Maker Pro
Maker Pro

Simple MIC Buffer Design Question

oneoldude

Jun 16, 2013
14
Joined
Jun 16, 2013
Messages
14
This is an unbalanced mic buffer for measurement mic use. It will be built into the mic wand and allow easy swapping of ECM capsules for loudspeaker testing and capsule testing and provide for longer cables than the ECM capsule alone.

The output will go to an unbalanced mic pre with switchable gain and from there to unbalanced sound card line in. All inputs and outputs in the chain are unbalanced.

I posted my first try in another forum. See post #37 here: http://www.diyaudio.com/forums/solid...quested-4.html

After a lot of reading and outlining the discussion of jan.didden and Amblifier in the thread above, I think I have come up with an improved circuit below.

Here are my calculations. Pardon the non-conventional presentation.

Ic = .0045 A = 4.5 mA (Should swing a bit less than 4.5V P-P into 1k. Input will probably never see more than 1V P-P [but you never know!] and load should be well above 1K)

Re = 9V/.0045 A = 2000 Ohm = 2k

2N5210 hfe = 250

Ib = .0045/250 = .000018 A = .018 mA

Bias string I = (10)(.000018 A) = .00018 A = .18 mA

Bias string Rb = 9V/.00018 A = 50000 Ohm = 50k

Rb2 = (4.5V+.65V)/.000018 A = 28611 Ohm = 28.7k

Rb1 = 50000 Ohm - 28611 Ohm = 21389 Ohm or 21.5k or Rb1a = 5k and Rb1b = 16.5k

MIC bias resistance 9k or Rm1a = 3.01k and Rm1b = 6.04k

RL and CL are loads for the sim. PS is either 9V battery or 9V regulated supply located at the mic pre about 40 feet from the buffer circuit.

Can C4 and or C5 be left out since C3 is very close to the rest of the circuit?

Any help or advice on improving this unbalanced buffer will be appreciated. The idea is to be able to get it to work well and to get it into a mic wand or connector.

Here is the circuit and its sim from LTS.

Thanks

NPN%2520Mic%2520Buffer%2520V%25202.PNG
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
It looks pretty good to me.

I might combine Rm1a and Rb1a, and C4 and C5, and take Rm1b and Rb1b from the same node.

Also I would check the DC voltage on the microphone and adjust the microphone bias resistor if necessary to get the recommended DC voltage.
 

oneoldude

Jun 16, 2013
14
Joined
Jun 16, 2013
Messages
14
I might combine Rm1a and Rb1a, and C4 and C5, and take Rm1b and Rb1b from the same node.

Super idea!

Of course the idea is to keep noise from PS from the front end of the circuit. Will combining them cause a problem? And is/are the filter/s actually necessary when there is a cap (C3) very nearby to provide low AC impedance after the cables? All of the circuitry to the left of the load will be in the mic wand.

Note that C3 could be used with a resistor as a pi section as well. If it did the equivalent job of the of the pi sections created by C4 and Rm1a and C5 and Rb1a or combined as one pi section as you suggest, then I could eliminate C4, C5, Rm1a and Rb1a (and properly adjust the the values of Rm1b and Rb1b). So, adding a resistor to the right of C3 could reduce my parts count by 3.

What I do not understand is whether a pi section placed on the right side of the collector would be as effective as the pi sections placed on the left side of the collector due to the intervention of the collector current which is 4.5 mA.. This could be a problem given the very low signal levels that will appear at the input of the circuit. Input voltages could be as low or lower than 1 mV RMS and peaks could be as high as 1 V RMS. Thoughts?

Also I would check the DC voltage on the microphone...

Typical ECMs (and the ones I am using) are rated at 10V max though they will run fine with higher voltages. It depends on the capsule's internal FET. Most data sheets show app notes with voltages from 2.2 to 5 for computer and cell phone use. But I understand that the capsules are more linear and have greater headroom with higher voltages. So 9V is within the spec no matter what. Bias resistor should be min 2.2k. But higher voltage and resistance is supposed to lead to greater linearity and better headroom. Typical use for hi qual mics is around 9k for nine volts, 12k for 12V and 15k for 15V. That is my understanding but I am not engineer. YMMV.

Thank you.
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
If the circuit is powered from a 9V battery that is not used for any other function, you don't really need the separate node for the input stage bias. If it's powered from a noisy supply, I would add a series resistor in the positive supply feed to the right of C3, as you suggested. Whether or not it's worth keeping the smoothed input node depends on how noisy the power source is, I guess.

The reason that a pi filter on the left of the transistor is more effective than one on the right is simply that one on the left can have a higher resistance without significant voltage drop, because of the lower current in the microphone supply circuit.

As you say, a higher DC voltage on the electret condenser mic will give better performance and wider dynamic range, and a higher impedance would be helpful too. You could use some kind of active circuit to supply the mic current, providing a fixed DC resistance with a very high AC impedance - something like a gyrator. See http://www.epanorama.net/documents/telecom/gyrator.html (which is linked from the Wikipedia page, http://en.wikipedia.org/wiki/Gyrator). I don't think electret microphones are considered worthy of such careful treatment though.
 

oneoldude

Jun 16, 2013
14
Joined
Jun 16, 2013
Messages
14
Kris,

If the circuit is powered from a 9V battery that is not used for any other function, you don't really need the separate node for the input stage bias. If it's powered from a noisy supply, I would add a series resistor in the positive supply feed to the right of C3, as you suggested. Whether or not it's worth keeping the smoothed input node depends on how noisy the power source is, I guess.

Thanks for the explanation. Many have said that the filter on the mic capsule is an absolute must to avoid motorboating when the battery weakens. I have never experienced such motorboating even with very old batteries. If I did, I would replace the battery.

The resistor to the right of the circuit could be made to do double duty by putting an LED in series with it. The LED would indicated the circuit was engaged and might indicate battery weakness as well. I would cut current thru the LED to a minimum ( but enough to power the circuit proper) so as to extend battery life and I don't like bright LEDs anyway. Will an LED introduce noise on the supply line? I don't know if they are noisy or not or whether the inclusion might worsen or improve noise filtration. Of course the supply voltage might have to be changed.

The reason that a pi filter on the left of the transistor is more effective than one on the right is simply that one on the left can have a higher resistance without significant voltage drop, because of the lower current in the microphone supply circuit.

Stellar! I have been wrestling with that for a while now. I knew of the lesser current on that side, but could not wrap my head around the reality. How simple things become when you know what you are about. Thank you.

As you say, a higher DC voltage on the electret condenser mic will give better performance and wider dynamic range, and a higher impedance would be helpful too. You could use some kind of active circuit .............. I don't think electret microphones are considered worthy of such careful treatment though.

Amen! Often simple circuits are extended as a design exercise by the designer to show his prowess. Things are often brought to extremes of complexity that my not be necessary. I believe in the axiom, "Just good enough is perfect". That is what I am trying to achieve with this little circuit.

What started me thinking is the M31 mic by LinearX. It is described on four tiny pages here:

http://www.linearx.com/products/microphones/m31/M31_1.htm

It seems to have a buffer and not much more at the mic end and can be powered by a 9V battery or regulated wall wart. See the one page PDF on the bottom of the last page describing the M31.

As a pleasant surprise it seems my design draws about as much current as they spec for the M31 so I figure I am pretty close to what they are doing. They would not tell me, I asked.

The beauty of this system, if it works as well as I hope, is that it will allow any hobbyist like me to have a competent and simple measurement system for very low cost even if you had your capsule calibrated or if you bought a calibrated capsule. Of course I have a calibrated capsule. But if I did not, one can be bought for $16 US. Not bad at all.

BTW, if you can make any suggestions to improve my simpleton's circuit, I am all ears.

I hope I have not bored you to death with my questions. But of all the folks I have run into on the net, your are the most pleasant, direct, helpful and understanding guru I have met. You are the kind of person the net was designed for.

I thank you and salute you for your patience, understanding and kind help. Kudos!
 
Last edited:

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Thanks for your kind comments :)

That mic looks like a good solution for your needs. But you should investigate how the response compensation is performed. Presumably the correction will need to be done in software.

The question is: are the compensation parameters specified by the manufacturer for that model of microphone, or do you have to test the actual microphone you are using to determine its response, then apply the compensation using the values you measured?

If it's done on a per-microphone basis, you will need to have, or at least borrow or hire, a sound generation system whose characteristics are extremely accurately known or controlled. Any change you make to how the microphone is mounted or supported may affect the response, especially at frequency extremes. You might want to buy several microphones and calibrate them at the same time, to avoid hiring the calibration system multiple times.

But generally, I think you understand the situation well enough to go from here.
 

oneoldude

Jun 16, 2013
14
Joined
Jun 16, 2013
Messages
14
Yes the mic cal files are called various things .cal or .dat or .frd. They are all text files that include freq, level and phase points. And yes, any competent measurement software can use them.

My cal'd mic and the capsule that can be bought are calibrated on a 0 deg incidence. That is pointing directly at the source. That happens to be the way I use them.

Thanks
 
Last edited:
Top