UM57 & UM70 observations

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Frequency plots and the effect of loading the microphone for the stock UM57...
http://www.xaudia.com/omnip/Mics/Gefell/Graph3.gif

And with a 47 uF cathode bypass cap, and the transformer feedback removed.
http://www.xaudia.com/omnip/Mics/Gefell/Graph4.gif

And seeing as we're on a new page here's the fully modded CMV mic again
http://www.xaudia.com/omnip/Mics/Gefell/Graph2.gif


These are all pretty flat in the big scheme of things - just 1dB drop across the range in most cases.

Also, it looks to me as though the CMV mic with the 100K plate resistor is more sensitive to output loading, which would make sense.
 
Good stuff! At first I thought the graphs look wrong, but looking closer, I see the level resolution is quite high, so they look a bit "dramatized".
The decibel numbers next to the graphs are the make-up gain, I assume. So @ 10k the stock circuit is about 1 dB louder than @ 600 ohms, correct?

So far, the stock circuit looks best, at least on paper.

I wanted to do some measurements as well, but strangely my UM57 doesn't work anymore. I only put in a different output cap, I don't know what happened. For some reason I have no anode voltage anymore. Maybe it is the PSU. I'm getting 125V unloaded, but when I connect the mic, the voltage is gone. Oh well, bad karma.

Did you connect the transformer to ground, when you put in the cathode bypass cap? Theoretically you could leave it connected to the cathode and still get more gain with the bypass cap.
 
Rossi said:
The decibel numbers next to the graphs are the make-up gain, I assume. So @ 10k the stock circuit is about 1 dB louder than @ 600 ohms, correct?

That's it. It's not real make up gain from an amplifer though, just a mathematically addition to get it on the same bit of paper.

The average circuit gains for the mics (20Hz-2KHz)
Stock UM57 = 3.283 dB
Cap & no feedback = 6.949 dB
Cap, no feedback & 100K = 8.925 dB

So far, the stock circuit looks best, at least on paper.
Certainly, but my ears are suggesting that having a little less bottom end feels a bit cleaner. But I don't really want to get into that yet!

Did you connect the transformer to ground, when you put in the cathode bypass cap? Theoretically you could leave it connected to the cathode and still get more gain with the bypass cap.

I've tried it both ways but these measurements were with the tranny to ground, like in the schematic a few posts ago. But I'm not yet convinced it's the best way.

I've just removed the cathode bypass cap from the modded UM57 so it is stock but with the transformer to ground. Will post results shortly.
 
My UM57 is running again. Turns out *three* wires in the connector were broken. Although apparently they made connection, when I held the connector to measure the voltage. Weird. But those GDR connectors suck...


I just tried a 47u cap across the cathode resistor with the tranny wired stock. I think I like it better like this than without the cap. Subjectively, the bass is a little tastier. Without the cap, my UM57 sounds a little forward in the upper register. The cap makes it sound a little smoother.

Before that I put in a new output cap. I originally replaced the lytic with an ancient  .47u mylar (?) cap, which sounded smoother than the stock lytic (which had shrunken to 150 n). The new one is supposed to be a Panasonic polyprop, but I'm not quite sure, it could be mylar, because it is relatively small for a 1u film cap. With that cap the treble seemed to have slightly more edge. The cathode bypass cap seems to take off that edge.

But that's all just subjective impressions. I think I have to live with it for a while.

 
I should measure the old electro in my stock mic. I have two more which were removed from the other two mics.

So the mods certainly bring up the gain, but do we want that? As always this depends on the use of the mic. With more gain the transformer will saturated earlier, so if you're recording something loud. I'm trying to get some numbers on this but my preliminary results indicate that it is the bass response suffers first.

I've just removed the cathode bypass cap from the modded UM57 so it is stock but with the transformer to ground. Will post results shortly

Here goes... look at the change at the low end!

http://www.xaudia.com/omnip/Mics/Gefell/Graph5.gif

But again the lines are shifted so we can compare the response.


Average gains for the circuits with 2K load

1.18 db: Transformer primary connected to ground.
3.28 dB: Stock UM57. Transformer primary connected to cathode.
6.95 dB: Transformer to ground & cathode bypass cap added
8.92 dB: Fully modded - xformer to ground, bypass cap, 100K plate resistor & 470Meg grid resistors.
 
For what I plan to use it on, a slight reduction in headroom won't hurt much. Also, I'm not sure to what degree that will really be the case. My feeling is that the transformer should be able to take some level. It is not that small (much larger than say a Beyer tranny, at least) and there's not that much bass response, anyway. In the mid band, transformers usually can take quite some level without too much distortion.

Another possibility is to do as some other guy did: put in a relais that reacts to phantom power. That way you can remote control the bypass cap via the phantom power switch.


Do you plan to do any measurements regarding the size of the output cap?

 
Rossi said:
In the mid band, transformers usually can take quite some level without too much distortion.

True enough. At the moment I'm looking at the entire circuit, and can't directly separate effects of the transformer from that of the amplifier -

Another possibility is to do as some other guy did: put in a relais that reacts to phantom power. That way you can remote control the bypass cap via the phantom power switch.
Yes, I thought this was a really nice idea.

Do you plan to do any measurements regarding the size of the output cap?

I do, and the cathode cap too. But one thing at a time!
 
Rossi said:
Do you plan to do any measurements regarding the size of the output cap?

OK, here goes.... these plots are all with the same tranny and tube. Your T&T may vary....

here's the stock UM57, 2K load, varying C3.

UM57_Stock_C3.gif


To get these I simply replaced the output cap with a decade box (0.1uF x 10) and turned the dial. The lines shown are in 0.2uF steps. I have the odd numbers too, but the graph was getting cluttered.

The plots look nicer because I've entered the 21st century and downloaded fuzzmeasure. Seems quite good once I'd spent an hour or so convincing myself the results were valid. Before this I was using a generator, scope, multimeters and pencil. (Although I have to say the pencil never crashes or hangs up!). Anyway, with the computer I can get more data, faster and don't need to type the numbers into the spreadsheet. The software does struggle with plotting more than 4 traces at once - suddenly gets very slow.

The big thing with the stock circuit is that as C3 comes down, the resonant bump comes up and it may feel like a smaller cap gives you a bigger bottom. As Rossi notes above this can happen when the original electrolytic dries out. So going with a bigger C3 here might push down and flatten out that bump. (I measured the original 1uF electro in one of the UM57s and got a value of 1.1 uF, which is pretty good after all this time.)

Rewiring the transformer to ground gives this chart...

UM57_trafo.gif


The resonant bump is much less pronounced. Presumably Neumann-Gefell-VEB-RFT did the transformer feedback trick to get the best specs with the parts in this system, including the capsule. But with it wired to ground the optimum value for C3 might  be about 0.4 uF (green line) , assuming you don't mind a little bit of HPF!

Two more charts, firstly with transformer feedback AND 47uF cathode bypass cap:

UM57_47uF_C3.gif


and then with the xformer wired to ground and 47uF cathode bypass:

UM57_47uF__trafo_C3.gif


Again, you can 'tune' the bottom end, balancing the rolloff from the cathode bypass against the resonance from C3 - bigger is not always better.

This software also generates phase data against frequency. I'm still digesting that.
 
Wow, great stuff! Truly awesome! Looks like I should check out some other values, although with my lower inductance transformer, the curves would probably look different. It's interesting how small changes can make a big difference. So it's definitely worth trying different values and even measuring them individually.

BTW My stock lytic (shrunken to 150n) was rated 470n as in the schematic, not 1u as in your mic. Looking at your measurements, the stock 470n value creates a rather large LF boost.

If I reed your graphs correct, 600n and transformer to ground is the most linear configuration. I think I'd like a very gentle boost of 1 or 2 db as in the green curve with the transformer to cathode + cathode bypass cap. Which value is that - 400n?
 
When I first looked at this, I thought the NFB level was dependent mainly on the inductance on the primary and that was how they had aimed to control it. However, that doesn't make sense because it should increase as the frequency drops if we just consider the primary inductance alone.

So, looking again, is the NFB actually mainly controlled by C3? As you get down to the bottom-end, impedance rises at C3, injecting less potential into the cathode from R3. This might explain why in the stock circuit you managed to hit a stonking +4dB at 30Hz by lowering the cap value, by not relying on RLC resonance alone.

Tell me to be quiet if that sounds like nonsense!

Roddy
 
Rossi said:
If I reed your graphs correct, 600n and transformer to ground is the most linear configuration. I think I'd like a very gentle boost of 1 or 2 db as in the green curve with the transformer to cathode + cathode bypass cap. Which value is that - 400n?

That's right - in the last graph somewhere between 0.6 - 0.8 uF looks flattest and the green line is 0.4. Of course with your transformer it may be a little different, and maybe each mic needs optimising, or at least tweaking to taste.

I've seen schematics for CMV563 with 0.5uF,  and UM57 with 1 uF for the cap.  Perhaps the transformer changed spec at some point?  I think those Gefell guys knew their stuff. The LF rise could be to balance the M55k capsule, perhaps?

rodabod said:
looking again, is the NFB actually mainly controlled by C3?

It looks that way doesn't it, or at least some balancing act between the two? I'm still trying to reconcile this with the slopes from the last two graphs. I think they are about right for RLC, but haven't studied enough similar systems to be certain.

I think you could make a nice resonant filter for a synth from this design!

rodabod said:
Tell me to be quiet if that sounds like nonsense!

;D
 
It did! Thanks! I'd seen something similar on ebay. I wonder how that mod measures. I'd love to see some frequency response charts.

Anyway, permission to publish these charts on other websites or use for commercial purposes etc is NOT given, unless you write nicely and ask. I suppose I should stamp that on the charts.

I have the urge to DIY an anechoic chamber. Might be getting ahead of myself here....
 
He uses Oliver Archut's circuit, so the only unknown parameter is the AMI transformer. I can't say I like that kind of business. I mean, yes, he's doing a service, and he's open about his sources. But I don't like people offering modifications without doing their own research. How can you offer something when you don't really know what you are doing? Soldering skills are not enough, imo.
 
Rossi said:
I don't like people offering modifications without doing their own research. How can you offer something when you don't really know what you are doing? Soldering skills are not enough, imo.

Sadly, this is very common, to varying degrees.
 
Great stuff so far here thanks for starting this Zebra! So has anyone done this project lately either a breadboarded P2P version or a PCB? I was looking at this to try as the parts are minimal for the circuit and this thread has re-sparked some talk about the mic. Also anyone have any comments of the "sound of this mic" in general terms without getting into a controversy!  ;D

 
mpc3000limited said:
So has anyone done this project lately either a breadboarded P2P version or a PCB?

Yes. You don't need a PCB - the early Neumann / Gefell ones don't use one. It's easy enough to build this circuit or something like it point to point. But the problem of course is getting the right transformer and the right capsule.
 
Hi Zebra
I´m reading and learning from this excellent thread quite a long time now. In the end, I decided to build a CMV 563 on my own. I studied the schematic and high resolution pictures and drew a component layout. Besides two resistors and the wiring everything is ready to go. Can anyone doublecheck my rough drawings?  Sorry for spamming your "observations". Maybe someone should open up a new "CMV building thread"

CMV_Layout.JPG



Thanks a lot.
regards
Bernd
 
Why would you want to build a CMV?

Most microphones like that are simple triode single stage transformer out designs no big deal.

There are more interesting designs to copy, or build something better.

 
Hi GUS,
honestly, it´s the first time that play around with a mic circuit. I have a CDK M7 clone capsule, a tiny Beyer mic transformer, EC92 and PC92 tubes and other parts at hand and just wanted to experiment a little bit. I thought the CMV might be a good start. I don´t expect too much concerning the sound quality of the microphone, but I simply had to build a microphone. For learning, nothing more.
regards
Bernd
 
Back
Top