Modding Neumann U87Ai to U87i voicing and headroom

[RuudNL]

I have even gone lower! In one of my microphones I have included a -10 dB/-20 dB switch.
0 dB = 60 volts, -10 dB = 19 volts. -20 dB = 6 volts.
But I think I never needed the -20 dB position...

 

[mean_street]

So, from what I understand there are several contributors to the headroom of the U87i compared to the U87Ai. The polarisation voltage makes only for a difference of about 2db, the biggest contributor is the difference of transducer capacitance due to different capsule construction. With the mentioned modifications the Ai would be close, but still less headroom than the original U87.

 

[kingkorg]

So, from what I understand there are several contributors to the headroom of the U87i compared to the U87Ai. The polarisation voltage makes only for a difference of about 2db, the biggest contributor is the difference of transducer capacitance due to different capsule construction. With the mentioned modifications the Ai would be close, but still less headroom than the original U87.

Nope, you get maybe 2db by lowering to 50v zener. To get more attenuation you need to lower the voltage (zener voltage) even more. Like @RuudNL suggested, there is really no lower limit.

 

[mean_street]

Nope, you get maybe 2db by lowering to 50v zener. To get more attenuation you need to lower the voltage (zener voltage) even more. Like @RuudNL suggested, there is really no lower limit.

Yes, I meant 2db at around 50V. I’ll try the zener mod and also will install the feedback cap and report back.

 

[MagnetoSound]

Maybe you don’t want to lose sight of the fact that improving the voicing and increasing the headroom of the mic are not exactly the same quest.

Once you have made all the circuit changes to that of the non-A variant, any audible difference you have left will be down to the tuning of the capsule. There is no further tweaking to be done in the electronics that will not produce an unintended compromise in sensitivity or noise performance.

If it was me, I would be focussing more on seeking the elusive sound of the vintage 87i, and would consider any improvement in headroom a natural bonus. But that’s me

 

[Gus]

Hey guys,
I’ve been using my U87Ai for all things, it sounds great on some but good on almost all sources. Two things that bug me with that mic is the really low headroom and the high mid/treble bump. I already changed C105 to 220pf, to get a little bit less high end response which worked out flawless. For the headroom issue I plan to install a 50V zener at place of D7, to lower the polarisation voltage for additional headroom and installing a cap between drain and gate like A. Grosser did in his mods to further gain some headroom. I would leave the FET as is for the moment.
Anybody got some more ideas or remarks doing those mods?View attachment 111444

I am surprised that A. Grosser used a D to G cap. JFETs do not have the transconductace of a BJT.
Low open loop gain with feed back is not as good as high open loop gain with feedback with the resulting distortion spectrum however, maybe it is fine in a 87Ai.

FWIW years ago there were thread(s) about capsule voltage.
Zebra50 even posted about a build with different heater and capsule voltages it used an old WE body IIRC.

IIRC the TLM50 has a voltage pad at 23VDC.

I use capsule voltage adjustment in most of the microphones. I built one that goes from 5VDC to 70VDC.

The Brauner Phantom C had switches inside to adjust the voltage to the capsule if I traced it correctly from picture in a thread at this forum

IMO the newer 87 is a better circuit than the older one and a voltage control is what I would try.
The S/N at lower voltages is not as good but you are most likely recording something loud so this not matter much.

 

[mean_street]

Maybe you don’t want to lose sight of the fact that improving the voicing and increasing the headroom of the mic are not exactly the same quest.

Once you have made all the circuit changes to that of the non-A variant, any audible difference you have left will be down to the tuning of the capsule. There is no further tweaking to be done in the electronics that will not produce an unintended compromise in sensitivity or noise performance.

If it was me, I would be focussing more on seeking the elusive sound of the vintage 87i, and would consider any improvement in headroom a natural bonus. But that’s me

The 220pf I interchanged with the 160pf is a voicing change, the polarisation voltage should influence both headroom and voicing (?). But I’m conscious that they are two different tasks, yes.

 

[Андрій Чубака]

the polarisation voltage should influence both headroom and voicing

Most like you can't get any better sound from mic with decreasing polarization voltage. Sound just bit more fluffy and weak in this case. So this is not a solution for increasing sound quality of the mic. I don't know why you decided about it is need for you. But as my opinion it is not really working. About modification mic in 3000$ price this way i can't understand it at all. With such expensive products like this mic i am preferred don't create any additional movement of soldering iron on pcb. Just it is possible in case when absolutely sure about changes will be for better sound.

 

[Gus]

You reduce the voltage so you do not overdrive the gain circuit.
You set it for what you are recording at the time to optimize the operation of the microphone.

You need to think of a microphone like a tube guitar amp. You sometimes have two volume controls in a guitar amp.
Think of the capsule voltage as one (it adjusts the output of the capsule) and the fixed (or variable if one is clever) gain stage.

There is more to this.

 

[living sounds]

Read that thread already with great interest. Did you find out what transistor and cap Andreas Grosser was using on the mod?

The answers are the the thread I linked above.

I did mod some clone mics that used correct headbasket geometry pretty good capsules, with high quality transformers from UTM and basic vintage U87 circuitry. It was crucial to get the biasing right (I remember doing a lot of tweaking and listening last year) and keep the crappy electrolytic on the output, maybe its ESR ensures stability, it didn't sound right with "better" caps.

After all that, the clone mic sounded very similar to the U87ai modded by Andreas Grosser, as long as you addressed it straight from the front. Not quite as detailed, clear and nice as the real U87, which I would put done mostly to the capsule and maybe the grill material. Still have to get and try it with one of Soliloqueen's U87 capsules.

 

[mean_street]

So, I’ve had some time today to record some samples, compared the mic with the 160pf vs 220pf in the filter network and also played around with the pad.
With the stock value of 160pf the mic sounds a lot brighter, especially on vocals it sounded more like a polished pop vocal track. With the 220pf it becomes apparent why some people call U87s honky. The mids are much more prominent, a more natural representation of the source.

I’ve also recorded two drum takes, one with and one without pad. With the pad engaged, the mic sounds even more honky and both low and high end get rolled off and lose detail. That additional headroom would come in handy here.

 

[MagnetoSound]

That’s a fairly accurate description of what my vintage U87s sound like, and how they respond to internal padding.

Don’t forget that the bass roll off cap (C104 on your schematic, C5 in the non-A circuit) is different too - and in fact has evolved quite significantly through the years.

 

[mean_street]

So, I put in both the cap and the zener.
I put in the cap first, which reduced the output level by 3db. Then I put in the zener (47V) which reduced the output again by 2 to 3db. After I had everything put together I tried the mic as front of kit and overhead mic on drums. In both cases the additional headroom came really into play, hearing snare and bass drum hits with transients was really refreshing.
I also recorded some vocals and compared them to a take with stock values.
As I’ve said before the Ai-configuration sounds more modern, more intimate, more low mids. The i-configuration (with the lower polarisation voltage) sounds clearer, maybe a little bit honky, more sub bass.

 

[Андрій Чубака]

As I’ve said before the Ai-configuration sounds more modern, more intimate, more low mids. The i-configuration (with the lower polarisation voltage) sounds clearer, maybe a little bit honky, more sub bass.

But you can use equalizer for change frequencys balance and sound will be not a honky. Sound from large condencer mic with high voltage on capsule contain a lot of different tones inside of recording you can just get it on top with eq without any looses in sound quality. At least for me it is working.

In any case if you playing on drums maybe for you is really important this more low voltage on capsule. Sorry i have no time read all this topic. Good luck

 

[RuudNL]

Don’t forget that the bass roll off cap (C104 on your schematic, C5 in the non-A circuit) is different too - and in fact has evolved quite significantly through the years.

Good point. This capacitor has been 33nF and 47 nF and is responsible for the attenuation of the sub-audio frequencies.
But measuremenst show that the attenuation already starts below 60 Hz!
In my (original Neumann) microphones I have increased the value of this capacitor to 150 nF.
See the graph for the difference.

 

[MagnetoSound]

It’s also been as low as 15nF (way back in 1968), increasing feedback at the low end even more, although R9 is much smaller to compensate. I have never tried this configuration myself. It will change the slope somewhat, giving possibly slightly more overall gain. Perhaps they increased this resistor to improve headroom. C3 is also smaller at 3.3pF, while C6 is a whopping 820pF. My feeling is that a lot of these changes compensate one another. As I say, many changes over the years.

 

[mean_street]

At the moment there is enough low end for me with the mic, especially as Front of Kit mic the U87 captures more sub bass then before. Could be a combination of two things, polarisation voltage and that I don’t have to use the pad any longer for this. Same goes for the bleed from the hihat, which is clearer sounding. We will record some cuts with my band tonight, so I will test out this configuration a bit more.
I will record a bit with it first and then I decide if there is any need to change out C104. If anything I would cut some low end and would lower the value to U87i-value.

 

[RuudNL]

C6 is a whopping 820pF.

820pF/1.1K vs. 160pF/6.8K gives about the same time constant.

 

[MagnetoSound]

820pF/1.1K vs. 160pF/6.8K gives about the same time constant.

Close enough, yes.

But the lower resistance shunts more of the overall feedback to ground, increasing the overall output. Together with the smaller value of C3, that’s a much livelier mic, in theory at least.

My suspicion is that the original intention was to produce a U87 that more closely matched the output level that users had come to expect from the U67, until they found that it ran out of headroom on loud sources as close mic’ing came into vogue.

All conjecture, of course, but it would explain the subsequent changes if they were made primarily to implement greater headroom.

 

[ZHOUJUN]

I have repaired several 87AI and 87I. I think the biggest difference between them is the capsule, so changing the circuit components alone should not achieve the desired goal. The best way is to find a vintage K87 capsule.