Wow, about -65 dB THD+N for the BKs, more like -73 dB for the GRAS, much of which may be owed to its lower bandwidth. I didn't expect the preamps to add that much noise, but again, this is probably down to the 100 kHz bandwidth in your measurement?
How did you feed a signal? Dummy parallel capacitance (how large?) to simulate a capsule and then a series capacitance to inject a signal? That would be a good way to measure distortion in the preamp, but it will not allow you to measure distortion due to charge transfer into the not-high-enough input resistor.
The distortion due to input impedance is explained in on page 2 of this document
https://www.by-rutgers.nl/PDFiles/KE_4-electret.pdf *, which is a page from an AES paper by Holger Pastillé. I don't have it but it is also explained (in German) in his PhD thesis:
https://depositonce.tu-berlin.de/bitstreams/03deab4a-673a-4f5e-896b-840697dad3a7/download
* If that fully bootstrapped circuit is really what is used in B&K preamps (I have a schematic of only one but I believe it was differen - need to look up), then it should have significantly lower distortion than the GR circuit that bootstraps only the input of the FET.
Last edited:
I'll explore the different preamps more later. There are a handful if issues with measurement preamps having to do with common standards.
In the meantime here is a plot of a Stax "IEM" (SR001-MK2) driving a GRAS RA0045 711 coupler with a 1/2" pressure mike at 92 dB SPL. The second harmonic is around .03 % if I have my sums right. I was astonished when I saw that.
In the meantime here is a plot of a Stax "IEM" (SR001-MK2) driving a GRAS RA0045 711 coupler with a 1/2" pressure mike at 92 dB SPL. The second harmonic is around .03 % if I have my sums right. I was astonished when I saw that.
Suprised at how high the 2nd harmonic was? It could be down to the STAX driver or the preamp of unknown type.
It would be interesting to see harmonics like in this picture for various preamps with a dummy capsule and a series capacitor to inject a signal. It is something I can probably do soonish.
The next level would be to include the charge loss due to the input resistance. This requires having an acoustic signal with very low distortion. I remember reading a hint in some B&K paper that they use a long tube that is damped in the places where harmonic resonances would have their highest air velocity, but I never found the paper where they actually present that setup.
The other thing I have been wanting to try is a pressure chamber:
https://www.artalabs.hr/AppNotes/AN5-MicroMeasChamber-Rev4Eng.pdf
Apparently, the excursion of the driver is extremely small, resulting in very low harmonic distortion. I don't remember where I read this. The setup could be further improved by using a better driver and current drive.
It would be interesting to see harmonics like in this picture for various preamps with a dummy capsule and a series capacitor to inject a signal. It is something I can probably do soonish.
The next level would be to include the charge loss due to the input resistance. This requires having an acoustic signal with very low distortion. I remember reading a hint in some B&K paper that they use a long tube that is damped in the places where harmonic resonances would have their highest air velocity, but I never found the paper where they actually present that setup.
The other thing I have been wanting to try is a pressure chamber:
https://www.artalabs.hr/AppNotes/AN5-MicroMeasChamber-Rev4Eng.pdf
Apparently, the excursion of the driver is extremely small, resulting in very low harmonic distortion. I don't remember where I read this. The setup could be further improved by using a better driver and current drive.
This is the thread about the pressure chamber:
There is even a plot in the ARTA manual towards the end of section 5.3. The venerable WM60 does ok. The Gefell MK221 is good but not stellar. I wonder how much of that 112 dB data point is lifted by noise.
There is even a plot in the ARTA manual towards the end of section 5.3. The venerable WM60 does ok. The Gefell MK221 is good but not stellar. I wonder how much of that 112 dB data point is lifted by noise.
Today I bumped into this paper from Antonin Novak; "Reduction of nonlinear distortion in condenser microphones using a simple post-processing technique"
https://www.ant-novak.com/publications/papers/2025_jasa.pdf
https://www.ant-novak.com/publications/
I thought it would be a very interesting read for here.
https://www.ant-novak.com/publications/papers/2025_jasa.pdf
https://www.ant-novak.com/publications/
I thought it would be a very interesting read for here.
The site is not accessable from here or its broken. I would like to see it. Bill Waslo made an interesting pressure chamber with a distributed coupling that should generate very high SPL's. I have not tried it myself.
There is an issue of the non-linearity of air affecting the distortion. As the spl goes up it becomes significant. The pressure can increase a lot but can't go below zero. Is there a calulator for that?
There is an issue of the non-linearity of air affecting the distortion. As the spl goes up it becomes significant. The pressure can increase a lot but can't go below zero. Is there a calulator for that?