M31 Preamp

[oneoldude]

This post replies to criticisms made by the self proclaimed audioguru (Uncle $crooge) re: my attempt to make the rightfully famous Linkwitz pre amp a bit more flexible by adding a switch to allow the MIC capsule to be biased either from ground to V+ or from ground to V- and to provide for a set-and-forget gain adjustment.

Here is the Linkwitz circuit. Scroll down to "Microphone"

http://www.linkwitzlab.com/sys_test.htm


Herr McDuck, you are at it again. All negativity, erroneous criticism and no real constructive help.

Only this time its worse. You intentionally failed to show my entire schematic in order to make an incorrect criticism. The entire schematic is the first post on the one page thread I point to below. If you look at the entire schematic you will see that the 10K resistor you claim is missing is next to the capsules in the top part of the schematic. But you chose not to include that part of the circuit and chose to make an invalid criticism again.

Here is the page:

http://www.htguide.com/forum/showthread.php?41165-Linkwitz-Pre-Amp-modification

Also, as usual, you do not do your homework. The first post in that one page thread was a schematic for discussion. It was not a final. Had you scanned through that one page thread you would have found the as-built schematic toward the bottom of the page. But you did not bother to even scan down one page and chose to give incomplete information to criticize, as usual.

In the as-built, the cap you circled in red (C2) does not even exist. The reason is mentioned on the as-built schematic. That should send you into convulsions. But Linkwitz apparently came to the same conclusion and did not use that coupling cap either. No cap at all will certainly widen the lower bandwidth in that portion of the circuit!

In any event you take the same position for C1. I just used what I have on hand and don't see a problem in having a very wide low end bandwidth for an instrument MIC. In Florida we don't have earthquakes and I don't measure during thunderstorms or even with air conditioning running, etc. This is something I discussed with you in another thread. Hmmmm Did you forget that?

Last, the motorboat filter you describe is not used by Linkwitz either. That is probably due to using a battery supply. It might be a good idea from a belt-and-suspenders point of view. But to do it would require a voltage divider and filter cap on the 10K MIC bias resistor. That lowers the voltage on the MIC and thereby reduces MIC linearity. It is an engineering compromise I do not know how to make or calculate. So for now, I think I will stick with a real pro, Siegfried Linkwitz, rather than incorrect, misleading, nebulous and unhelpful criticisms from you.

If you respond with anything other than a schematic of what you consider a high quality unbalanced circuit for ECM measurement MIC pre amp use that has not already been mentioned here and a description of how to make the engineering compromise entailed in the motorboat filter that you suggest, I will put you on my ignore list and not deal with your diatribes again.

Frank, sorry I misdescribed your filter. I was going from memory and got it wrong. I would still not use your filter for measurement use. I would shoot for a straight line to at least 20 Hz.

Thanks for putting up with this response guys, I think it will be the last re: Herr McDuck aka Uncle $crooge aka audioguru.

 

[ronv]

JFET Amp

Don't be to hard on Audioguru. He really is smart, just cantankerous like a lot of us old guys.

In any case take a look at this little JFET amp. It has a little gain and a lower impedance that the mic alone. It would probably drive your cable and still be ok at 20 KHz.
Since it is only 3 pins you could probably do surface mount if you had to.

 

[audioguru]

It is rare that the 10k resistor that powers or loads an electret mic is mounted at the mic. Usually it is in the preamp circuit.

Linkwitz also missed adding a filter to the DC that powers the mic. The 9V battery voltage will be bouncing up and down when the opamp plays high level bass sounds. Then the opamp amplifies the low frequency bounces and you have an output that motorboats, especially when the battery internal impedance rises as it runs down.

This is the electret mic preamp that I have used for many years. Its 10k resistor that powers the mic has a filter.

 

[oneoldude]

Ron,

Thanks. I will model with it in LSP with the WM61A LTS model I pointed to earlier.

Unfortunately, I have a prejudice against FETS because they often need to be selected to be sure they will work properly, if at all. That is a shame because the unskilled like myself often have trouble using them. FET selection is discussed in the John Conover article I previously pointed to. All but one of his circuits are FET or OPA based. I hope you have read it. I think it is very well done and way over my head.

Yes, BJTs are a bit noisier and don't have huge input impedance, but you can plug in most any gen purpose low noise NPN/PNP into a low noise NPN/PNP circuit and it will at least work and often work very well. That is a nice feature for a hobbyist like myself and others. OPAs are better and even easier to use but sometimes their bulk makes them inappropriate for the specific purpose.

I have tried to model simple and small CC and CE circuits for my purpose. But I am no engineer so all my efforts have failed.

And that brings me to the only BJT circuit shown by John Conover. It is about three quarters down the page under, "Other Interfaces for the Panasonic WM61A" see here:

http://www.johncon.com/john/wm61a/#FigureVIII

Unfortunately it is a P48 circuit not 9V and 15V like I am seeking. It has very low parts count (5) and should easily fit into an XLR3 connector. Heck, depending on what Conover means by "small" re: the zener capacitor, it could possibly fit inside a 1/4" MIC tube. I have modeled it as a P48 circuit with the WM61A model in LTS (both with and without the zener). It seems to have, if anything, too much gain and seems to only barely push about 40 feet of cable. Which is on the margin of "Just good enough is perfect".

I have tried to fiddle with it to get it to 9V and 15V but my ignorance prevents me from being successful. I have 2N5210s and 2N5087s for either polarity of operation. But I have no clue how the circuit works or whether it is even worth consideration re: noise, headroom, distortion and the like. But it sure looks like a real cutie in a tiny BJT bikini.

And if worthy, it might even run with the WM31 measurement MIC this thread is about.

Thanks

 

[ronv]

I think that circuit is just to drop the 48 volts to a level that the WMA61 can tolerate.

I don't know as you would have any problems with the JFET, but you do need to use the one called out.

But, here is a BJT. see what you think of it. It is a low noise transistor, but it has a very wide beta range.

 

[audioguru]

Ronv, we NEVER bias a transistor like that!

Your transistor has severe distortion because it is saturated and its operating point is much too low. Its output level is almost nothing.
Also your transistor circuit is affected a lot by beta range and will be even more distorted if the beta is higher.
I fixed it by using negative feedback. It has a voltage gain of 42 when the source impedance is high at 5k ohms.

 

[ronv]

Is there some difference between my post and your second picture?

 

[ericgibbs]

Is there some difference between my post and your second picture?

hi ron,

The only difference I see is the series input resistor.??

and you are showing that in your signal source as 5K.

E.

 

[ronv]

Thanks Eric,

My eyes aren't what they used to be (actually they were never good) but I didn't see it. Maybe it was that I put the 5K in the signal source or it might have been my crooked line for the bias resistor.

I do that just to check Audioguru's astigmatism.

 

[audioguru]

Ronv, I am sorry if I am wrong. This morning I checked and double-checked that your base bias resistor connected directly to +15V, not to the collector.

I got my astigmatism corrected a few years ago during my cataracts operations and now I see perfectly without glasses (except I use reading glasses for closeups).

 

[oneoldude]

Hello Herr McDuck,

Thanks for the circuit you posted. I like the fact that it only uses one 9V battery. How about a few questions about it?

What do you typically use the pre amp for? Recording, measurements?

How long a cable do you typically run from the MIC to the pre amp?

How long a cable do you typically run from the pre amp to wherever it goes?

Do you use this circuit for MLS/FFT type speaker measurements?

At what gain do you set the pre amp for MLS/FFT if you do measurements?

What is your typical SPL at the mic if you do MLS/FFT measurements?

I note the circuit gain is about 28 dB at the pot/trimmer midpoint. Do you ever overload your sound card line-in with a mid point setting?

Indeed, where do you typically set gain in dB if you do MLS/FFT measurements?

I see that your OPA bias resistors are an order of magnitude larger than Linkwitz talks about and that your feedback pot/trimmer at midpoint is 2.35 times an order of magnitude larger than Linkwitz talks about. He was concerned with noise. Have you had noise problems with this circuit?

 

[oneoldude]

Ron,

Re: your BJT circuit. I incorporated the load resistor and the model for the WM61A and ran it with a load in LTS. I ran it from both 15V and 9V. Interesting. It seems to be a self biasing circuit. Kudos.

Attached is the 9V simulation cause that is the one I will build.

A couple of points though.

The green trace is from the mic side of C1. The yellow trace is from the transistor side of C1. That makes no sense to me unless the mic model is a snafu.

Of course the red trace is the output.

I used a high capacitance load like you showed me elsewhere. The FR seems unaffected at that capacitance. Kudos. What is the output impedance anyway?

If I figured right, it looks like the current load is a bit more than 5 mA. If more current will help performance that is fine cause max battery life is not an issue. I guess noise might be the issue though.

If I did the sim properly it looks like the gain is more than 40 dB. Gulp! I estimated the gain of the M31 mic from its specs, and if I did it right, it comes to around 20 dB. When I spoke to LinearX they did not tell me much but did say that for loudspeaker measurement into a PC line-in, an attenuator would be necessary with the M31. That suggests that Linkwitz was on track with a 10 dB gain for his pre. Any chance of reducing the gain to about 10 dB?

If this little circuit comes in with relatively low noise, flat FR over 20-20K, and passable distortion, it could be a slam dunk winner. And I don't mean a candidate for mayor of NY.

This is really neat. Thanks.

 

[audioguru]

The gain of the transistor with negative feedback from its collector to base depends on the source impedance. I do not know how much is the simulated source impedance and I do not know if it is correct (it might be zero).

 

[ronv]

The green trace is from the mic side of C1. The yellow trace is from the transistor side of C1. That makes no sense to me unless the mic model is a snafu.

I doubt the mic is flat down to 1 Hz, but the drop is due to the coupling capacitor.

I used a high capacitance load like you showed me elsewhere. The FR seems unaffected at that capacitance. Kudos. What is the output impedance anyway?

1k ohm

If I did the sim properly it looks like the gain is more than 40 dB. Gulp! I estimated the gain of the M31 mic from its specs, and if I did it right, it comes to around 20 dB. When I spoke to LinearX they did not tell me much but did say that for loudspeaker measurement into a PC line-in, an attenuator would be necessary with the M31. That suggests that Linkwitz was on track with a 10 dB gain for his pre. Any chance of reducing the gain to about 10 dB?

If you add a 33k between the 6.8 Ufd. cap and the base of the transistor that will give a gain of about 15 Db. That would get you an output of about .325 volts RMS at a sound pressure level of 100. That's about the same as my neighbor. Or 1 meter from the speaker at the disco.


View attachment 75107[/QUOTE]

 

[oneoldude]

Thanks Ron,

I went to work on your circuit and came up with the attached. Three questions though.

Did I commit some sort of design sin by adding an emitter resistor to adjust gain? That results in two degenerative feedback systems and I don't know if that causes problems or improves things.

I lowered output impedance to be able to handle 40 feet of low capacitance cable (33p/ft). That caused current through the transistor to roughly double. Will that create a large increase in noise? Or is the MIC capsule still the greatest contributor to noise swamping out the active circuitry?

What did I screw up or am I getting close?

Thanks

 

[audioguru]

The 2N5210 transistor has a very wide range of hFE so some with high hFE will be almost saturated in that circuit.
I would use a 2N3904 that has a narrower range of hFE.

The emitter resistor reduces the gain and reduces the distortion which is good.
The circuit produces low noise.
R1 has no filter so the transistor will amplify the 9V jumping up and down and maybe cause motorboating when the battery impedance rises as it runs down.
The battery needs a bypass capacitor.

 

[oneoldude]

The 2N5210 transistor has a very wide range of hFE so some with high hFE will be almost saturated in that circuit. I would use a 2N3904 that has a narrower range of hFE.

Excellent! I have a bunch of 3904s. I thought they would be too noisy. I will use them instead.

The emitter resistor reduces the gain and reduces the distortion which is good.

Excellent! That is what I hoped. I was afraid there would be some phase anomalies due to two different degeneration paths.

The circuit produces low noise.

Excellent! I hope noise is low enough in spite of the extra current.

R1 has no filter so the transistor will amplify the 9V jumping up and down and maybe cause motorboating when the battery impedance rises as it runs down. The battery needs a bypass capacitor.

This will not be powered by a battery. I will use an analog wall wart Like the M31 MIC this thread is about. But, thanks to Ron, there will be a cap near the circuit to provide low AC impedance.

There is still a nagging question though. It is what the gain of the circuit should be. Most of my testing will be around 94 dB SPL which equals 1PA. The capsules have a specified sensitivity of about -44 dB (0dB=1V/Pa, 1 kHz which is at 94 dB SPL). That represents about 5 mV at 1Pa. This circuit has a gain of about 19 dB. So at 1PA the output of the capsule + pre amp will be around 44.5 mV. I assume the capsule will clip at 125 dB SPL(like in close mic'ed measurements). So the range of usable signal from the mic + pre amp will be from 44.5 mV to 1.58 V. Is this about right? Should the pre amp gain be lower or higher? Or is it like Goldilocks porridge, just right?

Thanks

 

[ronv]

What is the line in for your sound card? Some are around .35 volts and others 1.5.

 

[oneoldude]

I wish I knew. There are no specifications I can find re: input capabilities for that card on the net. It is now a legacy piece. It is an M-Audio Transit USB sound card. In the day it was used for hi quality recording on the road. It is full duplex stereo line-in line-out with optical line-in and optical line-out as well. Nowadays they call them USB audio interface devices but the new ones have more features to them. I have an email in to M-Audio support. I will see what comes from that.

Did I hurt anything by the changes I made?

Thanks

 

[audioguru]

Then without a filter feeding the mic powering resistor you will probably have lots of nice HUMMMMM.
Why not make the circuit correctly??