Bat detector

Hi Zach,

Thanks for sharing! Lets hope this encourages more people to hear bats and other animals. On your great webpage I also see you have experience with hydrophones, one of my hopes is that we can also develop a "hydrophone" mode for this detector. That is one of the reasons for me to use mics that are in mounted in a microphone connector so I can (whenever I build a hydrophone) just connect a cable on that point.

looking forward for the coop,

Cor
 
Looks very nice Zach.

If you want to try and get rid of the 23khz noise, it might be a good try to use an other microphone. In the pictures I see you use the ICS microphone which seems to have a similar issue as the SPU. The frequency does seem to be slightly different though.

I also have a recording made with the same type of microphone using an alc4040 (from a maono elf pro AU-410H, I do not recommend dismantling one!)
(I hope to try and see if I can also use ALC4042 and ALC4050 soon, these are available as micro USB boards)

This first one is in a quiet room without switching powersupplys and lights off.

23khz.jpg

This second one is with my faucet aerator (perlator) noise generator.
My recorder (raspberry) was quite close as well and I did not bother about the led lights.

23khzgen.jpg

You can see that signals around 23khz are always more present that other frequencies.
The microphone is amplified by a similar circuit as what I use in the teesny.

Anyway, I feel it is safe to say that the 23khz is not coming from the teensy or how it is built. We should search for it in the microphone or maybe in the amplifier circuit. It would be nice if someone had an other type of microphone like Knowles FG and record some files with the teensy.
(I protected my microphone, amplifier and codec in epoxy resin, so I can not test without the amplifier)

Kind regards,

Edwin
 
Wrapping up code development for V1.0

Hi,

Edwin (THANKS !!) and I are currently working on testing a completely renewed version of the TeensyBatdetector. Several of the requests posted in the forum have found a place in this version (v1.0).
You will see that for instance setting the date/time has become a simple procedure as you can now edit those on a lower level (hours/minutes or year/month/days) in the menu.
Another major change is that you can start recording in WAV, this makes the usage of the files in Audacity a lot simpeler.
Since this is a real update there are a lots of additions and changes. The interface has - we hope - become more intuitive, clearer.
One of the real new features is a high-pass filter that you can use and adjust. In our tests this has proven helpfull when you are in a noisy environment. Think of a warm summer evening where crickets and grasshoppers are filling the air with noise but you want to hear a possible distant bat passing ... the hipass allows you to suppress the cricket-sounds allowing you to hear those distant bats.

Those that cant wait for the release (hopefully available in a couple of weeks) and are interested in testing the current beta please PM me.

regards,

Cor
 
Last edited:
I just want to thank Cor, Edwin and the other contributors for this project. Years ago I lived in Africa and had a lot of interest in bats, and used my own simple bat detector following bats around the bat-shelters I placed around my house. Years later I moved to Norway, but because of ill health it was out of reach to do similar things over here. Stumbled onto this thread by chance, and after following this thread for about 2 years, and first had just a simple amp and 40kHz ultrasonic transducer running on veroboard connected to a teensy, but could not dtect anything. After a number of ups and downs in my health, revisited the forum, and saw ther was a PCB available. Ordered the PCB and destroyed the first mic that I tried using. now I finally have something going. Still need to do a proper enclosure for it, but yesterday I heard my first 2 bats!!!! Will have to start reading up what bats occur in the area where I stay, and based on that, see if I can hack together some bat-shelters.

Hendrik
 
Hi Hendrik,

Nice from you to share the experience ! The PCB from edwin is a great addition to this project as it makes the building so much easier.

regards
Cor
 
Hi Hendrik,

Nice to hear you have something working now.
The detector has improved a lot since the veroboard versions.

Not just in hardware but also in software. It amazes me what the possibilities are and how Cor is able to program it. The PCB helped us to eliminate a whole lot of noise by keeping data lines shorter and using big gronde planes on the PCB. A separate power supply for the audio parts and current restricting resistors in the datelines almost eliminate all the unwanted noises you would probably have picked up with the first veroboard versions.

If you look at the recordings they seem as good as most commercial available devices.

So for anyone that tried the veroboard version a few years ago, try to build it again today with the improved hardware and new software version 1.0beta. I think you will be amazed.


I am planning to work on the construction manual, adding some new photos and getting a small error out of the schematic. If you place the resistors as seen on the diagram it will work correct but the schematic I used in the pdf has a small error.

I also tried to make a video that shows how I build the detector myself. Nut quite satisfied about the quality, but I think we can use it and shooting the video again I will and up with another spare detector.

https://youtu.be/mgacnVbqvVM

Oh, about the enclosure. I like to use the boxes as described in the construction manual. The circuitboard was mad to fit that enclosure, it is about the only one that has the exact with for the Teensy 3.6 board.
Recently I did an very unwanted accidental droptest. About 6 meters to the hard concrete pavement. Expected a shattered Teensy bat detector only to discover a few cracks in the enclosure and a broken off screw support. A few droplets of superglue later I was chasing bats again. All electronics was nicely protected.

Have a bat day!

Edwin
 
I hope to contribute a little bit to the project once I have everything running. Have been a RF design engineer for 20+ years, so I have a bit of experience with noise issues, and also do a bit of software. If my brain allows it, and with all the other projects I am trying to work on, will,look a bit at the noise-issue from the mic. Seems not to be so much a conducted issue, as it goes mostly away when you put a dummy-mic on, so, I am starting to wonder if it might be some sort of piezo-electric effect from the display that gets picked up by the microphone. (Did a project with tracking filters, in which I had to switch small capacitors rapidly between 200V and 0V and you could clearly hear the SMD-capacitors singing).

Hendrik
 
Hi Hendrik,

Thanks for sharing that article, I have seen it before. We are interested to hear from you how good that SMX-US microphone works in the current setup. The article mentions that these Mics are not produced any more, did you find a source ?

greetings
Cor
 
Last edited:
Why not simply try to mount a Knowles FG-23629 element in stead of the MEMS element.

It seems available form Mouser and Digikey, the element is much more expensive than the Mems elements but they seem to be widely
used in batdetectors. These wholesalers do have quite large small order costs/shipping charges of will only deliver to a business address.
(buying two or some extra parts at mouser seems a better choice since above €50 shipping seems to be free be aware price is without vat)

Maybe you have an other source near you.

Being a cheapskate and qiuite please with the Knows MEMS element I never tried a FG-23629

It seems to me the FG-23629-P16 is a leaded version.

Edwin
 
Hi Hendrik,

Thanks for sharing that article, I have seen it before. We are interested to hear from you how good that SMX-US microphone works in the current setup. The article mentions that these Mics are not produced any more, did you find a source ?

greetings
Cor

Hi Cor, found a UK source on eBay. I assume it is a copy of the real one, but it is cheap enough to take a chance. Still looking for the pinouts though...

My current MEMS microphone I have now mounted on a 10cm cable, instead of plugging it directly into the PCB, and it made a huge difference in the digital noise coming from the display. Now have to get my 3D printer running again to print an enclosure. Plan to put a connector on the thing that is compatible with those from Wildlife Acoustics, so that (in theory), one can just swap microphones.

Hendrik
 
Hi Edwin, have a few similar electrecets here that I plan to test too, but is better to have a "real" bat microphone, to compare them to. Just need to schedule the bat-detector in between my other projects. Unfortunately it seems as though there is not so much bat-activity here, but have also picked up some moths. Will make a bat-stimulator and test the mics with that.

Hendrik
 
Interesting to see how it performs. I am using plugs on my detectors also, thusfar I have used a GX16 connector but the one that I am currently building I am trying to see if using a 3.5mm stereo-jack works as planned. Another option I might use is the GX12 connector. The nice thing about the 3.5mm jack is that its nice and small and I use a complete metal version which also shields the mics. Unfortunately the mics just dont fit inside so I have them sticking out a bit with a microphone cap on that (tie-wrapped).

Cor
 
Cor, was planning to use GX16 connectors, which are cheap, then I saw all the mics from Wildlife Acoustics use the Conxall Mini-Con-X series connectors. Guess there is not a standard for bat-detectors, as I have also seen some bat mics with 3,5mm connectors. Guess one must plan on having a few different adaptors.

Hendrik
 
New teensy_batdetector V1 beta sourcecode and hex released

Hi,

I have just uploaded the most recent code for the teensy_batdetector. Its available from https://github.com/CorBer/teensy_batdetector
A lot of work during testing of this version in previous months has been done by Edwin (PE1PWF), I want to thank you for spending all that time and also coping with my truckloads of mails :)

A few other members of this forum have been testing and reporting on the pre-release version and I also want to thank you for that.

As usual with freeware, this is not a finished product :) But I think its another step forward in building an easy to use detector.
Now the release is done ... next up we will try to provide a bit of a manual so you all know whats new or changed in the code.
Some new features are not directly visible, for instance when you start the detector up with the left-pushbutton pressed it will skip loading settings from the EEprom.

As usual a pre_compiled hex is available to make installing easy. For those that want to play with the sourcecode, be aware that I am not programming in a TeensyDuino setup.
My programming setup (I am working on a linux mint 19.3) consists of the free Microsoft Visual Code Editor using the PlatformIO plugin (which allows me to program many different boards).

The source and hex-files from the previous versions have been moved to another repository on github (https://github.com/CorBer/OLD-teensy_batdetector).

Have fun listening to Bats !

regards
Cor
 
Last edited:
Hi Hendrik,

I looked at the Mini-Con-X connectors before but found them way to expensive.

regards
Cor
 
Cor, agreed. Those mini-con-x connectors are too expensive. Will just make some sort of adapter mini-con-x -> GX-16
Thanks for the new firmware. Will download it and try it out.

Hendrik
 
Thank you for releasing the final 1.0 version of the software Cor.


I am compiling in Arduino on a windows machine, on my Google drive you will find a complete windows package of the final version.

Simply extracting the zip file to "My Documents" should put the files in the right place.

You can also find a short manual about how to update the software and a user manual for the 1.0 version software.

Kind regards,

Edwin
 
Why not simply try to mount a Knowles FG-23629 element in stead of the MEMS element.

It seems available form Mouser and Digikey, the element is much more expensive than the Mems elements but they seem to be widely
used in batdetectors. These wholesalers do have quite large small order costs/shipping charges of will only deliver to a business address.
(buying two or some extra parts at mouser seems a better choice since above €50 shipping seems to be free be aware price is without vat)

Maybe you have an other source near you.

Being a cheapskate and qiuite please with the Knows MEMS element I never tried a FG-23629

It seems to me the FG-23629-P16 is a leaded version.

Edwin

Hi all
I am new to this group and this project seems to me very interesting.
I work with ultrasonic microphones since 1998, this is my website www.dodotronic.com
Regarding the FG-23629: I use this sensor in the Ultramic384k. If compared to MEMS it is very insensitive to water and dust, moreover it has a more flat frequency response while MEMS have a strong response around 50kHz

Ivano Pelicella
 
Hi all, found a informative article about microphones for bat-detectors. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6338251/

Have also ordered a SMX-US bat detector microphone, which I will compare to the current 2 I have (a standard piezo-transducer, and our "standard MEMS-microphone".

Hendrik

This test is well done but it lacks of high frequency test, and after 50kHz we can find the main differencies.
For a good test it is important to have a reference with which to compare the results.
I have got this equipment:
SENSOR = B&K 4138
PREAMPLIFIER = B&K 2669
AMPLIFIER = NEXUS 2690
WH3219 = frequency extension(140kHz -1dB; 200kHz -3 dB).
Calibrator = Temna 72-2680 94dB / 114dB
sampler ADC: Ultramic384K 384kHz sampling rate 16 bit
Software = Multi-Instrument Pro 3.7
frequency generator Owon AG2052F

Beyond 60kHz SPU0410LR5H-QB greatly decrease the sensitivity while the FG reach more than 90kHz
The Vesper VM1000 has got a very scarse sensitivity in ultraound field
Ivano
 
Hi Ivano,

Welcome in this project, I hope you can share your experience with the teensy-batdetector as soon as you have build it. I am especially interested to hear/see how the different microphones, that you allready have tested it seems and use for the dodotronic mics, work on the system. Most people are currently reluctant to buy an FG microphone due to the fact that the price is so much higher than an SPU microphone. Sometimes the "best" is not different enough from "good", but we simply dont know. Nobody has mounted (as far as I know) a FG microphone on this setup. That difference should be very clearly visible in a) signal quality (less noise) and b) sensitivity (spectral range). But this test needs to be done when mounted on our teensy_batdetector, as it is a part in a chain of electronic devices and software that convert the ultrasound to something we can hear/see. Maybe the weakest link in this setup is not the microphone but the audioboard (since we are using a normal audioboard at highend samplerates, it was never built for that purpose).

I hope you can help us forward on that question,

regards
Cor
 
It would be great to have a nice measurement on the unit Ivano. Parts are posted and I understand Ivano alo has other projects unfinished. I just hope that your curiosity will make you build this detector soon.


I only used the SPU0410LR5H-QB microphone and did a few tests with the ICS-40730, this ICS microphone seems to be loved for the good signal to noise ratio. I mainly found a big hole near 40-55khz depending on the way it was mounted so I do not use it anymore.

The microphone mount that suits me best is a shallow cone that reduces the wall thickness of the case to 0mm where the microphone acoustic port is. Deep holes or narrow cones seems to distrurb the recordings. For simply using as bat detector that will also work fine but for clean recordings the shallow cone and a good seal between microphone and the housing is best.

Specially the seal is important and use low gain settings close to the bats give good recording results.

The sensitivity of the SPU0410LR5H-QB seems to have been described by knowles up to 80kHz, the curve does not quite follow the line described by you Ivano, but I know a lot depends on the way the microphone is mounted.

Because the sensitivity tends to fall on higher frequencies, the ampliefier circuit in the detector was designed to counteract that effect. The image is from knowles but I added the red line, the red line is just about the gain of the amplifier. You can see the amplifier helps to amplify the higher frequencies more than than the lower frequencies. (it also attenuates the lower frequencies)

SPU0410LR5H-QBwith amp.png

I know this does not give perfect results but it well help to get closer to a flat response.

I hope you can provide some information on how well or how bad the frequency response is. If you turn up the microphone gain you can clearly see the "bump" in the graph around 23kHz. That one is hard to get out but I have not found that to be a problem in my recordings. One will not suddenly see other species in the spectrograms. Using the highpass filter in the software can help reduce the 23khz bump.

To illustrate how my recordings look I have this picture with two parts of recordings showing a pipistrellus and myotis. (no filtering or cleaning afterwards)

pipistrellismyotis.jpg

I also might want to try the FG-23629 element or Momimic, but I think it is wise to wait what results you get Ivano.

If the FG-23629 has a good flat response we might want to change the amplifier.
C1 (4.7nF) and C3 (15nF) should both be at least 100nF to give the same gain over the whole ultrasonic range. (for low frequency sounds 1000nF is recommended)

I guess the Momimic on your website could be used directly connected to the mic-in of the teensy audio board so one does not even need the TL972 amplifier circuit at all.

Kind regards,

Edwin
 
Back
Top