Bat detector

Ok, I resorted to open the case and press the button, so that part is OK

I successfully uploaded the hex to the Teensy and will try it tonight !
Nice splashscreen but a bit too short to read I think

More when I have had time to use it.
Thanks
 
Just loaded the new development version on my two teensy detectors. Very nice splash screen at startup, but it does disappear a bit too quickly.

On my second detector, when powered on (with a previous software version) it didn't get past the splash screen at all. I loaded the new software and the same thing happens. It stays stuck at the splash screen.

It was working OK the last time I used it (some months ago), so I'm not sure what is wrong with it. Any clues?
 
Hi

We can easily change the startup-time, but ... its a tool so we dont want to stare at that beautifull photo too long :) Edwin has asked the photographer (www.paulvanhoof.nl) if we could use one of his images for this purpose. Since we aim to make this "ultrasonic world" available for many he agreed.

It might be interesting to compile for the detector with trouble a debug-enabled version to see where it hangs. As it happens to that machine its most likely a difference in hardware.
I will compile a debug_version shortly and allow you to see what that tells you.
 
The debug_enabled version is uploaded. This will send messages over the USB_serial directly after startup (normal serial settings with 9600 baudrate). If you can show what thrtat reports for
the problematic detector we can hopefully solve where it hangs. Have you tried to start that detector with the leftbutton pressed at startup ? Thats a "full reset"
 
This is the output on my machine:
Code:
********** START DEBUGGER ************                                          
SET GAIN:MIC                                                                    
   gain: 30                                                                     
EEPROM LOADED                                                                   
SDCARD FOUND                                                                    
   RAW/WAV files: 107                                                           
LEFTBASEMENU 10                                                                 
RIGHTBASEMENU 4                                                                 
 SETMENU0  10                                                                   
 SETMENU1  8                                                                    
 SETMENU2  8                                                                    
 SETMENU3  11                                                                   
***SETUP AUDIO***                                                               
 DETECTOR SampleRate: 281000                                                    
 inMixer from (gain=1) Mic                                                      
 outMixer from  heterodyne                                                      
 CHANGE DETECTOR: TExpansion                                                    
 outMixer from  granular
 
Many thanks for the debug version...I'll load that in a minute and see what output I get.

I tried the left button at startup and got the 'resetting to defaults' message. Still hangs at the splash screen. I went back to version firmwarecorbee_v100.hex to try that. It gets past the splash screen and displays the waterfall....but nothing else works...no buttons, no display of audio on the waterfall...:(
 
We can probably move this over a PM thread not to share all the debug with others.

Looking at the debug output, I have come to the conclusion it is a hardware issue with the SD Card reader on the Teensy. Without the SD card inserted, everything works OK (except for record!).

If I make any progress in resolving this issue, I will post back here.
 
Looking at the debug output, I have come to the conclusion it is a hardware issue with the SD Card reader on the Teensy. Without the SD card inserted, everything works OK (except for record!).

If I make any progress in resolving this issue, I will post back here.

Thanks for sharing the results of our testing with the forum, I hope you can resolve this matter.

regards
Cor
 
I had a similar issue. Yesterday I powered op a detector that I left on autorecord the night before. I had almost 700 recordings. Inserting an empty card (OR removing the 700 files card) solved the hanging in the bootscreen.

Maybe a card check and message (insert empty SD-card or remover card) could be a solution.

Kind regards,

Edwin
 
I also see something about the pressing of the button on the teensy for programming.

Teensy also has a reboot exe, if you look at my google drive there is file called updateXP this contains all files and a small instruction for updating using windows XP or more recent versions.

It has a driver (already built in, in win10) Upload tool and reboot/reset/restart tool (forgot the actual name)

This way you can load a hex file without touching the teensy board.

Edwin
 
Man so many messages in a short time.

I had to read back more messages that expected to see a bunch of questions from David.

The questions about the signal strength are quite difficult, how much Pa do you create being a bat shouting quite loud on 20mtrs distance..... I have no idea, translating this to millivolts and decibels will be hard since we do not even know where to start with.

The easy approach is looking at the opamp which has a 18:1 resistor ratio which means the voltage amplification should be 18 times. 2 millivolts of input should become 63 millivolts output.
With the caps in the circuit amplification is less on low frequencies and more on higher frequencies. It should translate to about 16dB on 40kHz and 20dB on 100kHz (-30db on 1kHz)

The ICS microphone is supposed to be the microphone with the best s/n ratio according to other builders, you can use it in a balanced configuration. The single ended output is a bit less than the SPU if I recall correctly but measured over the positive and negative output I guess it should give more signal.
I am not a big fan of the ICS element since it is less sensitive around 50kHz (a 15dB "gap" somewhere between 40 and 60) but you can see the species of bat well enough in the recordings. The simple test I did, did not show much better results on high frequencies, I think the SPU does perform better.

I also tried to use and FG element but these do seem to be quite insensitive compared to the ICS and SPU elements, so you need to amplify the signal more to get a similar signal strength.

The third question was about the LC filters used.
There has not been done any calculation on these filters. The LC filter kind off has a "knee" in the filter curve which you can calculate and it helps you to determine what values are high (or low) enough to filter out some unwanted noise.
The higher the capacity the lower the frequencies are that are attenuated, the higher the resistor is, the bigger the attenuation will be. You still do need to have enough voltage for the circuit to work so that kind of limits the value of the resistors used.
Since the microphone and amplifier use only very little current it is no problem to use a higher value resistor. The other parts use more current so it is better to use only just a few ohms so there is less voltage drop.

The first attempts of recreating Cor's ideas showed a whole lot of noise on the display and in the recordings. I just tried a lot of filtering in different ways. To be honest I don't know how much these LC filters actually help to reduce the unwanted noise because they cost little I never tried to remove them from the circuit.

Edwin
 
I had a similar issue. Yesterday I powered op a detector that I left on autorecord the night before. I had almost 700 recordings. Inserting an empty card (OR removing the 700 files card) solved the hanging in the bootscreen.

Maybe a card check and message (insert empty SD-card or remover card) could be a solution.

Kind regards,

Edwin

HI all,

This message of Edwin actually was pointing at a flaw in the updated code, not one you will encounter fast however. Since we allow users to play all the recordings that are present on an SDcard we also need to know the names (at least thats easy). In the latest update (october 2020) we reserved space for 1000 filenames, this however costs a large chunck of RAM memory. In the update we are (still) working on this space has been reduced to 500 filenames. And Edwins issue is related to this. He had recorded over 700 files and at startup the system would simply hang. Turns out that one line of code used the wrong logic to test if more than 500 filenames were allready read and writing into the wrong memory (over the edge of an array) always creates mayhem. Last evening we have cleared this issue so in the real update this will not give problems.

cheers
Cor
 
Many thanks Edwin, those answers make a lot of sense and I hopefully understand better how everything works. The comment about the sound pressure generated by a bat was just to highlight how impossible it is to work from the mic sensitivity figure. (Although I was not expecting an answer, it is an interesting question which I sure some scientists have studied.)

It is very reasonable to try to reduce all sources of noise, as long as the cost (in money and power consumption etc) are not too high. Has physical shielding around the audio components been considered and is it likely to be of benefit?

Hopefully I will have my detector working soon and I can start playing, although the weather is looking very cold for the next few days here, possibly with show. If I was a bat I would be staying somewhere warm

--
David
 
Questionaire: default startup settings

Hi,

We (edwin and I) are finishing up the last bit of work on the update. One of the things we would like to hear from active users in this forum is the default settings they tend to use (and save using the EEprom).
What are your values for:

VOLUME (0..90)
GAIN (0..63)
DETECTOR (heterodyne, auto-heterodyne, Time_expansion, Freq_divider)
DISPLAY (none, spectrum, waterfall)

regards
Cor
 
Hi Cor,

Not much luck yesterday night as temperatures took a drop from previous days: only two calls recorded during the evening...
And it's not going to be better for some time.

My settings as they are now:
Volume = 50
Gain = 20
Detector = A-Heterodyne (at start, then, maybe switch to TE but not always. I never use the others two)
Display = waterfall almost exclusively, sometimes spectrum, never none

Regards,
Thierry
 
Short answer ... yes. But you would probably need at least 3 microphones to detect the direction properly ...with 2 you cannot triangulate.
 
Volume 55
Gain 24
TE
Waterfall
Samplerate 176k (better use of display width with most common species)
Recordings 352 (281 should be enough, 384 trends to be a bit more noisy)
Playback 32 (about 1/10 th of 352)

Edwin
 
Hi,

The files recorded are stored with a name like 606C9BFB.wav
I understand that there might be a limitation related to the 8.3 filename standard, but if not, would it be useful to have the option to name the files according to the date and time of recording ?
For example: 20210331_212446.wav
In my opinion (which is worth very little :p) it would facilitate file management on the computer and also some uses require that naming convention (for example the BTO Acoustic Pipeline)

What do you think ?
 
Hi Thierry,

After the upcoming update I will see if I can start to incorporate the update of sdFat that will allow the usage of large filenames. Until we are sure that this is possible we are locked into the 8.3 format and (see messages in the forum) there are easy ways of converting these filenames to YYMMDD_HHMMSS.wav as the current HEX name is a direct conversion of the time in seconds.

But be aware that things are more complicated. If you want to store and play up 500 files from an SDcard and be able to directly use them on your detector the length of the filename is going to cost you RAM memory. Your proposal for YYYYMMDD_HHMMSS.wav will thus cost twice as much memory as the current setup. RAM is limited on our T3.6 so its something to think about before acting.

The BTO acoustic pipeline does not need filenames like this as long as they use the GUANO specification (which we allready do), this was tested by the people running the BTO pipeline allready.

cheers
Cor
 
Last edited:
Hi Cor,
Thanks for the very quick and extensive answer (I'm beginning to wonder whether you are waiting in front of your screen, just in case someone has a question ;))

Well, it doesn't seem like a good idea then: no use to waste RAM on this, especially if there is a better way.
Strange: I tried to upload some files to the BTO pipeline, but they were croppings from complete files that I made via Audacity. Maybe the Guano data got lost in the process.
I'll check again.

Regards
Thierry
 
Hi,

The guano is part of the latest pre-release update (which you allready had downloaded), the guano will stamp with a time for instance and more variables.
If this doesnt work I would like to know since we have been testing this after a request for it arrived. We then shared a few testfiles with the people from the BTO and they confirmed that
the files (hexnamed) where passing without problems.

cheers
Cor
P.S. I tend to work often on the laptop and also get notifications :)
 
Back
Top