Bat detector

Introduction

Hi All
I wanted to express my amazement and gratitude to everyone who made this project happen. Outstanding work on all fronts. Electronic, software and mechanical. I found this project a few weeks back while buying some Teensy parts for another project. Long made short, I got far enough along to program it and see if it was alive just a few minutes back and DANG! The thing fired up the first time. I am a couple of backordered pushbuttons and a knob or two from being truly finished, but it works!! I'm looking forward to reading the manual, playing with it on the bench, looking over the code as well as taking it out in the field, well, the streetlight in front of my house, and discovering whatever is out there.
Thanks!
Steve Bennett
 
I have been using the Knowles mic, successfully, and am now testing the ICS 40730 mic.

I have soldered it, I believe correctly, but get no signal. I am getting 2.5 V on the inverting output, which doesn't seem right.

The unit is glued with conductive epoxy to an aluminum holder. The case is 0 ohms to the Ground electrode. I soldered a lead to Ground, just in case, after I got no signal in initial tests.

Any suggestions would be appreciated.

IMG_2103.jpg
 
It looks like the red wire is ouput -
(Datasheet shows top view, terminal side down)

Are you using both outputs into a differential amplifier?
If you are just using one output, the other should be floating.

Kind regards

Edwin
 
Tolles Projekt, lies sich gut aufbauen.
Allerdings habe ich Mängel festgestellt.
1. Stromverbrauch zu hoch, das Display sollte abschaltbar sein, nicht nur die Hintergrundbeleuchtung sondern die Spannung des Displays. Der Stromverbrauch wird dadurch um 60% gesenkt.
Wenn der Detektor über Nacht im Freien aufgestellt wird braucht man kein Display, es stört nur wenn da im Wald etwas leuchtet.
2. Fat32 kann nur 4GB große Dateien schreiben.
Dies ist bei Daueraufnahme nach 1,75 Stunden überschritten und die Aufnahme unbrauchbar zerstört. Es sollte möglich sein eine kleinere Datei zu schreiben und dann eine neue Datei anzufangen.
3. Ist die Pause bei dem automatischem Aufnahmemodus von den minimal einstellbaren 5 Sekunden nicht zu lang.
4. Gibt es ein Handbuch in anderen Sprachen wie in deutsch.
 
Hi Helge,

We know the display uses a lot of current. We have seen the question before to turn off display power completely.
This gives a problem because it can not be turned on. The display needs initializing and that needs a trigger. It is initialized on power-up so after switching off the display power and turning on you need to cycle the power to get it working again.
I do not know if we can find an easy fix for that. The best way now is wiring the backlight bright/dimm function to bright/off or dimm/off so you save some power and do not have the backlight on. (and power by al larger powerbank perhaps)
Maybe a new version gets adjustable backlight and a baklight off switch, but maybe we could find a way that the display on-off switch also triggers the processor info initializing the display again. Anyway, this is something to think about.
The detector was designed as handheld "live" detecting device but recording options do make it more versatile. It was not intended to be used as stand alone logger.


I guess a max recording limit could be built in, but I'm not sure how difficult that is, the higher the sample rate the shorter the maximum duration. I guess this is a nice ome for Cor to think about. Although I can not imagine why one would not use the autorecord function if one wants to record over a loner period.

The pause can be shorter I guess, maybe we should set that a 1sec for minimum, I don't think there is any reason to prevent us to do so. Audiomoth seems to need 5 seconds but I don't thing we need that. If you can compile the code yourself you could set it to 1 second.

Sorry, there is no German manual.
Maybe you can help us make one. I can send you a .doc file of the English version if you want.
 
Hi Helge,

Thanks for the questions, my german is so bad I wont even try to answer in your native language ;)
1) As Edwin allready stated, its possible to simply add a switch to the screen. For a complete standalone recording that should be easy to implement. Just setup your detector with the recordingsettings, start it and switch the screen off (hardware switch). When you pick the detector up you will need to switch the whole unit off before switching on the screen and the unit. Otherwise communcation with the TFT will not work. But thats the easiest way around your problem. Since the unit was designed mainly for "live" detection and walkarounds we never spend a lot of time on this issue. You state that even without the backlight it would use a lot of energy, in the autorecording the screen is not updated continously and removing the backlight should then remove a lot of power.
I have never looked into the option to simply "switch" the screen OFF and then "live" restart it from the Teensy. That might be possible but it would demand a very clear procedure. Switching it off could be programmed as a default action directly after starting "auto-record". Switching on would demand a very clear procedure by pressing a button long or so. I will see if this is at all possible.
2) A single file can indeed be only 4Gb large but this unit was never ment to record long files. In fact I would strongly advice against doing just that. What if you get something happing with your detector like a low-power issue or a glitch in the software ... you would spoil a large file. Also in processing files from bats etc using only chunks of recordings is efficient and you dont record all the time were no animals were present.
3) We can reduce the minimal time to a lower setting, I am working on an update and that will allow you to set the minimal time between recordings to be 0 seconds. The next setting is 5 to keep things simple. But at zero the recordings can start directly after each other after the previous recording has been saved.
4) As edwin stated, maybe you can help us there.

regards
Cor
 
Danke für die schnelle Antwort.
Ich habe R12 ausgebaut und als R9 eingebaut.
So habe ich ein gedimmtes Display und ich kann die Hintergrundbeleuchtung so mit dem S3 ganz abschalten, das spart schon 35mA.
Das helle Display brauche ich nicht.
Eingebaut habe ich auch einen 5Ah LION Akku, der reicht für die Aufnahme von 12 Stunden.
Dabei werden etwa 42GB Daten aufgenommen, ich verwende eine 64 GB U1 Speicherkarte.
Bestellt habe ich ein Poti mit Schalter, der die Leitung von 5V_IN zum Display inklusive Reset und zur Hintergrundbeleuchtung unterbricht.
Das spart nochmals 35mA und das Rauschen und sonstige Störungen in der Aufnahme sind deutlich geringer.
Laut Datenblatt sollte es besser sein direkt am Ausgang von IC2 einen kleineren Elko mit einem parallelen kleinen Kerko einzusetzen.
Dabei sollte der Spannungsregler oberhalb der Nutzfrequenz schwingen.
Ich werde dies mal testen.
 
Horn

Ich habe ein Horn montiert, das ist von einer kaputten Fahrradlampe.
Dieses ähnelt einem Fledermausohr und bringt etwa 10db.
 

Attachments

  • IMG_20210310_231006.jpg
    IMG_20210310_231006.jpg
    37.6 KB · Views: 111
Hi,

I have tested the usage of the sleep functionality of the ILI9341 display, that seems to work as planned. It will NOT switch the backlight off but it will keep the TFT alive but asleep (alas using less energy). So in order to achieve a real dark screen when leaving the detector out at night a direct switch on the TFT backlight is by far the easiest. Alternatively you could control the voltage on the backlight with a potentiometer as a strong dimmer. In the next update this option will be available, its ment to work ONLY in autorecord mode. You can bring the TFT back alive at any moment by pressing the left-pushbutton (next to the left rotational-switch).

regards
Cor
PS. I will add (or better said reinstall) some code to use PWM (from pin 3 on the Teensy) to allow more control of the LED backlight. For those ILI9341 TFTs that do not have a build in transistor to use PWM a bit of extra electronics is needed to steer the LED. See https://forum.pjrc.com/threads/5686...creen-question?p=209521&viewfull=1#post209521
 
Last edited:
These are examples of common 2.8" ILI9341-SPI screens
old setup without transistor, so needs additional hardware to use PWM controlled backlight
Selection_671.png
new setup with added transistor can be directly driven by PWM
Selection_670.png
 
Ist da ein Fehler im Schaltplan?
Das Mikrofonsignal wird über R2 und C6 gegen Masse Kurzgeschlossen.
C1 sollte wohl an Pin 2 des TL972.
 
Hi Helge,

Please communicate in English, that way everybody on this forum can understand what is going on. Ive pulled your text through google translate and that shows:
Is there a mistake in the circuit diagram?
The microphone signal is short-circuited to ground via R2 and C6.
C1 should probably be on pin 2 of the TL972.

If you cannot write english good enough I suggest using a translator before posting,

kind regards
Cor
 
Hi Frank,

Nice info, i tried to reverse the question from Helge with deepl back to german and that looks ok to my dutch knowledge of German.
Liegt ein Fehler im Schaltplan vor?
Das Mikrofonsignal wird über R2 und C6 gegen Masse kurzgeschlossen.
C1 sollte wahrscheinlich an Pin 2 des TL972 liegen.
 
Last edited:
Ist da ein Fehler im Schaltplan?
Das Mikrofonsignal wird über R2 und C6 gegen Masse Kurzgeschlossen.
C1 sollte wohl an Pin 2 des TL972.

while I completely understand German, would you please add a link to the schematic you are referencing to?
Edit: OK, I'm Bavarian
 
BTO acoustic pipeline

Hi,

One of the users of TeensBat approached us earlier this week. He was wondering if the wav-files of the detector could be directly loaded into the BTO-acoustic pipeline:
https://www.bto.org/our-science/projects/bto-acoustic-pipeline

This pipeline allows people (not for all countries I think) to upload a wav-file and analyse it for bat-species (and other wildlife). For this it needs at least a timestamp embedded in the GUANO-section of the wav-file or a proper-formatted filename with a timestamp. Our filenames are currently hexadecimal-coded(seconds) to make them unique and this seemed a good test for the implementation of the GUANO standard that we will share in the upcoming update. After a few tweaks in the code (our GUANO wasnt yet 100% according to the specs) the files saved from TeensyBat were valid to be used with this pipeline. This includes also the GPS-data (if a GPS is implemented).

regards
Cor
 
Last edited:
I just stopped working on improving GPS reception, I seem to have aquired a small fortune of GPS units over the last weeks. And only two out of nine different pcs worked well enough.

I somehow even fried an audioboard in the process and discovered an extra cap I placed on the audioboard had more effect than I realized.

Anyway, the two best GPS units are both 18x18mm antennes fixed directly on the GPS unit circuitboard. This eliminates an antennawire that can pick up signals and helps to put the gps receiver as far way form the Teensy as possible. I still do recommend to place a metal shield over the Teensy and since the GPS module uses lot of current it is smart to be able to switch off the unit. The metal shield I used is a simple U shape that leaves open ends for the USB connector and micro sd-card.

The GPS module I told you about earlier works well enough (Beitian BN180) but it seems a Ublox (Neo)M8 with the antenna on the PCB works even better. The downside of that one is that the PCB is a bit longer, the antenna is thicker and the connector just a bit larger so it hits the battery. I was only able to put it in place with the connector removed and the wires soldered to the pcb.
Since I solder the shield to the mainboard, I only need to attach the VDD and TDX wire to the board.
This UbloxM8 is sold with a logo HGLTECH and description High Quality 8M M8N GPS Module. Where I do occasionally lose GPS reception the Beitian this Ublox keeps receiving but for the ease of mounting the Beitian is better.

Oh, I did mention an extra cap helped on the audioboard. With the replacement Teensybat I buit after damageing my old audiobard I did not replace a capacitor I soldered over the 3.3V and GND.
In my earlier experiments this did not seem to help but I now discovered a tiny bit of noise upon the writing of data to the micro SD-card.

After placing a capacitor the noise seems completely gone. You can easily place a cap afterwards, it fits on the 3 pins behind the 3,5mm jack. (GND VOL 3.3V)

Place a low ESR electrolyte or tantalum cap here is you also see some low frequency clicks in your recordings. Anything form 22-100uF would probably be just fine.

Edwin
 
Hello Edwin,
where is the GPS connected and is the code in the hex file in the current Teensybat1.0final.hex active.
 
Helge, if you look a few pages back you will find post 688 or 687 which has some info on the GPS. I am only using the power (5v) ,GND and txd Line of the GPS module.

The GPS uses a lot of curent so I use a switch to turn it on when I want GPS data.

The GPS part is not in the 1.0 version. We are still working on on the code for some other extra features.

You also had some questions about the microphone amplifier. The trick with the 4.7nF capacitor and 390 ohm resistor attenuate the power frequencies. It creates a high pass filter which keeps low frequency out of the recordings (like human voices) ultrasound is not that much affected but higher frequencies are ampilified a bit more to counteract the power sensitivity of the micro phone on higher frequencies.

Yes you can use te non inverting input of the opamp but I don't think the idea you suggested will work. You need some more modifications. But it should work fine as if is en the schematic is fine.
 
Back
Top