UM57 & UM70 observations

Hi!

I've spent my spare hours over the past few days tinkering with a couple of Neumann/Gefell/VEB UM57 mics, along with a CMV563/551 and an adapter (courtesy of Thiersch) to connect to Gefell UM70 and M71 capsules. I've been trying out some of the published modifications on the web - more of that shortly - but I wanted to note down a few of the differences and similarities whilst they're fresh in my head. I guess there is nothing earth shattering here and these mics are documented elsewhere. But Le bon Dieu est dans le detail*, und so weiter.**

Firstly, the two UM57s...




Both mics appeared to be extremely original when I first received them. Mic on the left is VEB badged and serial number in the 4900s, on the right is Neumann badged, and in the 4090s. The VEB has a circuit board and a different layout from the point-to-point wired earlier Neumann. On the P2P layout, the output cap is on the reverse side of the board. The orange poly cap is a replacement (by me) for a faulty electro. The circuit board on the VEB mic is marked both 563 and 57, suggesting it was used for both.

The VEB came with an RFT tube that was marked as 'M' grade, presumably tested and selected in the factory. I haven't seen this extra red stamp before. This seems to be as quiet as the NOS Tele tube I fitted to replace the standard RFT in the 'Neumann'.



In fact the mics sound extremely similar in this configuration, despite the differences in layout and components. To my ears the difference between the old (but good) electro cap and the new poly cap is not obvious.

Getting late here and whiskey calls. I'm going to post and save this - more about testing some of the mods and capsules in the morning...

z50


* Gustave Flaubert, allegedly.
** move along, nothing to see here, etc

So, the next up are the capsules. The head assembly of the UM57 can be swapped by removing three screws, and this in turn can be dissembled by removing three more screws, one of which is behind a neumann seal. If it's still there, think before you break it because collectors still look for this - it means at least the capsule will be there.

I had a UM70 and an M71 capsule too for comparison, along with a CMV563 mic body and adapter - the mutant mic is in the background of the photos above. What I hadn't done before was dissected the UM70. This is slightly harder as two plastic plugs need to be pulled out or drilled out to remove the patter selector ring. Once this is done everything is straight forward. Three screws release the switch assembly, and the capsule head can be lifted off carefully. Make sure you have a clean area because springs can pop out, and you don't want to lose them.




There's something here that I haven't seen flagged on the web before  (I try to do my homework). Once the switch assembly is removed, the capsule head connectors are the same as the UM57, and the heads are directly interchangeable.

This explains a couple of odd looking UM57s that I've seen for sale with sloping head grills. There are internet rumours that Gefell replaced some heads this way, and that may be the case, but doesn't have to be - anyone with a screwdriver could do this - I did this in the photo below. (And you can swap the capsules and grills over too, if you want to do that and hear the difference between the slope and the round grill.)



[Circuit mods coming soonish...]

Hi,


  all good stuff so far! I am keen to hear your opinions.


  Go Zeb!


      Kindest regards



        ANdyP

Z-50 --

This is great, keep 'em coming.  I'm in the process of building tube electronics into two MV-692 empty shells for use with UM-70 capsules and this most interesting.

Crazy Joe

Nice contribution!  Keep it coming!

nice - i like them eastern mics quite a lot (esp. mechanically)
i think i can guess the mods you are planning to do  8)
could you measure the inductance of the transformers?

-max

Thanks guys. More stuff later when I've done a day at the studio.

ioaudio said:

nice - i like them eastern mics quite a lot (esp. mechanically)
i think i can guess the mods you are planning to do  8)
could you measure the inductance of the transformers?

Click to expand...

Obviously I like this stuff a lot too! I've three or four working transformers - I hadn't thought to measure them but will do so. I see a lot of stuff written about the quality of this transformer, or lack thereof, but I'm trying to see it for myself because I don't think these are too bad.

No secret about the mods - they're things published over at PSW mic lab and in another thread here, adding a bypass, moving a transformer wire, changing a few resistors. But I want to record with them some more before I start saying that one circuit is 'better' than another. You can certainly get a bit more gain out but I'm wary of the 'louder=better' effect and want to look at the S/N ratios etc.

Joe Tritschler said:

I'm in the process of building tube electronics into two MV-692 empty shells for use with UM-70 capsules and this most interesting.

Click to expand...

Good plan. I'd thought about this but decided to keep my MV690 as a sand-state alternative for comparison. But there's certainly room in there for a nice tube circuit like the M582. Did you keep the transformers? I wondered if they could be made suitable for a tube output.?

More soon.

z50

zebra50 said:

I see a lot of stuff written about the quality of this transformer, or lack thereof, but I'm trying to see it for myself...

Click to expand...

great - let us know your findings. a lot of the written stuff underlies a certain view of east-german politics 
it comes down to the lamination used,the alloy, the stamping quality and especially the final magnetic heat treatment.
if the lams are from the better quality, the transformer is quite ok, but its not in the bv8 league.

removing the feedback/bypassing/100k ok + give the ef95 a chance.

This is a great thread, and it may actually save me some time. I have a UM57, which I gave quite a bit of TLC but kept in its original configuration, except for a film output cap (same value, though). I found a good tube for it, too. I quite liked it when I heard it solo on my own voice, but when I used it on my friend the other day, it sounded murky and even slightly distorted (which may be just a subjective impression). So I'm thinking about departing from my original plan and putting in a new or modified circuit.

Max, interesting! I was actually considering an EF95 as well. I think it's a good microphone tube, and while not pin compatible, obviously, it is also seven pin.

I'm not sure about the transformer, either. Oliver Archut keeps saying it is bad and inconsistent. That may be true, but then again he sells replacement transformers. I'm gonna measure the transformer in UM57 later on so we can compare notes. At least we should be able to find out how much they vary in inductance and DCR. If, in fact, it turns out to be bad, we'll challenge Max to wind us a better one!  ;D

FWIW, I have a (Thiersch refurbished) UM70S head (not the separate version but the head of the non-modular mic of later manufacture). I once put it on my UM92S, but if I remember correctly, only one side of the capsule was connected. So these aren't fully interchangeable. I haven't tried swapping heads with my UM57, as the capsule is still in good condition. The UM70S and UM92S capsules, as I said in the other thread, had warped diaphragms.

Max, if you want a UM57 tranny for measurements drop me a PM.

Oh boy, my meters are having a bad day or maybe the transformer inductance fools them.
As best as I can tell, DCR is about 2k : 25 ohms
Inductance on the secondary is 1.2 H, the primary is off the scale (my meter only goes to 20H)

The transformer is labeled Ü551
Serial early in the 1300s, Neumann badged.

For comparison purposes I also measured my UM92S (which sounds better)
I didn't take it apart, so I only have measurements for the secondary.
DCR is about 70 ohms, Inductance is 2.2 H

I believe CJ checked one of the transformers and posted about it in a cmv thread.

I worked on two CMVs.  One sounded better with the bypass and one sounded better stock both had the output cap changed to Mylar film.  The resistors looked like the green ones like in this thread pictures and were fine they seemed to have no excess noise.  We wanted to keep the microphone close to stock so I did not adjust the operating points of the circuit.

Also the two power supplies were different(no internal amps in them) and one sounded better.  So IMO question what you read on the web about microphones

The EF92, 1/2 of a 12AT7.

http://www.mif.pg.gda.pl/homepages/frank/sheets/093/1/12AT7.pdf
http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6AB4.pdf

Wow, loads of great discussion already. I managed to get a bit of time today to experiment and take some measurements.

But first off, a question. Am I right in recalling that bypassing the cathode resistor will change the phase of the signal?


Next, some transformer measurements for Max. Like Rossi, my meter does not go very high, so I couldn't get readings on the primaries.

UM57 SN 490x (VEB):
Primary - 1K9 ohms, Secondary 2.10 H, 21.4 Ohms

I put a circa 1V signal across the primary at 100Hz, 1kHz and 10kHz and measured the ratio. Measured volts ratios were 9.00, 9.07 and 8.91 - so a Volts ratio of 9.0:1

For the others I simply measured the secondary
UM57 SN 409x (Neumann):
Secondary = 1.84 H, 21.9 Ohms DCR

CMV551 No. 23x:
Sec = 2.12 H, 24.1 Ohms

(Note that this CMV551 arrived with a CMV563 circuit. It looked like it had been done at the factory.)

I do question what I read, which is why I'm doing this!


I should say a little thing about the mics... basically I have three to compare with slightly different circuits.

1. UM57 (VEB) stock circuit, all original parts, sealed capsule and original electro output cap, RFT (mic grade) tube.

2. UM57 (Neumann) is nearer stock, but with an added 47uF cathode bypass cap, and with the transformer rewired to remove the feedback. This is a very simple and easily reversed mod. Film output cap.

3. CMV551 with UM70 & M71 heads. This has had all of the resistors & caps replaced, 47uF cathode bypass cap added, Rp changed to 100K, transformer rewired to remove feedback, 100M resistors upped to 470Meg. The capsule voltage was set to 60V to match the UM57s in cardioid.  

The next thing i did was to connect my frequency generator ("Jenny"!) and inject circa 200mV to the head amps for the three mics. I measured voltage at the output across the frequency spectrum so I can calculate gain.

I'm drawing up some dB graphs now.

zebra50 said:

But first off, a question. Am I right in recalling that bypassing the cathode resistor will change the phase of the signal?

Click to expand...

Wouldn't think so; your just shunting the AC impedance of the cathode resistor.

Stubo, while you've got the generator on the grid there, I usually measure the clipping threshold of the amp (choose your own reference level) and frequency response, etc.

If you test either circuits in fixed-bias mode, then consider swapping valves at some point to see for any changes in case either are biased a bit "off".

rodabod said:

Wouldn't think so; your just shunting the AC impedance of the cathode resistor.

Click to expand...

Mmm... Getting my head in a twist vis a vis phase shift around the roll off point. Some goodish info here

http://www.freewebs.com/valvewizard2/ChoosingBypassCaps.pdf

Anyway, these three mics do NOT all have the same polarity wrt to the SM57 test. Could be a Gefell thing, could be a Stubo thing.... More work required!

Stubo, while you've got the generator on the grid there, I usually measure the clipping threshold of the amp (choose your own reference level) and frequency response, etc.

Click to expand...

Damn fine idea. I should have thought about that before I put them back together for audio tests... It's only 3 screws and a couple of clips.

If you test either circuits in fixed-bias mode, then consider swapping valves at some point to see for any changes in case either are biased a bit "off".

Click to expand...

Makes sense. I'm trying to reduce my variables but they keep on creeping in. A plan for later is to wire in the decade box and tweak the bias on the bench.

Fncking graph programme just dumped my freak data. At least my bit of paper is crash proof. Arse... start again...

Sooo....

Let's remind ourselves of the circuit, and the mods -
(A) cathode bypass cap (47uF),
(B) remove the feedback, and
(C) Increase R4 from 47K to 100K.



And here are the frequency plots I got for three variants - the signal was inserted at point A. Measured at 25, 30, 40, 60, 80, 100, 125, 160, 200, 250, 300, and so on up to 20K.



***Revised graph***
*** Note that these plots are with a 2K resistor as load for the microphone output***

Note the lines have been moved up so we can see the frequency responses all at a good scale on the graph. Adding the cap and removing feedback gave us additional gain compared with the stock mic of about +4dB at 1kHz. If we switch the plate resistor to 100K too we get about +6 dB.

The stock circuit is pretty flat, but a gentle roll off appears at the bottom end when the cathode bypass caps are present.

Remember these are the plots for the head amp circuit, not the microphone response including the capsule.

Of course I checked the output of the signal generator at point A across the frequency bands - it was pretty flat throughout the experiments (192.1±1.0 mv). (I picked about 200mV because it didn't clip and it gave me convenient numbers to read on my meter.) I also put both the input and output on the scope so I could keep an eye on the waveforms and make sure they remained good sine waves, which they did.

(I'd be grateful for any comments on the test methodology and keeping errors down.)

Gus said:

I believe CJ checked one of the transformers and posted about it in a cmv thread.

Click to expand...

Here it is.

http://www.groupdiy.com/index.php?topic=9310.0

Unfortunately the picture links no longer work, so the thread is a bit lacking in content. CJ?

*** Edit: Revised graph. ***

zebra50 said:

If we switch the plate resistor to 100K too we get +6.5 dB.

Click to expand...

It also looks like the bottom-end is rolling off even faster due to lack of primary inductance for that impedance.

All the circuits show about 4db roll off at the top end, from about 4k up to 20k. Is this the o/p transformer?

Click to expand...

Can't think what else it could be since the feedback is linear from what I can see. Odd. Intentional?

You could disconnect the cable and drive the transformer with your signal generator to see how it responds with that. It might be handy to overlap your response curves even though it throws away the nominal gain information.

I have a DIY mic of this mod circuit using the transformer CJ tested and an M7 lolipop , but I have two more changes that Olly recommended . R2 up to 1G  ,C3 up to 1.5 uF  (this likely gets the low end closer to stock) I prefer this mic to the CMV563 mics I sold.

Guys, I'm using a different power supply today and I'm getting different numbers so have pulled the graph above.  I have loads of data to crunch - will give an update shortly when I've worked it all out...

OK, I've revised the graph above - I had some problems measuring my signal generator's response at higher frequencies for my baseline measurement, which had introduced the steep roll-off at the top. I think it's sorted now, and the charts look a bit more believable. Errors here are probably significant (log of something divided by something small) so view bumps and humps with a pinch of snuff etc.

Anyway, next up is the CMV fully modded mic at different loadings.... Nothing too surprising here.... if you load down the mic you bngger up the top end. I think we all knew that.

rodabod said:

It might be handy to overlap your response curves even though it throws away the nominal gain information.

Click to expand...

I've 'shifted' the plots by a fixed number of dB so we can look at them together.


I'm just about to measure the other mics for comparison so stay tuned.