Mic Biasing

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

fredwreck

Member
Joined
Dec 21, 2015
Messages
12
Im trying to Bias a U87 clone im building from Mic&Mod. 2 questions.

1.  Should I disconnect the capsule before I bias with a tone generator and oscilloscope.  And if so, why do I need to disconnect the capsule wires?

2.  When sending tone.  How many volts of sine wave tone should I be sending from my Function Generator to the mic when Biasing.

Thanks
 
There is no need to disconnect the capsule.
If you use the test input or "Messeingang" (that is in parallel with the 560 ohm resistor), start with 100 mV  1 KHz and slowly turn up the level of the generator until you get noticable distortion.
Adjust the source resistor for minimum distortion, raise the level of the generator again a bit until you reach a point where no improvement is possible.
Take care that the distortion is caused by the microphone electronics and that you don't overload the input of a microphone preamp!
You may find a different DC voltage than 10,5 or 11 Volt on the drain. The DC voltage is just a guideline.
 
It might make things easier if you can open up a DAW and use an oscilloscope plugin on a channel assigned to the input where the mic is connected. If you take this route, make sure to also keep an eye on the input level and reduce the preamp gain so you don't clip the input / ADC ;)

The idea is to start with "a" signal amplitude, raise it until either the top or the bottom of the sine gets clipped, adjust the bias to "unclip" it, then rinse and repeat until both the top and bottom of the test-sine get clipped at the same time, roughly by the same amount.
 
Guys, since it's mentioned before by Abby Road Denfer that not all of FETs are meant to operate with 0v at Gate, should additional resistor be used from source to ground like with mxl2001, Perceptions, etc depending on FET for gain tweaking?

I am dealing with terribly biased st51 and cant make up my mind if i should keep stock arangement, or if i should i go with u87 arrangement.

I'm getting wide range  of results, and i believe it has to do with how much signal is coming from  capsule which is kind of hard to figure out.

Quote:

abbey road d enfer said:
That's the issue here. If you think you are abetter designer than teh AKG/Chinese guys, you must sustain that with some theory. Ther is "some" science involved there.

  Such as?

The FET in the U87 is a selected type, that can operate with 0V gate voltage. The FET in the P220 is very likely not capable of operating satisfactorily at 0V gate voltage (if it was, I doubt the designers would have added a resistor that increases the BOM).

Is it? 2SK30 is a very different animal than the one in the U87. If it was used with 0V Vgs, it would draw 3mA, which is not correct for this circuit. That's why teh designers had to add the bottom resistor.

What evidence do you have that it's not properly biased?

In heh absence of a proven schematic, I can only refer to the fragmented schemo that circulates on the internet, and I see that the polarization resistor is between teh capsule and the coupling cap. If you know different, please post a schemo.
 
Well, unlike most(?) other transistors, JFETs work a bit differently.

They need a negative voltage relative to the voltage on the source pin to turn off, the Vgs(off) or pinch-off voltage in the datasheet.
With the gate connected directly to the source (Vgs=0v), they've often been used as constant-current sources (even inside opamps) - that's what the Idss parameter is all about, the zero-gate-voltage current.


That being said, we really don't care if the gate's pulled to ground (0v), or to +20v or whatever - we only care about the gate-to-source voltage.

And as far as i recall, "even" the U87 has a source resistor for the JFET, adjusted / selected individually for each one. Soooo what "additional resistor [...] from source to ground" do you mean?

The only JFET mic circuit not using a source resistor that i can recall, is the stock one in the BM800.

Got a schematic for that ST51 circuitry?
kingkorg said:
Guys, since it's mentioned before by Abby Road Denfer that not all of FETs are meant to operate with 0v at Gate, should additional resistor be used from source to ground like with mxl2001, Perceptions, etc depending on FET for gain tweaking?

I am dealing with terribly biased st51 and cant make up my mind if i should keep stock arangement, or if i should i go with u87 arrangement.

I'm getting wide range  of results, and i believe it has to do with how much signal is coming from  capsule which is kind of hard to figure out.
 
Well it's basically image to the left vs image to the right.

Abbey Road D Enfer sugests that the FET has to be selected for arrangrment to the right like u87.

ST51 is almost exactly the same as GT55.  However they never changed component values, while they changed FETs, PNP many times, so voltages and current values are all over the place.

Simple biasing with injecting the signal doesnt work, as i need to adjust Drain and ''gain'' resistor as well - stock 47k which gives 4v at drain - low output, high noise with 2sk30 which wasnt the original FET used with this circuit.

Capsule is SDC (dont ask why) so i need more gain, which i get with u87 arrangement and 20k drain resistor, but then Abbey says i need a selected FET that can work in this arrangement. He seems to be right, because i manage to get maximum swing, symmetrical wave shape, high output, but then i have constant 3% THD at ALMOST ANY level injected starting from 100mV!

If i bias it the way it's usually recommended i get low THD, but S/N is, needless to say, not optimal.

All of this i believe relates to original post here, as it's the question of what FET is used and what capsule is used.

The mic really sounds great, but now i am chasing my tail wondering what is the absolut best solution for most gain, lowest THD, and largest S/N ratio.

And i think the offered usual biasing solutions with signal injecting are a bit over simlified. I just think there is much more to it. And i'm not even trying to go into de emphasis network rabbit hole xD

How does one know what signal level is produced by capsule in order to know how much gain it needs, and then adjust for lowest THD at that level?

Here\s GT55 schemo
https://groupdiy.com/index.php?action=dlattach;topic=63638.0;attach=44796
 

Attachments

  • Bias.png
    Bias.png
    5.1 KB · Views: 33
Well, regarding the version on the left, would a Schoeps-ish bias arrangement help, possibly?
Ie. a single source resistor, but a 1meg trimmer in parallel with it, and the wiper going to the 1G fron the gate.

Granted, that doesn't simplify things, but you'd have two variables to mess with. I might even go as far as having a fixed source resistor, and use a 50-100k trimmer for the drain resistor. The interesting trick then becomes finding the "right" balance between the two...
 
Now that's a great idea, i'm doing that next. Thanks!

BTW, these are the values around FET that gave me crazy low noise results, and best swing but very rich harmonic content.

Disclaimer:
DONT USE THIS IMAGE FOR ANY MIC MOD, I AM EXPERIMENTING WITH VALUES AND COMPONENTS!
 

Attachments

  • Spice.jpg
    Spice.jpg
    494.4 KB · Views: 49
I got that idea from some JFET guitar preamps i've been messing around with, it's pretty common to use 100k trimmers for drain resistors, to be able to bias each stage (properly?).

kingkorg said:
Now that's a great idea, i'm doing that next. Thanks!

Nooooot too sure how relevant those values are, though. JFETs are not tight-tolerance devices at the best of times (even within the same part number), so... If these work for this batch, they might not for another batch of 2SK30A's. Just sayin' :)

kingkorg said:
BTW, these are the values around FET that gave me crazy low noise results, and best swing but very rich harmonic content.
 
RuudNL said:
There is no need to disconnect the capsule.
If you use the test input or "Messeingang" (that is in parallel with the 560 ohm resistor), start with 100 mV  1 KHz and slowly turn up the level of the generator until you get noticable distortion.
Adjust the source resistor for minimum distortion, raise the level of the generator again a bit until you reach a point where no improvement is possible.
Take care that the distortion is caused by the microphone electronics and that you don't overload the input of a microphone preamp!
You may find a different DC voltage than 10,5 or 11 Volt on the drain. The DC voltage is just a guideline.

Thank you for the reply!
 
found a data sheet for the 30a from a company in China. I should check if it different from the Toshiba one
https://drive.google.com/file/d/0B-uz2Yms52D7eWNHT2VsNG5EZ2s/edit
It would be interesting to curve trace the different 30a JFETs from China and Toshiba etc.

The use of two resistors in  the source leg looks to me as a way to control the gain of the first stage.

If you look in your ST51s the unbypassed one are often the same or close to the same value. The bypassed one is for the biasing.

For AC the resistance value of the bypassed resistor becomes small
 
Thanks, i highly doubt it's the Toshiba one in the mic.

As the mic in question is meant for low level signals, recording at distance i went for maximum possible gain and S/N ratio. THD and high SPL weren't the priority here.

I went for u87 arrangement, no bottom resistor, used pots on both source and drain to get best results. Capsule is polarized at 80V. The mic has crazy high output.

The mic beats easily KSM44a regarding noise performance, and it's a SDC! Needless to say, it's not for close miking.
 
I adjusted  one of my 51s to the single bypassed source resistor at the JFET stage.
With the stock JFET I ended up with about 10K at the source with the stock 47K drain.
One of a few 51 circuit adjustments. I will probably change the JFET.

An issue with adjusting both the source and drain is the changing output resistance. With the PNP follower and the 1meg biasing resistor this does not matter as much as one JFET to transformer type circuits.
 
kingkorg said:
Thanks, i got almost the same result, exception being 56K on drain. Awesome!
The issue there is that gain decreases as the resistor increases. With 56k in the source and 47k in the drain, gain is less than unity. Not good for global noise performance and headroom. An FET that needs such a high source value is not good for the task. Biasing via a tap on the source resistor will not help either.
 
I guess i wasnt clear. Ended up with 10K at source, 56K at drain. Those values gave me maximum gain, but gate voltage is near zero, which you stated is not optimal for this FET, nor one source resistor like with u87. Or did i get that wrong?
 
kingkorg said:
I guess i wasnt clear. Ended up with 10K at source, 56K at drain.
Right, my mistake. Seems pretty normal.

Those values gave me maximum gain, but source voltage is near zero,
What is "near-zero"? With typically 0.4mA, the source voltage should be about 4V. If you have much less, current is too low, which impairs headroom.

which you stated is not optimal for this FET,
Did I?

nor one source resistor like with u87. Or did i get that wrong?
How did you came to conclude 10k was the right value?
[/quote]
 
abbey road d enfer said:
What is "near-zero"? With typically 0.4mA, the source voltage should be about 4V. If you have much less, current is too low, which impairs headroom.

Sorry, i just woke up when i replied. I was talking about GATE voltage which gets near zero when i use one resistor source to ground.

I got 10K at source by tweaking pot for maximum gain and headroom. I was tweaking drain pot at the same time. It's Sterling Audio ST51, not u87. Similar, but not the same as u87.

I am quoting you saying 2sk30 should have bottom resistor:


abbey road d enfer said:
Is it? 2SK30 is a very different animal than the one in the U87. If it was used with 0V Vgs, it would draw 3mA, which is not correct for this circuit. That's why teh designers had to add the bottom resistor.
 
Well, with the 1G resistor pulling the gate to ground, you would kinda-sorta expect there to be 0v on the gate ;)

Regardless of what the source voltage is... Unless, of course, it's connected in the middle of a two-part source resistor.

kingkorg said:
Sorry, i just woke up when i replied. I was talking about GATE voltage which gets near zero when i use one resistor source to ground.
 

Latest posts

Back
Top