Neumann Vintage U87 Clone : Build Thread.

[ln76d]

Still have problen to identify this issue.
Assuming that you have proper values of R17, R14, R12 i would try 33V zener instead 24V.  I will look more later.

 

[Neutrino]

Still have problen to identify this issue.
Assuming that you have proper values of R17, R14, R12 i would try 33V zener instead 24V.  I will look more later.

I've cross-checked the values of all resistors (including R17, R14 and R12) by their color-code and they look all correct.
Which values would correspond to the specs for the voltages I measured?

 

[ln76d]

Which values would correspond to the specs for the voltages I measured?

Don't understand your question... Please be more specific

Ok - two things which you can try - first use lower value of trimmer pot for FET bias - 12k-15k instead 25k. If this wouldn't help then try 33V zener instead 24V and set bias again. Note that original U87i (at least these revisions which i know) use 33V not 24V (which was borrowed from km84 circuit .

 

[Neutrino]

Don't understand your question... Please be more specific

You said that my readings are odd. I'm just curious which values you would expect. 

Ok - two things which you can try - first use lower value of trimmer pot for FET bias - 12k-15k instead 25k. If this wouldn't help then try 33V zener instead 24V and set bias again. Note that original U87i (at least these revisions which i know) use 33V not 24V (which was borrowed from km84 circuit .

OK, I will have to order these parts. In the meantime (for my understanding), would you mind explaining. a bit what it is that you are trying to achieve with the different trimmer and Zeners? To my understanding the value of the pot shouldn't really matter as long as it is set correctly.

 

[ln76d]

You said that my readings are odd. I'm just curious which values you would expect. 

Ahh, ok

For your setup:
-both sides of R14: ca. 22V and ca. 24V
-both sides of R2: at least 46V
-c9/c7 point: 10-11 V

With 33V Zener

-both sides of R14: ca. 28V and ca. 33V
-both sides of R2: at least 46V
-c9/c7 point: 10-11 V

OK, I will have to order these parts. In the meantime (for my understanding), would you mind explaining. a bit what it is that you are trying to achieve with the different trimmer and Zeners? To my understanding the value of the pot shouldn't really matter as long as it is set correctly.

For the trimmer i would suggest linear taper. Currently you have double value of what you really need. If you have logarithmic pot, then it is harder set it properly. Lower pot is just for more precise and easier setting.
Judging from your readings, you have higher current consumption than we should expect in this circuit. Making less voltage  drop with the zener you should be able to bias properly FET in the correct range. Since we expect more voltage flowing around then more precise pot will be also usefull If it will work, please write your readings in the same points as previously

 

[Neutrino]

For the trimmer i would suggest linear taper. Currently you have double value of what you really need. If you have logarithmic pot, then it is harder set it properly. Lower pot is just for more precise and easier setting.
Judging from your readings, you have higher current consumption than we should expect in this circuit. Making less voltage  drop with the zener you should be able to bias properly FET in the correct range. Since we expect more voltage flowing around then more precise pot will be also usefull If it will work, please write your readings in the same points as previously

Thanks a lot for these explanations. I got a 33V Zener and 10k trimmer from a local store (they didn't have anything between 10k and 25k) and will try what you proposed. For re-biasinig the capsule needs to be disconnected, right?

One thing that I don't understand is that with the 33V Zener we are deviating from poctops BOM. But other users got proper readings with 24V Zeners, so I suppose there must be something else wrong in my build, shouldn't it?

 

[ln76d]

Try 10k, but it can be to low. Usual i set bias properly with 12k pot.
You have only two active components inside the circuit - FET transistor and zener diode.
It looks like your FET is taking more current than in others builds, but you are not the only one, some guys also had this problem and bias was set in lower range (same with km84 circuit builds). From your voltage readings  looks like everything is fine - there's no any huge deviation - just looks like higher current.
If you are using calibration input (i don't know are the Poctop boards have it) then you don't have to disconnect capsule. Anyway - if it have calibration input or not - i always use signal applied direct into the input via 50-100pF capacitor. You don't need to disconnect polarisation voltage from the capsule, just gate connection.
In fact for U87 circuits which am building from the different donors, i prefer 33V zener (as in original U87 circuit) - it's really easier to set the bias in correct range.

 

[Neutrino]

In fact for U87 circuits which am building from the different donors, i prefer 33V zener (as in original U87 circuit) - it's really easier to set the bias in correct range.

OK, I did some testing and here are the results.

First I thought I try replacing the jFET as you initially suggested. Originally, I had one with an IDSS of 6.43mA and I replaced it with one with IDSS of 7.64 mA. Biased again and no big change:

phantom voltage: 48.0V
-both sides of R14: 20,77 V and 23,31 V
-both sides of R2: 44,7 V and 44,1 V
-c9/c7 point: 8,68 V
-R19/R18 point: 47,5 V

Then I replaced the 24V Zener with a 33V one, biased and here is what I got:
phantom: 48.4V
-both sides of R14: 26.24V and 29.54
-both sides of R2: 44.5V and 45.1V
-c9/c7 point: 10.64
-R19/R18 point: 48.0V

So the Drain voltage is now in the expected range, although the other voltages are still on the low side.
I finally did not replace the trim pot since I felt the 25k one was not really problematic to finely tune (it is a multi-revolution trim pot).

What I found, however, is that the soundcard oscilloscope can be amiguous at times: Depending on how high I set the input level at the sound card input, I end up with drain voltages between 8V and 8.7V after the biasing procedure (that is for the 24V Zener). I have attached the final screenshot after biasing.

 

[ln76d]

OK, I did some testing and here are the results.

First I thought I try replacing the jFET as you initially suggested. Originally, I had one with an IDSS of 6.43mA and I replaced it with one with IDSS of 7.64 mA. Biased again and no big change:

phantom voltage: 48.0V
-both sides of R14: 20,77 V and 23,31 V
-both sides of R2: 44,7 V and 44,1 V
-c9/c7 point: 8,68 V
-R19/R18 point: 47,5 V

Then I replaced the 24V Zener with a 33V one, biased and here is what I got:
phantom: 48.4V
-both sides of R14: 26.24V and 29.54
-both sides of R2: 44.5V and 45.1V
-c9/c7 point: 10.64
-R19/R18 point: 48.0V

So the Drain voltage is now in the expected range, although the other voltages are still on the low side.
I finally did not replace the trim pot since I felt the 25k one was not really problematic to finely tune (it is a multi-revolution trim pot).

What I found, however, is that the soundcard oscilloscope can be amiguous at times: Depending on how high I set the input level at the sound card input, I end up with drain voltages between 8V and 8.7V after the biasing procedure (that is for the 24V Zener). I have attached the final screenshot after biasing.

Now looks much better All voltage values are pretty decent. Due to higher current you have a little bit lower polarisation voltage, but still it should work better.
Let us know is there still any problem with distortions during recordings

 

[Neutrino]

Thanks a lot for your help!!! I really appreciate it. I will do some testing regarding distortion and let you know.

My understanding is that when the FET opterates at a higher drain voltage that should increase the Dynamic Range, right?

 

[Colorblind]

Hey Guys,

Looking to build a couple of these using Poctop's boards.  The 10uf electrolytic cap is on backorder from Mouser until mid-December.  Can anyone tell me if this will work as a substitute? 

https://ca.mouser.com/ProductDetail/Vishay-Sprague/516D106M063JL6AE3/?qs=sGAEpiMZZMtZ1n0r9vR22fPWwtj8kO8a7hrPz8OzbVg%3d

Thanks,

C

 

[ln76d]

My understanding is that when the FET opterates at a higher drain voltage that should increase the Dynamic Range, right?

Yes. More headroom.

Looking to build a couple of these using Poctop's boards.  The 10uf electrolytic cap is on backorder from Mouser until mid-December.  Can anyone tell me if this will work as a substitute? 
https://ca.mouser.com/ProductDetail/Vishay-Sprague/516D106M063JL6AE3/?qs=sGAEpiMZZMtZ1n0r9vR22fPWwtj8kO8a7hrPz8OzbVg%3d

Yes, it will be. 

 

[Colorblind]

Yes. More headroom.

Yes, it will be.

Thanks!

 

[Colorblind]

Regarding R18 & R19, I've read that it's important to match these resistors within 4% of each other.  I've got 20 resistors here (building 2 mics), and they all measure 2.19k according to my DMM.  Is it likely that my DMM is not accurate enough, or is it possible these resistors all are close enough to each other that I can use any of them?  I looked into the Wheatstone Bridge method, but I'm still a bit confused by it. 

Thanks,

C

 

[rockinrob86]

What's your DMM?

It sounds like it's not accurate enough, unless you ordered 1% parts.

 

[Colorblind]

What's your DMM?

It sounds like it's not accurate enough, unless you ordered 1% parts.

The DMM is this one:  https://www.cryptominer.ca/index.php/product/mastercraft-deluxe-digital-volt-meter/

The resistors are 1%.  I just used the Mouser cart link/BOM that Poctop provided. 

 

[ln76d]

The DMM is this one:  https://www.cryptominer.ca/index.php/product/mastercraft-deluxe-digital-volt-meter/

The resistors are 1%.  I just used the Mouser cart link/BOM that Poctop provided.

Then don't bother and use these resistors - it will work

 

[Colorblind]

Then don't bother and use these resistors - it will work

Ok, thanks!  The weird thing is, the mouser BOM still calls ten of these per mic kit, even though they're 1% tolerance and only two of this value are needed to complete the build. 

 

[TillM]

Dan used to do it, that you can match them with a tolerance of 0,4%. (look also at the real Neumann u87 schematic)
For my first U87 build I don't do it. My second U87 has resistors that match.
I don't hear any difference or have performance problems.

 

[Colorblind]

Dan used to do it, that you can match them with a tolerance of 0,4%. (look also at the real Neumann u87 schematic)
For my first U87 build I don't do it. My second U87 has resistors that match.
I don't hear any difference or have performance problems.

Ok, that's good to know. 

Thanks again, y'all!