U87 circuit 04 mod (daughterboard)

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ewayz7

New member
Joined
Dec 9, 2020
Messages
3
So I've been scouring the web looking for a way to mod my u87ai by removing the daughterboard. I've and my u87ai for some months now and I'm just not satisfied with it., mainly due to the fuzzy uppermids which can be overly brittle/fuzzy, especially when using the hpf. I've read that the mod was relatively easy to do. I'm pretty handy with a soldering iron, just need a little direction with selecting the resistors. I read that circuit 06 is upgraded with a biased Fairchild Jfet (2N3819 ) Is there anyone who's pulled this off? Any info would help.
 
I don't mean to be discouraging, but you won't get anywhere by doing this. Maybe lower the resale value of the mic.

Whatever fuzzines you hear is not due to that daughtetboard. Regardless of what you read on the web, or what certain gurus might claim. Neumann selects 2N3819, are you sure you can get hold of a good batch to select from? You also need to bias that fet properly after replacing it. I don't know if de-emphasis part of the circuit might need some adjusting.

I would personally just sell that mic, and find one that I like better.

I'll also quote Uwe Sattler:


If it isn't broke, don't try to 'fix' it! Contrary to what may be expressed here and elsewhere, the single FET substitution with the daughter board is actually a real improvement. There is still effectively only one active FET (selected for minimal noise) in the signal path, the other seemingly cascaded one in conjunction with a dual PNP-pair automatically optimizes the operating point, and the single NPN forms the load resistance for the FET. Compared to the original sindle FET circuit the result is lower noise, slightly better dynamic range and/or lower distortion due to the automatically optimized and more stable operating point. Going out on a limb, I would argue to replace to replace any (defective) FET in the older U87A models with this daughter board!
 
Last edited:
The daughterboard was used in the versions #3 to #5.
From version #6 Neumann has gone back to a single FET.
(By the way: there is no need to change the de-emphasis network when removing the daughterboard and installing a FET. Of course the bias should be adjusted accordingly.)
 
Hey thanks for your responses
.
@kingkorg , I managed to track down the fairchild (onsemi) 2n3819 jfet. Also I actually saw that comment in another forum, followed by what Klaus had to say about the different circuits. For right now, I have flipped the capsule around yesterday and it sounds a tad cleaner albiet brighter, it's still there though. Selling is also an option, but with that comes option paralysis, there are so many supposedly good vocal mics out there that I'm not sure where to even start for a OSFA mic, especially under 2.5k. I have no way of actually testing most of them. There are shootout videos, and zen pro's clipalator but in the real world those are pretty much useless for a number of reasons. I've been through a number mics before I landed on this one, and I only purchased this one because the u87ai is tried and true.

@Ricardus, when I say overly brittle and fuzzy, what I mean is the mic sounds generally unpleasant in the 2k to about 5k area, or the high mids. Almost like a distortion, gainstaging is not an issue, and I've done the cable mod, I wasn't getting any rfi but weirdly enough the mic sounds way more "open" with the modded cable (shield to ground). Not sure which is correct, but the mic does seem to have a different tonality. Still didn't solve the issue in the upper mids though. I find myself being overly careful when I'm mixing to not emphasize that when processing which I try to keep very minimal, that being said I can't highpass above 70 cycles, anything above that and it goes to complete ****. Just trying to get a vocal to sound less dull without exposing the upper mids is more than a task for a mic that cost 3.2k new.
 
there are extremely good fakes floating around, can we get a picture?

i will say, i have a u87i, not an ai, but wouldn't not describe the u87 mids as fuzzy and crisp, but kind of in a good way. something about how it picks up sound picks up room modes/issues way way way more too.

you can kind of hear it here vs the gefell:


You can hear way more of the room's issues in the u87 recording compared to the gefell. more than just the 87 being more off target here which it is. Not sure why this is, but mine does it too. Sounds perfect in extremely well-designed rooms, but (at least mine) really doesn't like any less than that. When i used my 87 in my vocal booth, i had a narsty peak at 2.5-3khz that it took a lot of panel-shuffling and repositioning to get rid of. after that, the 87 sounded way better, and my other 2 mics sounded mildly better in the same position. it's like a magnifying glass. no clue.

it could just not be to your taste/use case, but we should verify it's real first to begin troubleshooting this.
 
Last edited:
Sorry for a reply after so much time, but the OP's situation is somewhat rare, so I thought I could help.

ewayz7,

The circuit rev 04 and 05 are unfortunate ones, with several bad decisions cascading into a fizzy, nasty sounding mic that is mostly derived from the sub-board, which completely replaces a FET. This was due to Neumann's difficulty in sourcing high quality 2N3819s. However, there is a fix. I had mine modded by James Brian at the mic shop. He did the usual mod to remove/modify the high and low pass filters inherent to U87s that enable them to conform to German broadcasting standards, but make the mic sound less musical. He can also remove the sub-board, and add a true FET. He has one that he likes, but you can request an actual 2N3819, and he might be able to accomodate you. Also, I recommended another change that marked a first time mod for James. I asked him to lower the capsule polarization from 60v to 48v. (48v is what classic Neumanns used). This change "relaxed" the tone of the mic, and made it less brittle. There is a slight loss in sensitivity, but U87s are so over-the-top loud anyway, it actually helped. (I got the idea when I tried a Miktek that had the 48/60 switch, and noticed how much better it sounded in the 48v setting).

I love my U87 now, and it can hold its own against anything, including my vintage original C12 and U67, plus a whole bunch of modern mics I own. Try it!
 
I don't mean to be discouraging, but you won't get anywhere by doing this. Maybe lower the resale value of the mic.

Whatever fuzzines you hear is not due to that daughtetboard. Regardless of what you read on the web, or what certain gurus might claim. Neumann selects 2N3819, are you sure you can get hold of a good batch to select from? You also need to bias that fet properly after replacing it. I don't know if de-emphasis part of the circuit might need some adjusting.

I would personally just sell that mic, and find one that I like better.

I'll also quote Uwe Sattler:


If it isn't broke, don't try to 'fix' it! Contrary to what may be expressed here and elsewhere, the single FET substitution with the daughter board is actually a real improvement. There is still effectively only one active FET (selected for minimal noise) in the signal path, the other seemingly cascaded one in conjunction with a dual PNP-pair automatically optimizes the operating point, and the single NPN forms the load resistance for the FET. Compared to the original sindle FET circuit the result is lower noise, slightly better dynamic range and/or lower distortion due to the automatically optimized and more stable operating point. Going out on a limb, I would argue to replace to replace any (defective) FET in the older U87A models with this daughter board!
KingKorg,
This quote from Uwe represents a minority who actually believe the daughterboard makes the mic better. As Klaus pointed out to him, if this is so, why did Neumann return to a FET after they sourced the Fairchild 3819s? There were several good points made during the quoted discussion, but Uwe's wasn't one of them, unless all you're interested is headroom and self noise. As an owner of a rev 04 that has returned it to a FET state, I can telll you the improvement is substantial.
 
The vague sound reminds me of a project long ago. Later, I modified the type of resistor to solve the problem. It is a brown resistor produced by DALE as a bias resistor. I modified it to a common metal film resistor. The sound is much clearer. I don't know if it will help you
 
KingKorg,
This quote from Uwe represents a minority who actually believe the daughterboard makes the mic better. As Klaus pointed out to him, if this is so, why did Neumann return to a FET after they sourced the Fairchild 3819s? There were several good points made during the quoted discussion, but Uwe's wasn't one of them, unless all you're interested is headroom and self noise. As an owner of a rev 04 that has returned it to a FET state, I can telll you the improvement is substantial.
I personally don't value anything Klaus has to say, i'll take Uwe's opinion over his all day long. That is just me, ymmv, i am one of those empirically driven fools not interested in audio foolery.
 
The vague sound reminds me of a project long ago. Later, I modified the type of resistor to solve the problem. It is a brown resistor produced by DALE as a bias resistor. I modified it to a common metal film resistor. The sound is much clearer. I don't know if it will help you
Seriously?
 
Sorry for a reply after so much time, but the OP's situation is somewhat rare, so I thought I could help.

ewayz7,

The circuit rev 04 and 05 are unfortunate ones, with several bad decisions cascading into a fizzy, nasty sounding mic that is mostly derived from the sub-board, which completely replaces a FET. This was due to Neumann's difficulty in sourcing high quality 2N3819s. However, there is a fix. I had mine modded by James Brian at the mic shop. He did the usual mod to remove/modify the high and low pass filters inherent to U87s that enable them to conform to German broadcasting standards, but make the mic sound less musical. He can also remove the sub-board, and add a true FET. He has one that he likes, but you can request an actual 2N3819, and he might be able to accomodate you. Also, I recommended another change that marked a first time mod for James. I asked him to lower the capsule polarization from 60v to 48v. (48v is what classic Neumanns used). This change "relaxed" the tone of the mic, and made it less brittle. There is a slight loss in sensitivity, but U87s are so over-the-top loud anyway, it actually helped. (I got the idea when I tried a Miktek that had the 48/60 switch, and noticed how much better it sounded in the 48v setting).

I love my U87 now, and it can hold its own against anything, including my vintage original C12 and U67, plus a whole bunch of modern mics I own. Try it!
The filter that corrects the top end of a U87 is not a broadcast filter, it is to correct the inherent top end rise of the capsule.
 
I personally don't value anything Klaus has to say, i'll take Uwe's opinion over his all day long. That is just me, ymmv, i am one of those empirically driven fools not interested in audio foolery.

That's fine, but in this case, Uwe compomised his credibility by claiming that Neumann had not changed back to the 3819 when they in fact had. He also misread the circuit diagram, claiming it was for rev 07 (which didn't even exist at the time and still doesn't) and showing the older daugherboard when the data sheet clearly read 04 with a date consistent with the rev 4 release.

As for "audio foolery," the fact that the sub-board yielded a lower noise floor and higher sensitivity doesn't mean it sounds better. The fact that Neumann switched to the sub-board due to supply issues for audio quality 3819s, (the drying up of 3819s is well documented), and have now switched back to them now that they received a run from fairchild should tell you something about the audio advantages of a quality FET.

Anyone claiming sonic superiority (not "spec" superiority) of the Neumann sub-board over a good quality, well biased 2N3819 in a U87 is in a minority, including the manufacturer, who obviously heard enough advantage to go back. I can't believe Neumann would take a step backwards into FETs on their U87s without a solid sonic rationale.
 
Last edited:
The filter that corrects the top end of a U87 is not a broadcast filter, it is to correct the inherent top end rise of the capsule.
That's partially true. K870s are very bright capsule, and without filtering, the mic would be unbearably strident. However, the degree of filtering on stock 87s does take broadcast into account. The best mods for U87s include a different filter to allow more highs through, just not all of them. The result is a much bigger, wider, fuller mic with the sensation of being less harsh in the mids (which is really an effect of opening the bottom and top end).

I really appreciate Larry at ADK's approach to capsule design. He designs them to sound sonically correct without the need for too much filtering and negative feedback.
 
That's partially true. K870s are very bright capsule, and without filtering, the mic would be unbearably strident. However, the degree of filtering on stock 87s does take broadcast into account. The best mods for U87s include a different filter to allow more highs through, just not all of them. The result is a much bigger, wider, fuller mic with the sensation of being less harsh in the mids (which is really an effect of opening the bottom and top end).

I really appreciate Larry at ADK's approach to capsule design. He designs them to sound sonically correct without the need for too much filtering and negative feedback.
Guosheng Zhuang is responsible for the capsule design at ADK, not Larry IIRC, but they are indeed nice capsules. Warm also consulted him for the WA-87. Nice guy. Makes a mean M7. He actually sells his own mics on E-Bay:
https://www.ebay.com/str/3uaudioThe Warblers are really nice, and the parts are top-notch. The Warblers actually have a tunable HF attenuation, which makes them extra relevant to your post.
 
Last edited:
Back
Top