Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Paired resistor for balanced microphone circuit

Status
Not open for further replies.

Willen

Well-Known Member
Hi,
It's said that in a balanced audio circuit, both side of components should be paird/matched to supress noise perfectly. I am getting many circuit of MXL Condenser Microphone around internet. Here is circuit attached which has been improved in right side, and original is in left; taken from here: https://www.pa0nhc.nl/AudioTechniek/mxl990_sp1mods/mods_en.htm

Look at the R6 and R7, these are phase splitter FET bias resistor around drain and source. All people recommend to use paired 2k2 in both side which means these resistors will have equal (pretty close) resistance. BUT, look at the R13 and R14 which are in series and are connected as parallel with R6 (one of the matched pair of 2k2). Because of the parallel additional resistor in the source, the value of paird 2k2 will decrease overall. Look at the modified circuit, modifier added even lower value resistor which will decrease the value of paired resistor (ok, total resistance in source) even more. So, What is the worth of pairing these resistors in this case? Or what you suggest? Please!
 

Attachments

  • Modified schema SP-1 = MXL990.GIF
    Modified schema SP-1 = MXL990.GIF
    61.3 KB · Views: 276
Last edited:
The parallel resistors are over a hundred times higher value than the 2k2, so even using 1% resistors the end result is likely not changed more than the possible difference between any two 1% resistors.

It's nothing to worry about, with those proportions.

If you were manually matching resistors to try for perfect balance, you could add another high value one across the upper 2k2 to get an exact value, but I doubt you would be able to detect any difference in performance.

The end result is also dependant on the input balance and common mode rejection of whatever the mic is connected to.
That may be ultra-precise or very crude, from one device to another.
 
The extreme balancing is needed since the thyristor stage lighting dimmers spray buzzing all over the place.
I am glad to see that the mods also improved the max sound level and low frequency response.
 
Hm... then while matching these 2k2 resistors, upper one I need to choose around 0.5% higher than the lower one so values of both resistors will close little bit now.
 
You can add the same or similar resistors to the upper branch to get matched total resistance.
 
The parallel resistors are over a hundred times higher value than the 2k2, so even using 1% resistors the end result is likely not changed more than the possible difference between any two 1% resistors.

It's nothing to worry about, with those proportions.

If you were manually matching resistors to try for perfect balance, you could add another high value one across the upper 2k2 to get an exact value, but I doubt you would be able to detect any difference in performance.

The end result is also dependant on the input balance and common mode rejection of whatever the mic is connected to.
That may be ultra-precise or very crude, from one device to another.
Ok, however I will measure components value once simply to get little better couple. It makes me feel good at least :)

The extreme balancing is needed since the thyristor stage lighting dimmers spray buzzing all over the place.
I am glad to see that the mods also improved the max sound level and low frequency response.
Ok, If I tried to match transistors to cancel those probable buzz from XLR cable then what is better way to match: Matching BE voltage or matching hFE?
 
Matched transistors have their hFE matched.
Transistors in an IC have hFE and Vbe matched.

I never used a mic cable near thyristor dimmer cables so I never had buzzing.
 
One way to achieve perfect balance is to replace one of the 2.2K resistors with a 2.1K resistor and a 200 ohm trim pot in series.

It looks like the reason for R13 and R14 is to bootstrap the FET input for a higher impedance as seen by the microphone element. The 1 G resistor seems a bit large.

ak
 
The microphone is a "condenser" one. The very small capacitance of its vibrating membrane is in a voltage divider with a fixed small capacitance. Therefore an extremely high input impedance is needed for the Jfet. Electret mics have the 1G resistor built into the mic with the Jfet.
 
I haved hardly matched two transistors which have very close Vbe now. Their hFE are 480 and 488. In this application which is more important: Vbe or hFE?

Audioguru, I think months ago, maybe you said those PNP on the circuit has gain of 1. So I guessed matching Vbe is more important than matching hFE.
 
Last edited:
The extreme balancing is needed since ……….
I have used great care to balance the receiving side. (Diff amp or transformer; front ends) This removes much of the differential noise.
On the transmitting end, like this circuit, I see less need. If I designed it (and price was of no count) then there would be no DC offset and "signal" and "not-signal" would be the same but inverted.

I do not like it but often on the transmitting end there is "ground or shield", "signal", "signal-ground". So the signal is single ended. The receiver end still gets the audio and removes the noise. Yes I know there is 1/2 the signal.

My question is: So what if one wire has a gain of 1 while the other has a gain of -0.9? It is the receiving side where the noise if removed. That is where I put the 0.1% resistors.
 
My question is: So what if one wire has a gain of 1 while the other has a gain of -0.9? It is the receiving side where the noise if removed.
*Some* of the noise. And then more noise is added by the downstream electronics. If one signal line is low compared to the other, then the inverted-and-summed signal is lower than it could be for the same headroom. All added downstream noise is a constant, so the final output has a lower s/n ratio than possible with a correctly balanced signal pair.

ak
 
The AudioGuru was talking about noise picked up on long wire. I have 50 and 100 feet "snakes" that run from the stage back to me. There is 60hz hum and high frequency trash from light dimmers. (some times radio station noise) The noise from the wires is much worse then the noise floor in the amplifiers. I have run wires for a city block from a studio to the transmitter. (FM radio) In the last case I had 10Vp-p noise when the heating and cooling relays engaged. I have seen 2V of 100mhz from the transmitter. One consort hall was so noise with RF that I added one inch ferrite beads on each end of every cable and added diff-beads to all the connectors in the mixer.

The amp in the mic has a 25V supply so head room is not a problem. I think boosting the gain 2x or running diff will have the same 2x S/N improvement over single ended. (assuming no noise from wire)
It's said that in a balanced audio circuit, both side of components should be paird/matched to supress noise perfectly.
I agree "perfectly balanced" is needed on the receiving end. But "mostly balanced" vs "perfectly balanced" is of little effect on the mic end.
My mic preamps are diff. on the input and after the first amplifier they are single ended for the effects and mixing. The signal is not diff until the last amplifier(s) where it drives the output connector.
 
The Jfet with its source and drain resistors balances the circuit. The PNP transistors are simply emitter-followers with a voltage gain of 1 but an hFE current gain of hundreds. The Vbe of an emitter follower does not matter but its hFE affects its output impedance that should be balanced.

The electronic parts produce hiss noise that balancing does not change. Balancing reduces noise picked up by the connecting cable that has a shield to block most nearby noise anyway.
 
The Jfet with its source and drain resistors balances the circuit. The PNP transistors are simply emitter-followers with a voltage gain of 1 but an hFE current gain of hundreds. The Vbe of an emitter follower does not matter but its hFE affects its output impedance that should be balanced.

The electronic parts produce hiss noise that balancing does not change. Balancing reduces noise picked up by the connecting cable that has a shield to block most nearby noise anyway.
I have got tow types of transistors ZTX951 and 2N5087 for audio work. The ZTX951 showing just hFE of 35 on my DMM so I am not going to use ZTX951 which has pretty low hFE, (you are saying the circuit has hFE of 100). I will use 2N5087 which has same hFE of 498 on a couple but has little different Vbe. If Vbe is not more important here then I will not care about it here.
 
The 2N5087 is an excellent low noise, high hFE and fairly high voltage PNP transistor. The original 2N5401 has a higher noise level and much lower hFE.
 
The extreme balancing is needed since the thyristor stage lighting dimmers spray buzzing all over the place.
I am glad to see that the mods also improved the max sound level and low frequency response.
Hi, And do you think the modification lower the gain too? Recently I completed the project but getting little low audio even with full mic preamp volume in the audio interface.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top