M50c schematic

[riggler]

Anybody have a schematic for a Neumann M50c? Got my 5703 equipped M50b version design done, but through research found that there's a quieter filament section in the M50c...  ???

 

[mad.ax]

AFAIK, the difference between B & C versions is that the former is fixed bias while the later is self bias.
WRT bias, I guess the M49C and M50C are the same, so here is the M49C schematic if it can help...

Axel

 

[riggler]

Thanks mad ax! Decisions decisions. I think that the fixed bias will sound juicier, at least from what I've read. More to learn!

 

[abbey road d enfer]

With fixed-bias, both the HT and heater voltage must be super clean, because the bias voltage is directly derived from the heater voltage. With cathode-bias, only the HT must be super clean. As you know, making the heater voltage super clean requires huge caps and serious attention to wiring/grounding.
OTOH, with the little tube running class A, the difference in tone between fixed and cathode bias is microscopic.
IMO, the evolution from fixed to cathode-bias makes absolute sense.

 

[mad.ax]

FWIW, according to K.H., cathode-bias drawbacks are less sensible with triodes like AC701, than they are with penthodes, as VF14 for instance.
"C" versions of the Neumann circuit are usually considered different but not less valuable than previous versions.
It should be noted though that the value of the cathode bypass cap should be carefully chosen to keep the phase shift issues outside of the audio range.
Given the tolerance of lytics caps, perhaps some selection and listening sessions could lead to "better" results...

Axel

 

[abbey road d enfer]

Can you elaborate on the "phase-shift issues"?

 

[burdij]

From experience with building up a couple versions of this circuit, I find that the bias scheme that includes connecting the cathode to the filament raised above ground by a series resistance that is not bypassed is the circuit that most closely produces the sound that I associate with a U47, particularly on vocals. It is true that this is more difficult to implement due to the quality of filament supply necessary but the effect is a noticable one and certainly worth the effort.

It doesn't matter if the tube is a native triode or a pentode connected as a triode although the pentode seems to need to be of the class that has a seperate connection to the suppressor grid rather than a suppressor connected to the cathode internally.

I am also using this bias method in my M50 recreations.

 

[riggler]

Burdij,
might I get a look at your Pcb layout? I'm putting all high impedance pieces on top but am unsure which connections really need short leads/traces... I got some acrylic spheres but don't have a meter that can read into the gigaohm range, so I hope I have enough isolation. Would a glass mounting stud between the sphere and pcb plastic "cap top" help here?

Back to the topic... I think I might stick with fixed bias, how often should the bias be checked and adjusted?every 50 hours? 100 hours?

 

[mad.ax]

Can you elaborate on the "phase-shift issues"?

I was mostly quoting Klaus Heyne, the relevant thread is here:
http://recforums.prosoundweb.com/index.php?t=rview&goto=39486&th=3389

I'd like to learn more on that matter, whether it infirm or confirm the above...

For my part I believe that the value of the bypass cap should be determined by the lowest frequency one want the circuit to pass, but other factors that may have an influence on linearity should be taken into consideration. The interaction of output coupling cap and transformer for instance.
Also it is my experience that sometimes, say 3µF of difference on a 50µF bypass cap can make an audible difference (same brand, same quality)

Axel

 

[abbey road d enfer]

Please note that Klaus Heyne himself says: "The previous poster mentioned the low frequency phase shift associated with cathode bypass caps. That is of course no excuse not to apply this type of circuit, where applicable. Carefully choosing the right value and quality of this cap will yield phase shift-free results of high audio quality."

The phase-shift is the result of accumulation of high-pass filters in the chain:
1st filter: capsule capacitance versus input resistance (straight 1st order filter)
2nd filter: decoupling cap versus cathode impedance (1st order shelving filter)
3rd filter: coupling cap versus transformer (2nd order)
The latter is the one that may introduce the most phase-shift; the 2nd one, because it's a shelf, is the one that may introduce the less phase-shift.
Anyway, being tuned by design to a very low frequency, these filters introduce negligible phase shift in the audio range.
What's more, many researchers have demonstrated that the audibility of phase-shift at these frequencies is delusional.
And to top it all, what about the phase-shift introduced by a bona fide 80Hz high-pass (pop filter)?
Klaus Heyne's forum is full of threads explaining that negative feedback is devilish, cathode-bias is rubbish, only paper-oil caps are worthy, a mic without a transformer cannot be any good, and so on. Not the best way to acquire scientific knowledge, IMO.

 

[riggler]

Good points here! My most beloved Living Stereo recordings were recorded with fixed-bias mics. I've got no way to know what difference if any there would be if they were self-bias. We've got some people here with real world experience saying they hear a definite difference. We've got Klaus saying in a triode-equipped mic he hears no difference.

I'm going to build both and see which I like. I just don't want to kill a tube or capsule in the process, so on to learn how to implement...

You guys are all great man, I freakin love it here.

 

[zapnspark]

What's more, many researchers have demonstrated that the audibility of phase-shift at these frequencies is delusional.
And to top it all, what about the phase-shift introduced by a bona fide 80Hz high-pass (pop filter)?
Klaus Heyne's forum is full of threads explaining that negative feedback is devilish, cathode-bias is rubbish, only paper-oil
caps are worthy, a mic without a transformer cannot be any good, and so on. Not the best way to acquire scientific knowledge, IMO.

I'm glad you raised these points.
I read and enjoy Klaus Heyne's microphone forum.
Here comes the however
That forum has some of the usual audio and microphone folklore that is unsupported by research.

Cheers.

ZAP

 

[Rossi]

When people talk about phase shift and low cut (and time delay in the case of digital microphones or digital monitoring systems) I find that they neglect the fact that the situation matters a lot. When you stand in front of the microphone and listen to yourself over headphones, you can easily hear phase-shift in the form of phase difference to the original, direct sound. I can hear a 30 Hz hi pass even though my voice (in fact nobody's voice) extends that low. I don't think I could hear the same hi pass later on listening to the recording.

To me a 30 or 40 Hz Hi pass doesn't sound bad over headphones. I do dislike a low cut at, say, 80 Hz, though. When you stand in front of the mic, a 80 Hz hi pass makes your voice sound kind of indirect. However, the same hi pass may improve the sound to the guy in the control room.

I find that a 30 or 40 Hz low cut is a good thing for a mic in almost any situation. There's really nothing interesting going on below that range, and if there is, you use a different mic (e.g. true omnis for organ recordings). I hardly ever use a higher low cut when recording but cut excess bass in the DAW. I'd really like mics to have a 30 or 40 Hz low cut switch. The U47 cuts frequencies below 40 Hz and I see nobody complaining about phase shift.

 

[rodabod]

decoupling cap versus cathode impedance (1st order shelving filter)

Interesting. I'd guessed it would act like a slightly more complex (at least to calculate) 1st order, and not shelving in any way. I'll have to think about that. Ah, I suppose you don't get the usal 6dB/8ve slope due to the parallel resistance across the cap. Either way, keep that cap large enough, and decent quality. One thing I did consider is that any voltage formed across this cap (ie. distortion which is uncorrelated with the common signal) would effectively be amplified as a grid-cathode voltage difference. I'd need to prove that though.

Regarding that forum, there's  snobbery, etc. , but the amount of misinformation possibly isn't all that bad. I've recently been browsing Gearslutz (don't ask why... I was bored), and f**k me, there's some ******** right there. I almost view it as dangerous... There are people there recommending wasting money on something that they've never tried, that was recommended to them in the first place by the bloody company owner! WTF!

 

[zapnspark]

Just an FYI.
Matt Mcglynn's website (recordinghacks.com) has some great information on
the Neumann M50 (eyeball) microphone series.
Here is the link:

http://tinyurl.com/ycd6jex

Cheers.

ZAP

 

[abbey road d enfer]

When people talk about phase shift and low cut (and time delay in the case of digital microphones or digital monitoring systems) I find that they neglect the fact that the situation matters a lot. When you stand in front of the microphone and listen to yourself over headphones, you can easily hear phase-shift in the form of phase difference to the original, direct sound. I can hear a 30 Hz hi pass even though my voice (in fact nobody's voice) extends that low. I don't think I could hear the same hi pass later on listening to the recording.

Please note that, in that experiment, you are not anymore in a minimum-phase configuration, because the sound you hear is constituted of 3 parallel paths: electrical, acoustic and solidian. All three have a different time-delay: electrical is the shortest, then solidian (speed of sound in flesh and bones is highly variable) and acoustic is generally the longest (not only for the aerial distance between mouth and ear, but because of the time-delay introduced by the headphone insulation).
The importance of this multiple-path can be easily checked by reversing the polarity (phase) of the electrical path. This leads to a very noticeable change in the perceived sound.
That's why I consider assessing the quality of a microphone that way is absolutely ridiculous. However, that's the way it's done most of the time, particularly in MI shops.
Please note, I'm not criticizing your experience; on the contrary, I agree with you that, due to the complex relationship between the various paths, the phase-shift becomes audible and a factor that needs to be assessed.
One more time, this is a confirmation that the consequences of phase-shift, although being generally inaudible in a single-path (minimum-phase) linear system, become audible when multiple-paths are experienced, or excessive non-symetrical distortion.

To me a 30 or 40 Hz Hi pass doesn't sound bad over headphones. I do dislike a low cut at, say, 80 Hz, though. When you stand in front of the mic, a 80 Hz hi pass makes your voice sound kind of indirect. However, the same hi pass may improve the sound to the guy in the control room.

Since we are mainly concerned with the recording, this is the ultimate decision factor; but we cannot neglect the "comfort" factor for the talent. Talent's comfort is probably the paramount factor in a good take, over any mic choice or high-pass filter selection.

I find that a 30 or 40 Hz low cut is a good thing for a mic in almost any situation. There's really nothing interesting going on below that range, and if there is, you use a different mic (e.g. true omnis for organ recordings). I hardly ever use a higher low cut when recording but cut excess bass in the DAW. I'd really like mics to have a 30 or 40 Hz low cut switch. The U47 cuts frequencies below 40 Hz and I see nobody complaining about phase shift.

The problem is that very often , the choice is limited (to say the least); in fact, according to different situations, there should be different high-pass options. Sometimes you want a 1st-order high-pass for compensation of proximity effect, some other time you want a steep 30-40 Hz 4th-order for getting rid of foot noise, and maybe one day you want this 2nd-order 80Hz filter to get rid ot ventilation noise; who knows?
AKG seem to have brought a beginning of answer with the new 414's: 3-position, 12 dB/octave at 40 Hz or 80 Hz and 6 dB/octave at 160 Hz

 

[abbey road d enfer]

decoupling cap versus cathode impedance (1st order shelving filter)

Interesting. I'd guessed it would act like a slightly more complex (at least to calculate) 1st order, and not shelving in any way. I'll have to think about that.

Just think of the gain varying from fully bypassed cathode (s.Rp) to open cathode (Rp/Rk); at both extremes the phase is pure 180°. In the middle, the cap plays with the Rk//1/s to twist the phase up to 45° lead. (Veeeery simplified approximation, don't hang me for that)

One thing I did consider is that any voltage formed across this cap (ie. distortion which is uncorrelated with the common signal) would effectively be amplified as a grid-cathode voltage difference. I'd need to prove that though.

You are absolutely right. Anytime you let an electrolytic cap develop a noticeable voltage in the audio range, its non-linearities become apparent in the signal its stage carries. Electrolytic caps in the audio path should see unsignificant AC voltage across them. When I mentioned the three sources of high-pass filtering in a typical valve capsule amp, the 1st and 3rd one are not significantly non-linear. The 2nd one is. The fact that electrolytic caps do not age well has made them identified as a source of non-linearity (call it distortion, warmth, grit, character...), but good design practice and thorough maintenance should do them justice.

Regarding that forum, there's  snobbery, etc. , but the amount of misinformation possibly isn't all that bad. I've recently been browsing Gearslutz (don't ask why... I was bored), and f**k me, there's some ******** right there. I almost view it as dangerous... There are people there recommending wasting money on something that they've never tried, that was recommended to them in the first place by the bloody company owner! WTF!

I know, I know... Thruth is K.H.'s forum is not too bad (although I don't agree with him about his despise for NFB), and GS is an almost absolute dumpster. For this reason, K.H.'s forum is viewed by some as the paragon of undisputed truth. In fact, a number of false assumptions permeate and the least technically minded members get polluted by these. I do sometimes have a peek at GS, just to see what the latest BS is going to hit the fan.

 

[zebra50]

I haven't played with this circuit, but looking at the M49c schematic posted near the top, I'm wondering about what exactly is put into the 'calibration input', and to what end? To me it looks like it is there to inject some signal through C2 & C3 onto the backplate and adjust / select C4 to balance the two sides of the capsule? And if so you wouldn't need this with an M50 capsule.

Also, C2/C3/R3 is presumably to introduce a little bit of feedback to the backplate?

Finally for now, the Blue 'omni' mouse has a 'proper' spherical omni capsule rather than dual backplates, fwiw. I'll try to get some pictures.

 

[Rossi]

Please note that, in that experiment, you are not anymore in a minimum-phase configuration, because the sound you hear is constituted of 3 parallel paths: electrical, acoustic and solidian. All three have a different time-delay: electrical is the shortest, then solidian (speed of sound in flesh and bones is highly variable) and acoustic is generally the longest (not only for the aerial distance between mouth and ear, but because of the time-delay introduced by the headphone insulation).
The importance of this multiple-path can be easily checked by reversing the polarity (phase) of the electrical path. This leads to a very noticeable change in the perceived sound.
That's why I consider assessing the quality of a microphone that way is absolutely ridiculous. However, that's the way it's done most of the time, particularly in MI shops.

The multiple path situation was exactly my point, and yes, people aren't always aware of the fact that the phase correllation of the electrical path is a crucial factor when you stand in front of the mic. Which is why I think any preamp should have a polarity switch, even if it is just one channel.

I do disagree with your latter statement. Assessing the quality of a micophone over headphones is not ridiculous at all. Evaluating the comfort and subjective feel for the performer is absolutely important. The performance is the most important part in the whole equation. You can always make up for technical matters, but you can't really improve a bad performance.

Also - I know this is subjective, but having reviewed and owning a large number of mics I speak from experience - the sound quality of a mic is almost directly proportional to the effect the polarity switch has on your listening experience in front of the mic. At least in my setup that's an almost unmistakable indicator. My U47 for instance sounds fantastic and very "true" when the electrical path is (more or less) in phase with the acoustic path(s). When you flip the phase switch, it sounds weird, far away, and drops in volume. I have several other mics that come close, and they're all considered hi quality mics. With lower quality mics, you often have difficulties to tell which position of the polarity switch sounds more direct. There are of course other factors as well, but I think the comfort factor and the subjecitively natural sound for the performer contribute to the popularity of those mics. In the case of the U47, I think the 40 Hz low cut also helps a lot, when it comes to vocal applications. It also looks great.  ;D

 

[burdij]

Burdij,
might I get a look at your Pcb layout? I'm putting all high impedance pieces on top but am unsure which connections really need short leads/traces... I got some acrylic spheres but don't have a meter that can read into the gigaohm range, so I hope I have enough isolation. Would a glass mounting stud between the sphere and pcb plastic "cap top" help here?

Back to the topic... I think I might stick with fixed bias, how often should the bias be checked and adjusted?every 50 hours? 100 hours?

The PCB for the test mike is conventional (I don't have a picture handy, either). No part of the grid side or capsule circuit are on the PCB. High impedance junctions, if necessary, are supported on Delrin standoffs with terminals embedded in them. The capsule side of the high resistance bias network is supported on standoffs. The tube socket is a good quality ceramic. Before installing the socket in the board, I drilled out the pad for the grid pin to a large diameter so that the pin itself is surrounded only by air or ceramic. Acrylic has a high resistance value (comparable to glass) so if it is fairly clean, it will supply a high isolation. Using the resistance values from the original Neumann designs (i.e. 100-200 Meg Ohm range), you don't need to use heroic isolation measures. Don't, and I repeat, don't use Gig Ohm resistors in these circuits thinking it will make them sound "better". They will have an unfortunate effect on the sound quality. When installing a high value resistor, remove it from the factory packing with a pair of tweezers or wear clear rubber gloves (non-talc latex). If the resistor has been handled, clean it with low H2O isopropyl alcohol and thoroughly dry it. From that point on, handle it with tweezers (or gloves). After soldering, clean all junctions and the standoff areas with iso and dry (a can of dry N2 comes in handy here).

In this thread, there are some pictures of an earlier version of the M50 mike capsule. I used teflon covered or teflon sleeved wire to connect to the capsule.

http://www.groupdiy.com/index.php?topic=29692.40