Missing frequencies bands in the recorded sound spectrum with AudioMoth placed in the official case.

I had turned to this forum looking for an answer to the problem with my AudioMoth, a hint elsewhere in this forum regarding the seating of the unit inside the case led me to solving what is a different problem. I’m sure that other users of AudioMoths who are placing their AudioMoth units in the official cases may have the same problem as I had, so I thought I’ll share my solution.

Having recently bought another 2 AudioMoth units in October to add to my collection of 1 (bought this summer) I now have more options for detecting bats by distributing the 3 units around a small area.

I was very pleased with the first one, mainly used during Sept and Oct to ID bats. But when doing a daytime test on the 2 new ones in the garden last week I was so disappointed by the poor performance of one of them. A significant part of the audible range was greatly diminished, which if repeated in the ultrasound range would make bat ID a bit of a challenge perhaps.

I re-tested all 3 together in the garden again, just a short 10 mins test then listened (and analysed the spectrograms), my 3rd AudioMoth still performed poorly, it seemed that several frequencies bands were being attenuated by -10 to -30dB, this is absolutely unacceptable considering the purpose for which I had bought these units.

I decided to create a “standard” audio test file that can be used to repeat audio tests to try to get a “measured” record of the performance of the AudioMoth units as I investigated the problem of the missing frequencies.

Detail of my “standard” sound file used could be provided later in another posting if requested, along with the recorded spectrograms of before and after solving the issue.

Up to this point I had deployed all the AudioMoth units in their cases. Another 10 min test without using the cases revealed the issue seemed to be the official AudioMoth case since all 3 am units performed equally well without being placed in their cases.

So, another test with the units put back in their original cases again and then repeated with the cases swapped over showed that the issue was back with one of the AudioMoth units and not the case!!!

A closer look at the AudioMoth units, and how they fit in the official case had me thinking; the difference between the 2 new units and the old original unit was that the two new units had a raised solder blob on the terminals from the battery carrier (the red arrows in the image below). The first AudioMoth I bought had a very flat solder joint almost flush with the pcb board. I suspected that the raised solder blobs were keeping the unit from sitting flush with the aperture through which the sound enters the case.

Fig 1 End view of an AudioMoth unit showing the raised solder blobs

Careful filing off the top of the solder blobs using a flat needle file to about 1mm off the pcb board and with someone holding a vacuum cleaner close by, then blowing out any remaining filings with a can of pressurised air (from my camera bag) to complete the task. (I also covered these 2 solder blobs with a small piece of electrical insulation tape to prevent any accidental shorting of the battery terminals in future use after having read the post about the melted battery case).

The poorly performing AudioMoth unit was carefully placed into a case and tested with the other 2 units simultaneously.

The result of the next test, SUCCESS, problem solved.

All 3 AudioMoth units now behave in a similar manner, the 3rd am unit now sits in the case with the mic flat on the membrane covering the sound vent of the case (or at least I assume it does since I can’t see it of course since the unit sits on top of the membrane!).

n.b. all the issues and testing here has been in the frequencies below 14kHz since the bats haven't been out to play over the past few days so I can't comment on how this affects the ultrasound frequencies.

I’ll be interested to hear from anyone else who may have the same problem, and whether this post helps them.

Hopefully the manufacturers can be informed of this problem and arrange to have a little less solder/wire protruding above the pcb 😉

4 Comments

Alex Rogers

Nov 11

Hi Trevor, Can you confirm where all three AudioMoth came from? Alex

Trevor Shepherd

Replying to

Hi Alex, The AutoMoths came from different places. The first one I got was via GroupGets in United States, this being the one with flat, flush solder joint. That unit was fine.

The second and third AudioMoths I bought recently were via LabMaker in Germany, these are the 2 with raised solder blobs. The photo I posted above was what I call my "am2" unit before I filed the blobs flatter. The third one "am3" unit was the one that initially triggered my concern, it had a slightly more pronounced raised solder blob, perhaps about 0.5mm higher that I filed down before I thought about taking a photo and putting up a post about it.

We'll let LabMaker know as they should be keeping the solder joint height to below the microphone standoff dimension. What frequency range were you testing? We'll try to reproduce this effect since it hasn't been reported before.

Edited

My first use of the new AudioMoths was just last week, I had them outside early evening in their cases expecting to pick up just the sound of blackbirds etc (but also hoping for bats).

There were no bats detected. With the lower frequency signals, I generally listen to them at full speed since I can make more sense of these frequencies <12kHz.

The difference between the recordings were obvious, just by listening, the third unit did not sound right hence the investigation. I have no equipment that can produce calibrated ultrasound frequencies so I’m restricted to the human audible range. Since there were no bats during that first use then I had nothing I could use to assess the ultrasound frequencies. (apart from jangling car keys!)

So, to answer your question, I could only do what are semi controlled tests up to 14kHz.

I do have a set of sonograms from recordings of the am units inside the cases and also outside the cases that I could post here if they might be of any use.

Trevor