AKG Perception P220 to Neumann u87 5 min mod ( p200, p100, p400, p420? )

Help Support GroupDIY:

James-F

Member
Joined
Nov 16, 2019
Messages
17
High output from the same SPL and improved SNR is not a bad thing at all, unless you conduct an interview with a jet engine as Julian Krause nicely puts it.

Anyone knows what is the typical AC voltage that a 34mm capsule outputs at normal 60-80db conversation volume, around 50mV?
I need a rough estimation for LTSpice.
 

kingkorg

Well-known member
Joined
Apr 15, 2017
Messages
1,321
Location
Norway
Well, mics do get used on drums for example where levels at transients could potentially exceed that of a jet engine. I use condensers on snares a lot. I believe i remember someone once said they use this arrangement to be able wide range of fets.

When it comes to the capsule output, that can vary a lot depending on capsule size, polarization voltage, electret.
Even for spoken word if plosives get through you could be seeing several volts!

To test this you could make a buffer circuit with a fet input opamp and scope the output.
 
Last edited:

James-F

Member
Joined
Nov 16, 2019
Messages
17
I've simulated an entire MXL V67 mic in LTSpice and from what I see this additional tail resistor might be useful to calibrate how the feedback EQ behaves by altering the gain of that stage, but for that we need a reference from a real U87.
For example when feeding a U87 FET with 0.1vAC peak, what is the vAC peak at the drain (measuring through coupling caps of course)?
Once we know that we can calibrate any FET+PNP 2:1 mic from the East to have the same feedback behavior.
Maybe Neumann still using the 2N3819 to retain the correct gm value?
 

James-F

Member
Joined
Nov 16, 2019
Messages
17
Images;
1. Using the 2n3819 as reference in simulation I can see the U87 eq curve with -6db at 10kHz and -12db at 20khz, so that's a healthy rolloff.
2. Changing to LSK170B and biasing the rolloff is far less effective with -3db at 10khz and -6 at 20khz, I compensated capsule input voltage to hit 0db.
3. Adding a 3.9k tail resistor the mic loses -10db of volume and has barely any feedback EQ at all.

So it is all about the gain of the FET stage, I think we need to calibrate the tail resistor to match a U87 fet gain, because just removing it may have too much or too little effect on the feedback.
 

Attachments

  • 1.png
    1.png
    270.5 KB · Views: 24
  • 2.png
    2.png
    262 KB · Views: 22
  • 3.png
    3.png
    208.5 KB · Views: 23

kingkorg

Well-known member
Joined
Apr 15, 2017
Messages
1,321
Location
Norway
To be honest i don't see the point. The result will depend on the response of the capsule. No point in matching the circuit if the capsule behaves differently. The way i do it is since i have the acoustic measurement of u87 i bias the circuit for lowest thd, and adjust feedback network to give me the desired acoustical response of the whole mic.
 

ccaudle

Well-known member
Joined
Mar 18, 2010
Messages
125
Location
Houston
Why reduce gain in that stage?
In general the reason you use feedback to reduce gain in a stage is to reduce distortion, and to reduce variation of gain due to device-to-device variation. That applies to all devices, but JFET devices especially vary in gain and needed source to gate voltage for correct biasing, so you either test each JFET and throw most away, test each JFET and adjust component values on a device-by-device basis, or you design a circuit which is not as sensitive to device variations. AKG chose the last option, which is the only reasonable option for a mass produced device.
As pointed out previously that also gives the option to bootstrap the gate resistor (i.e. positive feedback to that resistor), which increases the effective impedance seen at the gate resistor to be larger than the 1G Ohm DC value of that connection. That reduces low frequency roll off from the capsule. There may be a distortion reduction at low frequencies as well, but I don't know how large that effect would be.
 

Latest posts

Top