Neumann Vintage U87 Clone : Build Thread.

[joulupukki]

Following up on this JFET selection discussion, I’ve tried a couple other JFETs in the mic and have settled on a 2N3819 that has an Idss of 9.6mA, biasing it for symmetrical clipping. I think it sounds pretty good. It’s a lot less woofy than the first JFET I tried but still keeps the raw clear quality.

During my swaps of JFETs I did hit two issues:

1. I accidentally pulled the solder pad off the underside of the drain pin connection. I was being a bit too aggressive with the tip of my soldering iron. I soldered the current JFET to the board’s top pad for now and am probably done swapping anyway

2. I discovered that the reason some of these weren’t sounding “right” to me was that I had a couple of subtle compressors enabled in my signal chain in the Apollo Twin. GRR! The recordings were sounding closer to how it sounded when I had the J305 JFET installed – which makes me think that I probably had the compressors enabled when the J305 was installed. So dumb of me!

 

[MicMaven]

which makes me think that I probably had the compressors enabled when the J305 was installed. So dumb of me!

Ha - No worries, old man. Just use your DAILY MULLIGAN and it will be like it never happened. Of course you only get one Mulligan per golf game, so my wife says we only get one Mulligam per day of married life. Still, it may be the easiest way out !!! / Happy trails to you, James /

 

[joulupukki]

Here's a quick sample of the DIY U87 compared with the Townsend Labs Sphere L22 with the LD-87 Vintage microphone model. This is using a 2N3819 JFET that has the lowest Idss value (9.64mA) of all the ones I had. I've biased it for symmetrical clipping.

View attachment DIY-U87-vs-TownsendLabs-Sphere-L22-LD-87-Vintage.wav

 

[egroys]

Here's a quick sample of the DIY U87 compared with the Townsend Labs Sphere L22 with the LD-87 Vintage microphone model. This is using a 2N3819 JFET that has the lowest Idss value (9.64mA) of all the ones I had. I've biased it for symmetrical clipping.

View attachment 126749

L22 feels scooped and has a crispy/brittle top that is unpleasant to my ear.
I have yet to try a modeling mic, but it really does come across like a veneer of a U87 here.

Thank you for sharing and enjoy your new mic!

 

[AdamFrisen]

For me, the version using 2N3819 sounds more polished and even than with J305. I compared by loading both in Audacity with loudness normalization.

I would prefer listening to long book readings from 2N3819.

I cranked up the silent part at the ends and the noise floor sounded different to me too.

 

[jeremynothman]

Here's a quick sample of the DIY U87 compared with the Townsend Labs Sphere L22 with the LD-87 Vintage microphone model. This is using a 2N3819 JFET that has the lowest Idss value (9.64mA) of all the ones I had. I've biased it for symmetrical clipping.

View attachment 126749

DIY sounds like it has its low pass filter set a bit lower (to me, in a good way; it makes it feel solid). The Sphere is sounding thin to me by comparison and seems to be adversely (negatively) affected by moving in and out of the proximity effect. Maybe you just moved less while recording with the DIY mic, or maybe it handles that better

 

[TimG]

For a Moderator:

Could someone make this an active link in the first post?
I hadn’t realized its importance. Maybe others too?

Read This First : Assembly Instruction , Excel BOM and Tips For Assembly and Important Considerations Here: https://cdn.groupbuilder.com/groupdiy/u/39511/58d1402a02019.xlsx

 

[EarthTone]

Hey guys! Quick question. I found a similar post in this thread but please, just to set my mind at ease: For C14 and C2 in this build, the Wima part was out of stock where I live. For now I just ordered this: https://www.mouser.com/ProductDetail/598-185103J100RAA-F

Will this Cornell Dubilier part be ok? Seems to be the same rating... 0.01uf 100VDC

Will it sound the same?

Alternatively I have a pair of Wima 0.01uf's but they are only 63V... Is that too low?

I can try to source the Wima 0.01uf 100V parts if necesssary. Just since I already have these, I wonder if one of these parts I already have will be OK?

Thanks!

 

[dannyjel]

Hey all, i have a question and i hope it's OK to ask it here, i'm trying to identify a component on the original U87Ai schematic,
i have attached a pic of the schematic with the component marked in a circle, can someone please help?

 

[abbey road d enfer]

It's a tapped inductor, part of the bias oscillator.

 

[GhostlyStudio]

I̶'v̶e̶ j̶u̶s̶t̶ f̶i̶n̶i̶s̶h̶e̶d̶ m̶y̶ U̶8̶7̶ b̶u̶i̶l̶d̶ u̶s̶i̶n̶g̶ D̶a̶n̶n̶y̶'s̶ b̶l̶u̶e̶ p̶c̶b̶'s̶ a̶n̶d̶ B̶O̶M̶, a̶s̶ w̶e̶l̶l̶ a̶s̶ a̶n̶ A̶M̶I̶ T̶1̶3̶ t̶r̶a̶n̶s̶f̶o̶r̶m̶e̶r̶, a̶n̶d̶ f̶o̶l̶l̶o̶w̶e̶d̶ t̶h̶e̶ b̶u̶i̶l̶d̶ g̶u̶i̶d̶e̶ o̶n̶ h̶i̶s̶ w̶e̶b̶s̶i̶t̶e̶. C̶h̶u̶c̶k̶e̶d̶ i̶n̶ a̶ T̶S̶C̶-̶2̶ c̶a̶p̶s̶u̶l̶e̶ f̶r̶o̶m̶ J̶L̶I̶ t̶o̶ c̶h̶e̶c̶k̶ e̶v̶e̶r̶y̶t̶h̶i̶n̶g̶ w̶o̶r̶k̶s̶ f̶i̶n̶e̶ w̶h̶i̶l̶s̶t̶ I̶ w̶a̶i̶t̶ f̶o̶r̶ o̶n̶e̶ o̶f̶ A̶r̶i̶e̶n̶n̶e̶'s̶ K̶8̶7̶'s̶ t̶o̶ a̶r̶r̶i̶v̶e̶ a̶n̶d̶ e̶v̶e̶r̶y̶t̶h̶i̶n̶g̶ a̶p̶p̶e̶a̶r̶s̶ t̶o̶ w̶o̶r̶k̶ a̶s̶ e̶x̶p̶e̶c̶t̶e̶d̶, b̶u̶t̶ t̶h̶e̶ o̶u̶t̶p̶u̶t̶ s̶e̶e̶m̶s̶ q̶u̶i̶t̶e̶ l̶o̶w̶?̶ P̶l̶u̶g̶g̶e̶d̶ i̶n̶ m̶y̶ A̶K̶G̶ P̶4̶2̶0̶ i̶n̶t̶o̶ t̶h̶e̶ s̶a̶m̶e̶ a̶u̶d̶i̶o̶ i̶n̶t̶e̶r̶f̶a̶c̶e̶ a̶n̶d̶ t̶h̶e̶ d̶i̶f̶f̶e̶r̶e̶n̶c̶e̶ i̶n̶ v̶o̶l̶u̶m̶e̶ i̶s̶ n̶o̶t̶i̶c̶e̶a̶b̶l̶e̶. I̶s̶ t̶h̶e̶r̶e̶ a̶ g̶o̶o̶d̶ p̶l̶a̶c̶e̶ t̶o̶ s̶t̶a̶r̶t̶ i̶n̶ t̶e̶r̶m̶s̶ o̶f̶ d̶e̶b̶u̶g̶g̶i̶n̶g̶ w̶h̶e̶r̶e̶ I̶'v̶e̶ l̶i̶k̶e̶l̶y̶ g̶o̶n̶e̶ w̶r̶o̶n̶g̶?̶

Sorry, cancel that... I think I misunderstood that the original U87 circuit has less output compared to the U87ai (which the P420 is modelled off). I'm able to give more gain and speaking normally into the Mic comes out fine / clean.

 

[GhostlyStudio]

I've just finished my U87 build using Danny's blue pcb's and BOM, as well as an AMI T13 transformer, and followed the build guide on his website. Chucked in a TSC-2 capsule from JLI to check everything works fine whilst I wait for one of Arienne's K87's to arrive and everything appears to work as expected, but the output seems quite low? Plugged in my AKG P420 into the same audio interface and the difference in volume is noticeable.

Is there a good place to start in terms of debugging where I've likely gone wrong?

For some wider info whilst I've been tinkering this AM.

It does look like the JLI TSC-2 capsule backplates aren't isolated so figure 8 not playing nicely...