Tweeter source for mic calibration

Member
Joined 2007
Paid Member
I have ordered a couple of ND20FB-4 tweeters in order to perform second teir calibration of MEMS mics using a known microphone for the range >2kHz. The idea is that I will put the tweeter on a flat baffle placed on the floor, position the micrphones in exactly the same point in space above the tweeter and perform gated measurments in order to calibrate them. I used Vitux-cad diffraction simulator to come up with this baffle design:

1646093371933.png


I think the ripple ~5kHz will not occour as the tweeter will not be omi at this frequency. Any comments? I haven't seen anyone calibrating microphones with this kind of source before so could be making some silly mistakes.
 
This is for a polar measurement rig I described in this thread:
https://www.diyaudio.com/community/threads/speaker-measurment-mic-arc.378894/
I have decided to go with MEMS microphones as I believe, although they have non flat response <100Hz and >10kHz, they are likely to be very consistent unit to unit and so I can probably use a single CAL file for multiple microphones. A few years ago electret mic capsules where cheap and of good quality but I didn't find anything with a flat response and also cheap that is still available. MEMS mics are very low cost <1$ and include signal buffering which reduces the amount of support circuitry. Assuming they can be calibrated flat 20-20kHz their main downsides appear to be:
1) not as omi directional in vertical plane as electret capsules due to the acoustic port been on one side of the capsule.
2) acoustic overload points around 120dB
3) non flat 20-20kHz response (but I think this is manageable)
4) need machine assembly due to bottom side pads
for the application of a polar measurment arc I think they are usable though!

Here is some information on an existing MEMS measurement mic:
https://audioxpress.com/article/fresh-from-the-bench-ik-multimedia-arc-2-5-mems-test-microphone
The mic I am currently look at is the SPU0410LR5H-QB:
1646132240866.png

For this mic the resonant peak is ultrasonic which should reduce error in the HF range.
 
Int. J. Metrol. Qual. Eng.4, 97–107 (2013)c © EDP Sciences 2013 DOI:10.1051/ijmqe/2013045 Acoustic metrology – an overview of calibration methodsand their uncertainties
https://www.metrology-journal.org/articles/ijmqe/pdf/2013/02/ijmqe130045.pdf this is a very nice article summarizing microphone calibration methods, I'm not sure if its pay walled. Secondary calibration using a mixture of presure response and free field is mentioned as something that is sometimes done. For the presure response the two microphones can just be held in close proximity to each other in the near feild of a woofer. I would make a calibration source where a small woofer presurises a small sealed box with the mics in (ensuring uniform presure) however this would require sealing the tips of the mics in the box which will be tricky for the MEMS mics as they are just going to be exposed PCBs not round.
 
  • Like
Reactions: homeswinghome
How about a mechanical sound reference for microphone calibration. The sound source doesn't need to be linear as long as its characteristics is known and is possible to build by anyone and succeed to within, say, 0,1 dB 20-20k. I wonder what materials could be used and how it could be built with repeatability... I'm envisioning some sort of hinge and clap... something falling by its own weight hitting something - all within some sort of box - or maybe a tube? With a hole for the mic... What items could be said to be standard, stable and worldwide available except for Big Mac...

//
 
I have used thunder to compare mics to each other. Unless the thunder is very close, there isn't a lot top end, but close strikes do OK. My supposition is that source is equal at both mics at the same time.
 
How about a mechanical sound reference for microphone calibration. The sound source doesn't need to be linear as long as its characteristics is known and is possible to build by anyone and succeed to within, say, 0,1 dB 20-20k. I wonder what materials could be used and how it could be built with repeatability... I'm envisioning some sort of hinge and clap... something falling by its own weight hitting something - all within some sort of box - or maybe a tube? With a hole for the mic... What items could be said to be standard, stable and worldwide available except for Big Mac...

//
well there is the piston phone which is a mechnical device, they are fixed frequency though: https://www.bksv.com/en/transducers/acoustic/calibrators/pistonphone

I already have a class 2 sound calibrator that can do 94dB, 104db and 114dB but only one frequency (1kHz). They are usefull to determine the absolute sensitiviy of the mic-preamp-adc chain in the field. The constuction of one is described here: https://www.mvaudiolabs.com/diy/1khz-microphone-level-calibrator/ I would just buy one though the OEM ones that get rebadged to £250 units are £60 on aliexpress, you have to choose the calibrator for your specific mic diameter though which is annoying as they could easily come with the inserts for all the diameter varients.
 
Pistonphones are not fixed frequency. You can change the motor speed. Directions are in the manual for the B&K. Its intrinsically accurate for low frequencies but stops at around 330 Hz. For higher frequencies I use a pioneer ribbon like the HPM900 I found on eBay. Its useful to about 70 KHZ. however above about 40 KHz there are a lot of interactions due to the size of the driver vs wavelength. If you do not have a reference mike you won't get far since none of these solutions will have intrinsic accuracy.
I have been working on a spark gap but even that is not too predictable. https://www.diyaudio.com/community/threads/spark-calibration-of-microphones.353494/post-6065519 I'll get back to that project soon.
 
  • Like
Reactions: Pano