Home › Forums › Mooshimeter Support › Dead batteries just after real logging gig
- This topic has 0 replies, 1 voice, and was last updated 6 years, 9 months ago by admin.
Last time i saw it (monday), app told that battery level is ~50 (percent? as it starts from 100). Now (wed/thu), when i could not connect and led was not blinking, battery voltages were ~1,10V both. Had not used meter for ~50 hours in between..
Meter was logging for sunday-monday night at my dads house*, log file had two extra headers in the middle (write errors/reboots/???), card did not mount automatically on ubuntu and dmesg told:123456789101112131415161718[2289534.046401] usb 3-3.2: new high-speed USB device number 28 using xhci_hcd[2289534.154513] usb 3-3.2: New USB device found, idVendor=05e3, idProduct=0736[2289534.154516] usb 3-3.2: New USB device strings: Mfr=3, Product=4, SerialNumber=2[2289534.154517] usb 3-3.2: Product: USB Storage[2289534.154518] usb 3-3.2: Manufacturer: Generic[2289534.154519] usb 3-3.2: SerialNumber: 000000000272[2289534.155452] usb-storage 3-3.2:1.0: USB Mass Storage device detected[2289534.155581] scsi host28: usb-storage 3-3.2:1.0[2289535.155423] scsi 28:0:0:0: Direct-Access Generic STORAGE DEVICE 0272 PQ: 0 ANSI: 0[2289535.155683] sd 28:0:0:0: Attached scsi generic sg4 type 0[2289535.345902] sd 28:0:0:0: [sdd] 15564800 512-byte logical blocks: (7.97 GB/7.42 GiB)[2289535.346731] sd 28:0:0:0: [sdd] Write Protect is off[2289535.346733] sd 28:0:0:0: [sdd] Mode Sense: 0b 00 00 08[2289535.347546] sd 28:0:0:0: [sdd] No Caching mode page found[2289535.347548] sd 28:0:0:0: [sdd] Assuming drive cache: write through[2289535.350964] sdd: sdd1[2289535.353484] sd 28:0:0:0: [sdd] Attached SCSI removable disk
but mounted just fine (in read only mode) when sudoed.
Before i started writing here, i formatted that card (and forget to save that log to my computer, still have it on my dads computer though).
Would ~empty batteries cause these symptoms and how accurately that battery level indicator code has been calibrated? Would it be possible to predict how much logging time there is, with current battery voltage and selected logging options when we start to log? Its probably impossible to say accurately, but even “over a day/week/month worth of batteries (and memory card) left” would ease the need of babysitting our loggers..
* – Found out that ~30y old freezer and fridge are pieces of shit.. Freezer shut down only once in a 20 hour log, fridge on the other hand, took 41 cycles in that time (which might be just fine, sadly i was logging only current+voltage, inside temp might have been nice). I calculated that those together would use ~1900kWh in a year. Same size new ones promise something like 350kWh/y. Savings would be about 15 euros in a month.. New ones would pay for themselves in six years just in electricity :D
I’m really happy you got good data out despite the logging issue! Yes, new appliances pay for themselves quite fast :)
Regarding the meat of the question: If the meter keeps trying to log as the battery dies, I think it’s possible this could cause the filesystem corruption you saw. The SD card draws the most power of anything on the meter when it is writing, so my hunch is that when the batteries are near death a write event might cause brown-out. I have some code in the firmware to prevent this problem in normal use but it’s not tested against all SD cards and circumstances (if the batteries are cold, they sag more, stuff like that).
Regarding battery levels: the indicator in the app is quite rough: 100% means 3.0V or more, 0% means 2.0V or less, and it’s linear in between. The meter shuts off around 2V.
Estimating battery life while logging: What was the logging interval set to? The longer the logging interval, the better the battery life. I am going to change the default so it’s no longer 1s, because this is actually quite battery intensive. Since the ADC needs to settle for at least 2 seconds before giving a good reading, a 1s logging interval means the ADC never turns off and draws quite a bit of power. Also, was the meter in the fridge/freezer when performing this logging? Cold temperatures also reduce the amount of power available from batteries (this is reversible when the batteries warm up though).
Hope this helps with diagnosis, I’d love to see logs if they are available :)
Hi James (and anyone else, who is willing to read all this nonsense ;))
http://imgur.com/a/nNPPw <– Thats all i have for now, probably can’t get to dads computer in little over week (and he actually don’t know how to use it :D ) for original log file. Do you really want to see it?
Mooshimeter was of course outside of fridge, in a wooden cabinet on top of it (maybe 25-30 deg c, somewhat over room temperature as those appliances were converting all that electricity to heat :D ).
I made a plastic box for current transformer and meter itself and cut an extension cord (with three outlets, which was nice in this case, so i just plugged both devices to it to log at same time) to go through it. I’m a bit paranoid for leaving any experiment in a state where someone could hurt themselves.. And all i had was uninsulated banana plugs, as i want to leave that box as easy to use again as possible, so could not use real measuring leads, which i would want to use elsewhere. I ordered safety plugs and will do that box setup again, but nicer and safer.
In (most of?) europe, we cant even tell which lead is neutral and which is phase, as our schuko-plugs will go to outlets in two ways. And for that reason, we cant even be sure that the contacts in wall outlets are in right order, as some electricians “know” that it makes no difference, so even marking my plug for right orientation wont do any good.. So i need safety plugs for all contacts, not just for live. And this time i used current transformer, which is meant to be used inside an electric cabinet, so it really does not have consumer level shields over it.
As for the battery estimations, i was on 10 sec intervals. I was hoping that the app itself would have some understanding of the situation.. So it would be able to tell that
a) you would have over week of log time left, no matter what you do,
b) couple (interval/range dependent?) levels in between
c) and a warning that “you should not log with these batteries, you are getting 5 hours at most and probably screw your memory card while doing so”
It may be hard for user to estimate anything from that percent value for battery, as we are seeing same kind of values with our phone batteries, which calibrate constantly and will go to zero about every day etc.. Not my oneplus-2 though, little use in good reception while on vacation, i can go 2-3 days.. But its still far away from mooshimeter and usage is so different, so mentally estimating wont work..
Very cool! I added the logging time estimation issue as an enhancement on github. It’s a good idea, I hope to get to it soon.
I was just curious to see the data… I’m always interested where meters end up and how they are used.
We have the same issue with hot and neutral sometimes being switched with plugs in the US even though officially the issue is solved. You reminded me of a very funny story about its development here (if you’re into nerd literature): http://amasci.com/amateur/whygnd.html