Shopping for a measurement mic

[Biblob]

There are no systematic issues with UMIK I'm aware of, the rate errorsnseen were isolated cases. REW can compensate for clock rate differences if required, the setting is under the Analysis preferences, but when measuring in a car there are so many close reflections that the acoustic timing reference could easily be affected.

Could you comment on what 617 said; that using a Umik-1 is not a good idea for designing a loudspeaker because of loopback/phase caused by it's own usb-interface.
Are these problems solvable, you think?

 

[JohnPM]

Could you comment on what 617 said; that using a Umik-1 is not a good idea for designing a loudspeaker because of loopback/phase caused by it's own usb-interface.
Are these problems solvable, you think?

I don't know what post or problems you are referring to, sorry. USB mics need a timing reference, the acoustic timing reference can provide that in the kind of environments likely to be used for design but an analog mic and a loopback connection might be more convenient. If making polar measurements the timing reference source would need to be a separate driver or speaker to the one being measured.

 

[617]

I can't speak to the theoretical equivalence of usb mics with some specific technique and an analog mic with a loopback, but I can say practically that the latter works very well - there are about 100 things that can go wrong in speaker measurement, and minimizing them is very important for idiots like me. Next time I do measurements I will try to compare the acoustic timing reference to the loopback method.

The other thing I really like about the analog mic is the ability to change the gain. I find that with averaging I can measure at pretty tolerable SPL, not sure how that works with a USB mic.

 

[Dave Zan]

Just FYI, I measured 90-100V equivalent on some better electret elements. If you have raw capsules without FET's you can buck the built-in polarization and measure the loss in sensitivity.

Thanks Scott for that update, it has improved since I last looked.
So the electrets are only 6 or 7 dB behind in polarisation.
Makes it plausible that part of the B&K max SPL capability comes simply from better front-end/pre-amp headroom.
As one would expect at the price.

Best wishes
David

 

[Dave Zan]

There are no systematic issues with UMIK I'm aware of, the rate errors seen were isolated cases. REW can compensate for clock rate differences if required... in a car there are so many close reflections that the acoustic....

The specific application I want to measure is a DIY clone of the Klippel near field scanner.
Link >here< if you haven't already seen one.
There was a discussion on DIYaudio and now there is a project on this forum, both easy to find.
I don't see any way that an acoustic reference will work and we need to maintain time lock over hundreds or even a few thousand measurements.
So perhaps over an hour.
I must admit I haven't reached the point to really study this issue closely but USB looked problematic.
I did have REW as the presumptive measurement system, so very interested to have your comments.

Best wishes
David

 

[scott wurcer]

Thanks Scott for that update, it has improved since I last looked.
So the electrets are only 6 or 7 dB behind in polarisation.
Makes it plausible that part of the B&K max SPL capability comes simply from better front-end/pre-amp headroom.
As one would expect at the price.

Best wishes
David

More importantly the capsules often use the common source connection (2-wire) which has relatively high distortion and limited DNR. The three wire mod (or a bare capsule) can be set up to be much better. This is all DIY mic building which does not work for everyone.

 

[Dave Zan]

More importantly the capsules often use the common source connection (2-wire)... The three wire mod (or a bare capsule) can be set up to be much better....

Yes, but it's true even for the three wire capsules set up as source followers, for instance in the Primo EM273 spec sheet.
The B&K set-up just pumps out more before it clips, partly because the Primo is spec.ed for only 5 V supply.
I understand you are a keen mic builder, have you experimented to increase the supply volts to typical capsules FETs?
Any idea what a reasonable maximum would be?
Obviously we won't reach B&K levels but any improvement would be nice.

Best wishes
David

 

[scott wurcer]

Yes, but it's true even for the three wire capsules set up as source followers, for instance in the Primo EM273 spec sheet.
The B&K set-up just pumps out more before it clips, partly because the Primo is spec.ed for only 5 V supply.
I understand you are a keen mic builder, have you experimented to increase the supply volts to typical capsules FETs?
Any idea what a reasonable maximum would be?
Obviously we won't reach B&K levels but any improvement would be nice.

Best wishes
David

Better than increasing supply is fully bootstrapping the three wire capsule. You can get over 8V p-p with virtually no distortion off of a 9V battery. OTOH the FET's used typically have much more than 5V breakdown voltage if you wanted to pursue that route.

 

[Dave Zan]

bootstrap...over 8V p-p with virtually no distortion off of a 9V battery.

Bootstrap appeals to me, do you have any links for this?
BTW- Bootstraps appeal because of their theoretical neatness, did you ever consider the distortion reduction 'trick' (for lack of a better word) in the AD797 as a type of bootstrap?

OTOH the FET's used typically have much more than 5V breakdown...

Yes, I expected maybe 15 or 20 V minimum, do you have any real world data?

Best wishes
David

Edit. Just found out a typical JFET is/was the Sanyo 2SK296, rated at 20 V.
Maybe replaced by newer types but probably these will be very similar.

 

[Colonel7]

No probs with a PC and a USB TRRS audio adapter.

Any specific one? Seems there's a thousand no name ones with lots of misspellings and claims. Turns out I got one for Christmas a ways back that I forgot about and it's not working right with android on my archaic phone.

 

[Blumlein 88]

Any specific one? Seems there's a thousand no name ones with lots of misspellings and claims. Turns out I got one for Christmas a ways back that I forgot about and it's not working right with android on my archaic phone.

Does it work to 8khz and die above that? Had one like that and it was the phone.

 

[Colonel7]

Does it work to 8khz and die above that? Had one like that and it was the phone.

After playing with it some more it does work above 8k but the Audiotool app is garbage (at least on Android). Maybe it's user error but in any case I'd rather use it with my laptop through usb

 

[Colonel7]

No probs with a PC and a USB TRRS audio adapter.

Thanks. Got one for my laptop and the IMM-6 works just fine with REW now.

 

[kipman725]

You can use a dynamic mic for distortion measurements, they won't have a flat frequency response but you can create a correction file using a mic that does. https://service.shure.com/s/article...really-loud-sounds-maximum-spl?language=en_US not as accurate as a top end high SPL mesaurment mic but a lot cheaper and good enough for development of speakers.