To test this feature, visit your live site.

AudioMoth recordings fail after firmware upgrade

Hello,

I have quite a few audio moths that were working well, and I upgraded a few to the newest firmware (1.10.1). I am now having an issue where recording fails intermittently (sometimes after a few minutes, sometimes after a few hours), particularly at higher sampling frequencies (>= 92 kHz). None of the devices on the new firmware have lasted the scheduled 48 hours (recording 55-secs ON/5-sec OFF). At first I thought it was an SD card issue, but it even happens with the supported SanDisk Extreme 128GB cards (as well as other SD cards of different sizes).

I've rolled back those moths to older firmware versions, but the issue is persisting. Is there a way to "hard reset" to older firmware versions? I'm happy to provide more context or other info that you need; just let me know.

All batteries are fully charged and the SD cards are fully reformatted in exFAT before each test. It's odd, becuase when I look at the files, everything looks good for a while (again, this is variable from an hour-ish up to ~40 hours) but then a few files are written that are smaller than the "normal" ones, until it fails. There's still plenty of space on the cards (>30 GB), but it almost seems like what happens when the cards do fill up.

Thanks for any advice!

6 comments

6 Comments

This comment was deleted.

Alex Rogers

Aug 29

Replying to

Hi Liam, I think this is the SD card being full. The last file that wrote started late as the file took longer to open that normal, and the next recordings suffered SD card write errors until the AudioMoth cancelled the schedule after five failures in a row. Make sure that you reformat the cards between use, and this should solve the problem. Alex

jack

Jul 17

Sorry this has taken so long to update, but we're finally getting around to testing this. So far what we've found is that on AudioMoths running 1.10.1, the SanDisk Ultra cards caused recording failure, but in a way that looks like a battery failure (the batteries still had plenty of charge). Keeping all things the same, when we switched in a SanDisk Extreme card, the AudioMoths recorded fine for >48 hours. One interesting thing to note is that in the EXIF data that the AMs write, the battery voltage was always about 0.3V lower than what the Configure app reported, so maybe the device thinks its batteries are dying and shuts down the recording? We're still testing this (currently running a batch using the SanDisk Extremes to see if they record the "correct" battery voltage).

I appreciate all the info above, Alex!

Alex Rogers

Jul 17

Replying to

The battery voltage reported in the Config App is made whilst the AudioMoth is running from the USB supply so it will always be a little higher than the true loaded battery voltage.

Whilst recording the AudioMoth monitors the supply voltage (the voltage actually supplied to the processor after the voltage regulators) and stops the recording if that goes below 2.8V. It should always be maintained at 3V. The SanDisk Ultra card has very high write currents which can cause short-lived drops in the supply voltage (particularly as the internal resistance of the batteries increase as they are depleted) and it seems that this is causing SD card write errors which also cause the recording to be aborted.

There seem to have been a few problems experienced with SanDisk Ultra cards, so we would recommend only using SanDisk Extreme cards.

Alex Rogers

Jun 05

Hi Jack, If the files are getting smaller then this is most likely due the batteries going flat. If you generate the summary of the SD card using the 'Summarise AudioMoth Files' in the Config App it will likely show the battery voltage decreasing and then the comment field will report that the AudioMoth is stopping the recording early due to low supply voltage. It will do this five times before stopping the recording schedule to prevent the risk of damaging the SD card. Can you share this SUMMARY.CSV file. Can you confirm what SD cards you are using and what batteries (make and model of both). Also, can you post the CONFIG.TXT file showing the recording schedule. The most likely explanation is low-capacity batteries or certain models of SD card that consume significantly more energy than expected. Alex

jack

Jun 05

Replying to

Will do, Alex - it will take a few days as I'm just relaying the info from a coworker who is preparing to deploy the equipment.

I know that we are using 2800 mAh rechargeables (not sure of the brand, but will find out), but we've also been able to replicate the issue with brand new Energizer alkalines.

We've tried SanDisk Ultra (A1 rating) and SanDisk Extreme (U3, A2 rating) cards in a 64GB and 128GB sizes, as well as Lexar cards of the same ratings and sizes.

Alex Rogers

Jun 05

Replying to

No problem. The SD card summary will reveal what is happening. SanDisk Ultra may not be fast enough for high sample rates, and some Lexar cards have previously caused problems due to their high write currents. We generally use SanDisk Extreme (red and gold) as these are reliable and have low energy consumption. There's a table here - https://www.openacousticdevices.info/sd-card-guide - that shows the energy consumption of different types of cards. The Config App values are calculated for a SanDisk Extreme 32GB card as these have the lowest overall consumption. The large SanDisk Ultra 256GB card we tested had very high energy consumption - between 2 and 5 times as much as a small SanDisk Extreme 32GB card. I can measure the exact cards that you are using. This somewhat similar issue - https://www.openacousticdevices.info/support/device-support/custom-mode-issues/dl-1eae128f-63e4-4450-8723-4bfcba8372d9 - turned out to be caused by high write currents. In this case, it caused the AudioMoth to reset. It is possible that it could also trigger a low supply voltage warning. I don't think your issue is due to the firmware update - much more likely to be something with the batteries or SD card.

Edited