lower microphinic effect. Pentodes have more complex construction and so more complex microphonic tones . Also, bigger tubes have higher microphonics (although, it makes that sound )

*Rp : I think, it's better to use circuit design with Rp lower than 10kOhm. High Rp in combination with inductance of transformer (rather, core losses in transformer) causes unpleasant distortion at low frequencies (in worst cases the transformer causes more distortion than the tube, even at low level signals).

Hope it will help you.
PS: There is no just good TUBE, there is good combination of tube, working point and transformer parameters.

 

[soapfoot]

this is a very interesting and helpful discussion.  I've learned a lot from all of your answers.  Thanks again.