C12, ELA M250e, 1k8 or 2k7 for cathode, higher or lower cathode bias??

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

musix2000

Member
Joined
Feb 16, 2017
Messages
17
Location
Hamburg, Germany
I`m building an ELA M250e clone with a CT12, JAN GE 6072a mic. grade and a 13114 transformer by cinemag. As in the original schematics I use a 1k8Ohms (1k82) cathode  resistor, instead of the no more produced 8M resistor I use one with 7M96(7)Ohms  For tests and some experiments I have the capsule of a Rhode K2 in it. Because Telefunken have a 2k7Ohms (2k67) resistor in their C12 and ELA M251e I`ve also tried this: with 1k82Ohms 1.065V and with 2k67Ohms 1.37V are coming from the cathode (B+ is 120,6V). It seems: higher resistor values > higher cathode bias voltage. I was told that the cathode bias should be 1.1V, otherwise, with more voltage, the tube could be damaged.  But I think the guys at Telefunken know what they do. They also have changed from the 8M resistor to a 10M (what I still haven`t tried, maybe this brings the cathode bias down to 1.1V??).  Whether in their C12  resistors were changed or not I didn`t find out, instead of fixed bias they also use a 2k67Ohms resistor.
I still haven`t tried the microphone under studio conditions, I`ve just finished it with the K2 capsule and it`s in my workshop for the next days.
Is it ok to have a higher or lower cathode bias? What is the influence to the sound?
Does anybody know about this?
 

kingkorg

Well-known member
GDIY Supporter
Joined
Apr 15, 2017
Messages
2,030
Location
Norway
I think you are being to picky about values. 8meg vs 10meg wont play huge role here.  Some use 1 Gig here. There are a lot of threads about what those components do and how to pick them.

You shouldn't look for voltages but for the sound. Those 1k8 and 2k7 are biasing resistors, look for adjusting bias topics. IMHO those two values won't make a world of difference if any. And those voltages aren't hurting the tube. Try both and use one you prefer. Ideally you should chose the value that makes least amount of THD and that could be neither of those.
 

RuudNL

Well-known member
Joined
Apr 26, 2009
Messages
2,801
Location
Haule / The Netherlands
My experience is that the value of the cathode resistor is not very critical.
Usually I aim for a value of about 1/2 of the supply voltage of the tube stage on the anode.
Tubes aren't exactly 1% tolerance devices!
 

Matador

Well-known member
Joined
Feb 25, 2011
Messages
2,544
Location
Bay Area, California
The cathode resistance sets the idling point of the tube.  Given most small signal triodes in these microphones have only about 2V of usable grid swing when run at 120V supply voltage, then a 1V bias point is a reasonable point for most tubes.  So in short, the stock value of 1k8 works well across many tubes.

If you want to be super pedantic, you would do a signal swing analysis on a scope, and bias the tube for minimum distortion (much like you would with a JFET).  But tubes aren't exactly meant to be perfectly clean so I wouldn't worry about it.

You won't damage the tube by adjusting cathode bias in any case.
 

musix2000

Member
Joined
Feb 16, 2017
Messages
17
Location
Hamburg, Germany
Thank you all for your answers, they make the thing much clearer for me!
I have a scope, would an adjustable resistor to find the lowest THD be a choice? Does lowest THD also mean best working point of the tube?
 

kingkorg

Well-known member
GDIY Supporter
Joined
Apr 15, 2017
Messages
2,030
Location
Norway
Last time I made a tube mic and adjusted it to lowest possible THD it cancelled 100% in null test with same capsule in the same body but with FET circuit. In other words the mic sounded exactly the same as FET mic when used in normal conditions.

Like Matador mentioned I don't think  tube mics are about lowest possible THD. 

 

musix2000

Member
Joined
Feb 16, 2017
Messages
17
Location
Hamburg, Germany
Hi Kingkorg....
This would mean that the cathode resistor is the component that makes the tube sounding like a tube or not, is this right? In this case an adjustable resistor would be a good choice to create the tube sound to it`s best. Values could be changed as long as the optimal sound is reached.
 

RuudNL

Well-known member
Joined
Apr 26, 2009
Messages
2,801
Location
Haule / The Netherlands
When I first started building tube microphones, I thought that there was something seriously wrong with the design!
My solid state microphones usually had distortion figures of 0.02% at high output levels, but for a tube microphone 0.8% seemed 'normal'. I analyzed the distortion and came to the conclusion that it was mainly second harmonic, probably the kind of distortion that people like about a tube microphone. Even top class tube microphones produce this kind of second harmonic distortion!

By the way: adjusting the cathode resistor for minimal distortion on an oscilloscope isn't very exact.
Before you start seeing distortion (on a sine wave) on an oscilloscope, the THD is already in the order of 3%.
Personally I use a distortion analyzer to find the bias point for minimal distortion.
 

Matador

Well-known member
Joined
Feb 25, 2011
Messages
2,544
Location
Bay Area, California
At low cathode resistance (meaning that as the cathode voltage approaches ground), ground-referenced input signals will tend to distort as the grid is driven more positive and the tube becomes progressively non-linear.  The same is true for operating a higer cathode resistances (meaning the cathode voltage is increasing), as the tube current will pinch-off as the plate voltage swings upwards and intersects with the 120V power rail.

So the first case is that the porridge is too hot, and the second is that the porridge is too cold, so we want to be right in the middle. :)   

c12_6072_bias.jpg


So here is a typical tube with a typical load line as you would see in a microphone application.  The cathode bias sets where on the red line the tube idles, so if you are far to the left of the line (low bias voltage) you have positive grid operation, and far right on the line is where the curves become 'bendy' and the tube current cuts off.  Somewhere right in the middle is where the -1V and -2V grid lines intersect, which tells you that you need to run a cathode voltage between +1V and +2V.

This chart is for a 12AT7, but you can calculate the exact resistance from any tube type in a similar fashion.  If you want 1V cathode voltage, follow the -1.0V grid voltage line down until it intersects with the red line, then follow it over to the left and see what plate current this happens.  On this chart, it's about 0.7mA, so 1V at 0.7mA means 1/0.7 = 1.428k ohm.
 

granger.frederic

Well-known member
Joined
May 22, 2010
Messages
458
maybe it's punctilious, but why not choosing the tube (and its biasing voltage) depending to the max output voltage of the transducer (just before the capsule starts to really distort, maybe around 1%...), given by the manufacturer (if he measured it...) at its optimal voltage polarisation (usually around 60v)?
So , we could have the lowest distortion and the optimal bias point at the same time...
 

Matador

Well-known member
Joined
Feb 25, 2011
Messages
2,544
Location
Bay Area, California
That's pretty much exactly what is done.

If we assume a typical 20 mV/Pa sensitivity rating (@94 dbSPL) for an LDC, then we need about 30dB of gain to bring this up to line level at that average sound level.  94dB is fairly loud already, meaning that we can get to standard line level under stock conditions, and the 22mV input swing is well within the operating point of a tube under most conditions.  Also remember that we have an output transformer that reduces the gain by 20dB, so the net gain of the system (as seen by the output jack) looks like 10dB, or an average signal level of 100mV under pure stock conditions.

If we switch to a super loud source (say 130 dBSPL), then the capsule might put out near line level by itself.  If 1V RMS, then the tube biased at -1V will still be 'fine', the plate will swing 30V peak-to-peak, and the output signal will be 3-4V at the output jack.
 

Shattersignal

Well-known member
Joined
Jul 14, 2012
Messages
114
Location
Left Coast, USA
Matador, have you ever experimented with different plate R values and bias points to deliver more gain from the tube stage?  The stock mic is a little quiet for my liking, although certainly workable.  I'd love to get an extra 10dB out of it, if at all possible, to bring it more in line with the output I get with my Manly Ref C.  I'm curious if you or others have any experience to share with gain modifications to the stock C12 circuit.

Matador said:
That's pretty much exactly what is done.

If we assume a typical 20 mV/Pa sensitivity rating (@94 dbSPL) for an LDC, then we need about 30dB of gain to bring this up to line level at that average sound level.  94dB is fairly loud already, meaning that we can get to standard line level under stock conditions, and the 22mV input swing is well within the operating point of a tube under most conditions.  Also remember that we have an output transformer that reduces the gain by 20dB, so the net gain of the system (as seen by the output jack) looks like 10dB, or an average signal level of 100mV under pure stock conditions.

If we switch to a super loud source (say 130 dBSPL), then the capsule might put out near line level by itself.  If 1V RMS, then the tube biased at -1V will still be 'fine', the plate will swing 30V peak-to-peak, and the output signal will be 3-4V at the output jack.
 

Matador

Well-known member
Joined
Feb 25, 2011
Messages
2,544
Location
Bay Area, California
Yes, there are several ways to increase gain:. You need to flatten the load line, which means higher supply voltage and higher currents.  Circuit changes are needed.  This also means higher plate resistor values, which means higher output impedance.  You can convert to a CCDA topology and use the 2nd half of the tube as a cathode follower, which means you can use a much lower ratio transformer.

To be honest I haven't found it necessary, as you can get 50dB or more from a mike pre these days.

What kind of sources are you recording?
 

musix2000

Member
Joined
Feb 16, 2017
Messages
17
Location
Hamburg, Germany
Hi Shattersignal ...
With a 1k8 resistor I measured 1.06V at the kathode and 61,7V at the plate,  with 2k7 1.35V at the kathode and 70,5V at the plate (ELA M250e).  You can use at the lowest  a 1k5 resistor and the sound is more warm with less output, higher resistor values give more dynamic, the sound becomes louder,  brighter and loses warmth ....
 

Shattersignal

Well-known member
Joined
Jul 14, 2012
Messages
114
Location
Left Coast, USA
Thanks a ton.  Yeah, I tried bumping up the plate R value a couple times, and I seem to like it at around 160k (up from 100k) on the otherwise unmodified PSU.  I mainly use the mic for vocals, but I track almost exclusively on awesome tube mic pres that offer about 40-45dB of gain on their own.  The output of the Manly is basically perfect for my ideal gain staging, in most cases.  The C12 is pretty damn great sounding (much nicer than the Manly!), but the rest of the recording chain would like to see a few extra dB off the mic. 

Matador, I assume the "bias" pot inside the PSU is the variable cathode R?  I dialed the bias V up to about 1.8 as well, but I can't quite tell if it is making much difference in sound--although it could be dropping the headroom, I suppose, although I couldn't hear it.

Matador said:
Yes, there are several ways to increase gain:. You need to flatten the load line, which means higher supply voltage and higher currents.  Circuit changes are needed.  This also means higher plate resistor values, which means higher output impedance.  You can convert to a CCDA topology and use the 2nd half of the tube as a cathode follower, which means you can use a much lower ratio transformer.

To be honest I haven't found it necessary, as you can get 50dB or more from a mike pre these days.

What kind of sources are you recording?
 

Shattersignal

Well-known member
Joined
Jul 14, 2012
Messages
114
Location
Left Coast, USA
Also, FYI, I just spent all day yesterday tracking vocals on the C12 in that modified state and I ADORE what the mic was bringing to the sessions.  It definitely seems to throw a few extra dB, and the sound is a bit more "modern" for lack of a better description.  I really don't need or want a really gooey sounding mic, since I mostly work on densely layered electronic pop.  The sound of the mic with the Campbell capsule is very very nice with the boosted plate and cathode R values, it would seem.  (It didn't seem overly loaded down, despite the increased output impedance of the tube stage.) 

I think I'll eventually experiment with really reworking the load line and implementing a higher B+.  Of course, that would require some pretty big changes, as the capsule polarizing V is also taken off the B+, which is currently dialed as low as it goes on the PSU, at 120v.  In fact, Tim Campbell recommends the capsule run even lower, at around (EDITED) 55v, for maximum sonic detail, which would mean a supply voltage of 110v. 

Matador, would there be a relatively simple way to drop the capsule V, separate from the plate V?  I would love to lower the operating V so the  capsule can run at (EDITED) 55v while not changing the tube supply V (or possibly while even raising the tube supply V).  What would be the location in the circuit for the addition of such a dropping resistor?  If such a thing were possible, I might just put the 100k plate R back in, and crank up the B+, while trying to compensate the capsule V to hang out at around (EDITED) 55v. 

Thanks a ton!     
 

Matador

Well-known member
Joined
Feb 25, 2011
Messages
2,544
Location
Bay Area, California
Shattersignal said:
Also, FYI, I just spent all day yesterday tracking vocals on the C12 in that modified state and I ADORE what the mic was bringing to the sessions.  It definitely seems to throw a few extra dB, and the sound is a bit more "modern" for lack of a better description.  I really don't need or want a really gooey sounding mic, since I mostly work on densely layered electronic pop.  The sound of the mic with the Campbell capsule is very very nice with the boosted plate and cathode R values, it would seem.  (It didn't seem overly loaded down, despite the increased output impedance of the tube stage.) 

I think I'll eventually experiment with really reworking the load line and implementing a higher B+.  Of course, that would require some pretty big changes, as the capsule polarizing V is also taken off the B+, which is currently dialed as low as it goes on the PSU, at 120v.  In fact, Tim Campbell recommends the capsule run even lower, at around 100v, for maximum sonic detail. 

Matador, would there be a relatively simple way to drop the capsule V, separate from the plate?  I would love to lower the capsule V to 100v while not changing the tube supply V (or possibly while raising the tube supply V).  What would be the location in the circuit for the addition of such a dropping resistor?  If such a thing were possible, I might just put the 100k plate R back in, and crank up the B+, while trying to compensate the capsule V to hang out at around 100v. 

Thanks a ton!   
I would first try switching to a 12AT7 tube, as you can get a bit more gain (it has a higher gm) from the circuit with only relatively few/minor changes.

Otherwise, you need to change the R12/R13 ratio, and add a resistance between the polarization pins P2 and P3 where the polarization switch attaches to the screw terminal(s) in the PSU.  This would create a voltage divider that would reduce the higher B+ back down to 120V max.
 

Latest posts

Top