Is there any new mic you can buy or a kit to build that sounds as good as a vintage U87i?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I wrote I didn't like the U87's high end... Well, I went back and listened to the recordings of the other mics, and found I didn't like theirs either. It turns out I was using a crummy converter. After trying another AD I find the modded U87ai to sound great overall. Pretty much the swiss army knife of mics it is known for. Just love the stiff, defined Neumann mids. No ringing, everything sounds balanced and dry and very detailed without the artificial sparkle of many modern mics.

The only fault of the modded mic is the increased proximity effect due to the extended frequency range. Well, you can't have it all... :cool:
 
Here's an image of the mic before and after the mod. The changes made I have noticed are circled in red.

R 103 went below the PCB before, now it's a different resistor and goes to a different solder point.

The FET has been changed - can someone identify the type?

A small cap (feedback?) has been added.

The polystyrene cap C105 on the far right has been changed (was physically bigger before).


I must say it doesn't look too clean for a high impedance circuit and the red plastic casing on the right was slightly damaged with the soldering iron.
 

Attachments

  • U87ai_mod.jpg
    U87ai_mod.jpg
    173 KB
  • U87_before.jpg
    U87_before.jpg
    121.4 KB
I've had a closer look at the mod. It looks like they replaced the FET for a BF256B (it says's Vishay on the invoice, but I don't think that exists and the markings look like Motorola), added a 339p COG cap between drain and gate and replaced the 160p LP filter cap with a 220p one. The cap mods are the same (filter part) or reminiscent (source2gate) of the U87i.
 
Last edited:
I've had a closer look at the mod. It looks like they replaced the FET for a BF256B (it says's Vishay on the invoice, but I don't think that exists and the markings look like Motorola), added a 339p COG cap between drain and source and replaced the 160p HP filter cap with a 220p one. The cap mods are the same (filter part) or reminiscent (source2drain) of the U87i.
339pf from drain to source?
Is this to add some distortion?

Nothing wrong with a BF265B however what company JFET is that(I see you posted Vishay)and what was the original? The original might have been fine.
Does the BF265B have a source resistor around 5k

No cleaning up of the flux?
An old joke about soldering "the bigger the blob the better the job"

I tune a 87 circuit a bit differently
 
Typo, meant between drain and gate, corrected. The original U87i had 470p + 10p in series between drain and gate (C1+C3), albeit in series with the 470nf feedback cap C9.

You can see the markings on the transistor in the photo above - if anyone knows the manufacturer let me know. To me it looks like Motorola upside down.

The original was a Fairchild 2N3819 and the mic sounded a little strained, now it sounds really smooth and clear.

There's a trimpot to set the source resistor value for lowest distortion and biasing was included in the mod, of course.


So how would you go about tuning a U87 circuit?
 
"By the way", as a direct answer to the question posed in the initial post in this thread:

https://groupdiy.com/threads/akg-perception-p220-to-neumann-u87-5-min-mod-p200-p100-p400-p420.67473/
Interesting, thanks!

Regarding the U87ai, I find it strange that Neumann uses a pricy, rare and huge BC components electrolytic capacitor for power filtering in one place, but a tiny, cheap, low cost Jamicon 85° electrolytic for the signal path (after amplification, but still).
 
Well, given the price points they sell their stuff at, BOM consolidation / optimization may very well not be at the top of their priority list...
 
II have measured and changed many capacitors over the decades. Jamicons I checked had good measurements.

Also don't get hung up on 85C and 105C.

A 3819 and 265 are the same JFET the B is selected.

Maybe all the microphone needed was to be biased.

Is this a ai or i model? I don't see a drain to gate cap in the ai schematic
339pf does not make sense drain to gate that would reduce the gain a lot
is it more like 3.3pf? that would make more sense

I also don't understand changing one polystyrene cap for another

http://recordinghacks.com/microphones/Neumann/U-87-Ai
No chance of me sharing in the forum how I currently adjust 87 or JFET PNP china circuits.

https://groupdiy.com/threads/reply-4-in-the-link-it-is-about-u87-microphones.80704/
 
II have measured and changed many capacitors over the decades. Jamicons I checked had good measurements.

Also don't get hung up on 85C and 105C.

A 3819 and 265 are the same JFET the B is selected.

Maybe all the microphone needed was to be biased.

Is this a ai or i model? I don't see a drain to gate cap in the ai schematic
339pf does not make sense drain to gate that would reduce the gain a lot
is it more like 3.3pf? that would make more sense

I also don't understand changing one polystyrene cap for another

http://recordinghacks.com/microphones/Neumann/U-87-Ai
No chance of me sharing in the forum how I currently adjust 87 or JFET PNP china circuits.

https://groupdiy.com/threads/reply-4-in-the-link-it-is-about-u87-microphones.80704/
Thanks!

It's the ai. I sent it there (brand new) to get it sound closer to the none-ai. The explanation given was that they would change the JFET to a lower noise type, therefore being able to lower the amplitude without a noise penalty to achieve lower distortion / a higher headroom. Changing the polystyrene cap does make sense, it is 220p in the i but 160p in the ai, so using a higher value will lower the LPF frequency.

I am currently working on a cardiod only mic with a U87ai circuit. Adding the 390p cap (yes, that's what it's in the U87ai markings are - KAG 3900, so it's a COG 390p cap) did indeed lower the gain a lot. This way I could tweak the source trimpot for very nice low distortion (since the FET doesn't get much signal), but of course the amplitude is way too low for an LDC .

What would be the point of adding in a 3.3p cap between drain and gate?
 
According to the schematic I have here, in the U87 (non-ai) there was a 10 pF negative feedback capacitor.
Technically speaking, it is between drain and gate.

The value change from 160 pF to 220 pF in the de-emphasis network will certainly change the sound character. It will cause the highest frequencies to roll-off at a lower frequency.
 

Attachments

  • U87.PNG
    U87.PNG
    41.7 KB
According to the schematic I have here, in the U87 (non-ai) there was a 10 pF negative feedback capacitor.
Technically speaking, it is between drain and gate.
That's the "i"-schematic I've got here, too. But does it do anything in a cardiod-only variant? I would have thought the 10pf+470pf caps were in series vs the front back plate and thus feedback would add up to 480pf. But I don't really understand discrete circuits...


The value change from 160 pF to 220 pF in the de-emphasis network will certainly change the sound character. It will cause the highest frequencies to roll-off at a lower frequency.
Yes.
 
Look at reply 72 in one of the links I posted
Also look at the KM84 schematic
 
Look at reply 72 in one of the links I posted
"Furthermore, 2 dB of the nearly 6 dB increase in maximum output are explained by elimination of the 10pF feedback capacitor, with another 3.8 dB gained from a reduction in THD by the reduced capacitive load on the FET gate."

If I understand this correctly adding a feedback capacitor actually increases THD and lowers the output. But lowering the output also lowers the clipping point and thus increases headroom, right? So what would be the reason for adding this cap when modding a U87ai?
 
Back
Top