I'd like to share some experience in using cards larger than 32G. I record at 384kHz or 96kHz. I have 4 audiomoths, one has a 64G card, three have 128G cards. Furthermore I have a 200G card for long trips. It was 80% full after the last trip). All my cards are sandisk SDXC. I have nearly filled all 128G card, and over run one 128G card. After getting full it created names to the FAT table, but the file lengths were 0B, which was fine. One set of batteries lasted the whole 8 night recording. I have formatted the cards to FAT32 in linux (Ubuntu) using gparted - takes a few seconds. This may be possible also in the linux bash available in the latest Win 10 updates later than 1709. (I have not quite tried it yet.)
There have been a couple of cases where some single recordings were missed when I had set 0 second interval, and recording of 300 or 600 seconds. Now I tend to set an interval of 2-5 seconds between the long recordings. These missed cases number less than 10 in over 1T data. One of my 119G recording failed, because I did not remember to reformat the card!
That was a user's fault. No files written what so ever. It could be useful if audiomoth could warn about a wrongly formatted card during the setup.
One thing to watch in any SD cards are the gold contacts. When you start to see dark lines in the contacts suggesting wear your card, a failure of your card may be immenent. I have had this happen on smaller cards (thrice in 16G/32G cards) in other devices. Audiomoth appears to handle the cards quite gently. I have been quite happy with large cards in the audiomoths.
Thanks for this. I will try some larger cards. I have been sticking to 32 Gb with a duty cycle of 15s record, 30s sleep for 4/5 nights recording about 5 hours a night.
Alex, I have comments about the non-trivialness of the triggering issue, but I will throw those comments in to the triggering thread in a few days with some images.
Great. I'm working on a version that will skip sections of the WAV file where the amplitude (or some other signal) is below a threshold. The WAV file will only include the active sections if checked in a normal application, but it will also store the length of the skipped periods so that they can automatically be expanded again. It should make recordings for bats much much smaller.
702 files, last 38 files of these 0 KB but showed a file name indicating time. 384kHz sampling, 295 s recoding, break of 5 seconds file size ~220MB. The bad files were mostly at the end: good ones, four bad, one ok, and 34 bad files. All good files were of the same size. Written amount on the card 139 GB (150 225 135 936 bytes). Quite a few bats and good rat recordings!
So far the largest number of files in one card is 476, but in a couple of days I should have my hands on a slightly larger set. Your suggested file name (ISO standard like) is very good (the old one was fine for mee too).
Hi Harry, Thanks for writing up your experiences. We're currently experimenting with a firmware update that will make the writing to SD card a little more robust - you can pull the SD card and put it back in again and the recorder will start normally on the next scheduled recording. Also looking at long file names so we can have yyyymmdd_hhmmss formated filenames. This requires the firmware to check for filename clashes so potentially slows down the file opening process. We're collecting some data to see what this means for the new firmware, but with the current firmware you probably need 2-3 second gaps between recording periods to allow the file to close and the next one to open. How many files do you end up with on the large SD cards? Alex