M31 Preamp

[ronv]

I think it is going to be gain challenged now. But we should agree on some specs.

The card probably is .9 volts peak to peak (. as opposed to 3.4 volts. So should we target .9?
I think Panasonic has the WM61 at -35 dB which I make out to be 35 mv pp at a SPL of 94. so a gain of about 25?
But the gain also depends on the impedance of the mic and the input to the circuit so we need to nail down the power and the resistor to the mic then we can play with the preamp to get the desired results.

 

[WTP Pepper]

A simple inverting opamp cct with diode feedback to soft limit and a gain of about 40dB would suffice for most apps.

 

[audioguru]

A simple inverting opamp cct with diode feedback to soft limit and a gain of about 40dB would suffice for most apps.

The circuit will be used to measure the frequency response of speakers so you do not want any limiting.
The speaker will be close to the mic and be loud so a gain of 40dB is too high.

 

[oneoldude]

Ron,

Yes the WM61A is spec'd at -35 dB. But I calc that at 64 dB SPL the output will be more like 50 mV PP than 35 mV PP.

In any event, I am not building this mic pre for the 61A. I am using the WM64PCT that is spec'd at -44 dB. Data sheet attached.



I have been all over the net trying to get a handle on what the input overload is on soundcard line in. All I get is confusion. The manufacturers do not spec that cause the typical customer has no clue what the spec might be for. There appear to be no standards at all on that parameter. I have emailed M-Audio to see if they even know the spec for their own product. I am waiting for a reply.

So, for now, with no target at the receive end, I propose a thought experiment.

To use a laptop or equivalent for measurement you must use an outboard (USB) sound card cause the internal sound cards are usually crap and sometimes not even duplex which is necessary for proper measurement. Often they do not have line-in so you would have to use mic in which is usually crap and usually mono and not suitable for proper measurement. So we are now talking about a USB stereo duplex sound card with line-in.

We know the USB port provides no more than 5V to the sound card. If we assume the sound card will use all the room 5V provides and that it will accept a rail to rail signal, then we can hypothesize that the max input possible is 5V PP or 2.5 V P or 1.78 V RMS.

Using that as an underlying assumption for this thought experiment, let's compare the output of the WM61A against the WM64PCT and see if we can figure out why Linkwitz did what he did when he set the gain on his preamp at about 10 dB and see if that helps us with the WM64PCT.

The next image is a Table I put together that shows the output of the WM61A and WM64PCT at 1 Pa (94 dB SPL) and at 125 dB SPL. Then I show the result with +10 dB gain a la Linkwitz, then at +20 dB and finally at +18 dB. You will note I left out the 61A after + 10 dB cause it blows up the 5V PP limit at and above +10 dB. Note I used +10 dB for Linkwitz because that is what he calls it . Actually his gain is X3 which is +9.54 dB but 10 is close enough.

On my table I highlighted where Linkwitz crosses the thought experiment peak to peak and where the 64PCT crosses it as well. This suggests to me that the circuit gain should be around 17 dB. Now this does not deal with the reality of sound card overload and presumes Linkwitz understood what he was doing. But its better than nothing.

What do you think?



Last I attach some spec snips I took from the LinearX site re: their mics. The variations are not much different from other companies. It makes me think all the mics are buffered but some are amped and others are not. Hmmm.



Thanks

 

[ronv]

I guess it all depends on what you want to do.

But the more I look at it the more I think the SPL to voltage conversion is peak to peak. So based on this calculator a WM61 would output 17 mv At an SPL of 94. I don't think it will work at 125 because it's amplifier will clip. So that may be what leads to the modifications or a lower sensitivity microphone.
http://www.sengpielaudio.com/calculator-transferfactor.htm
I guess I don't know what your specs are for the range.
Maybe someone else has some input.

 

[oneoldude]

That is a great site. It is the site I used to calculate my table. I did not know they had a sensitivity conversion. It would have saved me a bunch of time. I checked my table with it and got the same numbers I busted my butt for.

On another page on that site they make a point that in audio and recording engineering, AC voltages are always RMS unless identified otherwise. That is why I calculated the way I did.

So it looks like a 17 dB gain will get me where Linkwitz wanted to be. Now where do I want to be?

 

[ronv]

One way to find out.

 

[oneoldude]

That's right. There is only one way to find out. So I looked for my breadboard to hook 'er up and could not find it. Geeeze! So I ordered one up and am waiting for it.

In the meantime, your most recent version is yet another variation. I don't know the math and theory so could you help me out here.

What I have done is use LTS to model the circuit and move things around to see how they work. That is how I came to the emitter resistor degeneration to easily change gain.

I changed the collector resistor to change output impedance for longer cables. When I did that it increased current through the transistor. I don't know how far I can go with it. Is it to the point the transistor frys due to too much current?

I changed the collector to base resistor and found it affected the operating point. So I shot for an operating point about 1/2 of V+. Then I did transient analysis in LTS and moved the operating point around to get equal gain on both polarities of the output signal and try to get the circuit to voltage clip equally on both sides of the output signal by slowly increasing input voltage while I moved the operating point around.

That all makes sense to me. But am I doing the right things?

I still do not know the best current to use through the transistor and where the tradeoff is between current, noise and output impedance (bandwidth). Any thoughts while I am awaiting my breadboard? Or is this simple circuit now about as good as it can be for the purpose?

Thanks

 

[ronv]

Not sure what changes you are talking about.

I just moved the signal source so it would have the 5K in series like the actual mic. This gets the correct input impedance.
I showed the little filter for the mic voltage and added a filter cap to the 9 volts. Is it 9 volts?
I then needed a bit more gain to get close to your numbers so I made the emitter resistor a little smaller.
I don't see any advantage to going lower on the collector resistor. It can easily drive the cable.
And yes, I think you did the right things.

 

[oneoldude]

Not sure what changes you are talking about.

I just moved the signal source so it would have the 5K in series like the actual mic. This gets the correct input impedance.

Yes that is what I was talking about.

Yes it is 9V.

And thanks for the clarification.

Now I have to wait for the mail. I'll let you know what happens when I breadboard it.

BTW, I got a reply from M-Audio. Turns out that M-Audio was sold and the new owner knows nothing at all about any products the company made before they bot it. How's that for cutting ties to your past?

Thanks

 

[oneoldude]

Thanks for all the input guys.

I did build a couple mic preamps and decided to go with the Linkwitz version with variable gain.

Here is the result: