I have a couple of emonTH (V2, I believe),… and having run for 18months or so,. their supply voltages are now showing up as 2.1volts and 2.7volts…
I was wondering what the lower limit was on there ‘usufulness’ for reliable operation of the actual Hum., Temp. sensor was. as I can find no references in the docs,… as far as I can see…
Is there a default alarm post in emoncms for these devices going low batt. ? I assume not, unless I have missed something…
Many Tx
No there isn’t, and you assume correctly.
The operating voltage is supplied by a boost regulator, so it holds the voltage up quite well but then drops rapidly. This is a pair of white ‘Eneloop’ rechargeable cells:
and the last 24 hours in detail:
So depending on what you have in there now, it’s probable it will give up before too long - mine collapsed from 2.2 V to 1.2 V in 14 days, 1.6 V to failing in under 24 hours.
Sorry that’s not all that helpful, but very much depends on the chemistry and the discharge characteristic of the battery, and they’re all different when it comes to the details.
Ahh tx Robert,… That is useful information…
As I said I have two Temperature units,… and this is their track…
The ‘interesting’ bit is this is your ever fav. batteries,… sold as ‘Simply Duracell’,… their shelf life is until March 2027,… so well within their useful life what ever that is.
Both units were install/commissioned at the same time (from the same packet,… so one might ‘assume’ the same production run, or at least close),. but one set seems to be reaching EOL sooner than the other set,… I wondered if there was a temperature component to the one set falling of quicker than the other,… but as you can see they both follow a similar temperature profile…
But tx for the feedback
The non-transmitting sleeping currents are quite small, but over time large compared to the current taken doing useful work, so a small variation asleep would account for a large effect on the rate of discharge of the batteries.
I think I’ve read that normally, dry cells are rated down to 0.8 V for end-of-life - so 1.6 V for the two.
I might have expected the cooler one to last longer, but maybe the warmer one has more chemical activity - who knows?
I think @awjlogan has a number for the ratio of sleeping energy to ‘working’ energy…
Hi @diyhouse - thanks for the plot. The bedroom ones really drop off, looks like a defective cell, to be honest - the blue lounge curve is much more typical, as are Rob’s - a long plateau followed by a sharp drop when they’re depleted.
Typically, colder cells last longer (chemical storage, so with Arrhenius as a guide, you typically double the self discharge rate for every 10ºC raise) but they’re very close in temperature over their lifespan. Hotter cells can provide higher peak current with the same rough relationship as the reaction can run more quickly.
Re. the standby/active duty cycle, you can approximate it to a first order with a weighted average of the currents in various phases of operation. The TH2 has a standby current of approximately 60 uA for 54.995 s, an active current of about 1 mA for 1 ms, then a Tx current of 42 mA for about 4 ms. This gives a rough estimate of 4 years battery life from typical 2200 mAh AA alkalines - very much in line with observations.
So, I doubt there’s a big enough variation in the TH2 construction to account for your variation. Your readings imply 120-240 uA standby current. You can expect a little variation in the capacitor leakage, standby currents etc, but it would be in the ±5% range at most, not 2-4X. Altogether, I think your batteries probably aren’t very good, sorry!
Tx for the comments guys,… interesting,… I have just measured the individual cells in the bedroom sensor,… and they measure out as 1.063volts… and 1.072Volts so not bad…
also thought prudent to measure the lounge cells,… and got 1.4volts… and 5mv,… yes double checked,… 5mv… so I think we can say that is flat…
I am now trying to understand my graphs to see if I have miss-labelled the feeds etc,… but they seem OK
latest graph as follows:
will pick up more later,…
1 Like
Well Ok,… have been thinking on these results and graphs for a day or so,… and they just don’t add up,…
Have been back and remeasured cell voltages,… been as careful as I can to not disturb the physical/mechanical connections.
Have measured the:-
Lounge monitor cells at 1.180volts and 1.414volts sum = 2.594volts
Bedroom cells 1.050volts and 1.060volts sum = 2.11volts
The long term graph looks as follows:-
And short Term as follows:-
Looking at the new voltages in the lounge monitor,… where the total is 2.594volts,… one cell is doing decidedly better than the other,… which ‘is’ the mate of cell that was showing 5mV,… So I immediately think ‘operator’ error,… but as I said I did measure several times,… and,… graph monitoring bares this out,… so I have no idea why a cell would show close to zero,… yet a few hours later,… after,… I had played ( measured, and remade mechanical connections ),… it comes back to life.
But if we take today’s measurements as more fact,… then there is better correlation, although it is still interesting that one of the cells is doing significantly better than the others, at 1.414volt,… ( saying sqr rt 2, is too obvious ),… the other 3 , are all in the same ball park,…(120mV range).
But we shall see how their full life pans out over the next few months…
Many thanks for your feedback so far, but I leave this for the greater minds to ponder, and possibly comment.
Hi Mark - just to check, is this the same battery that you measured as 5 mV, or a fresh one?
Hi Angus,. yes,… this is the same cell that was reading 5mV…
just thought I’d see how the land (voltage) lies ATM,…
I don’t believe it??? at 3:00am the cell went down,… so temperature monitor is now essentially running of 1 cell,…
and in case you were wondering,… I was not tampering with the Lounge monitor at 3am… 
Why a cell would exhibit such behaviour is beyond me,… but evidence would suggest it really is,…
Me thinks I need to change both cells,… unless you would like to leave as is for now,… in the interests of science,… as they say,
EDIT:- added temperature to the graphs right axis.
Rgds Mark
1 Like
well the batteries have finally thrown in the towel you might say…
on the bedroom unit the cells got down to 90mV and 820mV… and transmissions failed,
whilst the Lounge unit cells got down to 1.35v and 0.2mV… but still going
Both units now have new cells,.. with a shelf life ( shown on the box ) up until March 2028.
this is the last graph track up to failure.
and from start date to end date,.
Result: 644 days
It is 644 days from the start date to the end date, end date included.
Or 1 year, 9 months, 4 days including the end date.
Or 21 months, 4 days including the end date.
Clearly there are some cell in balance,.. I can only speculate this is poor cell storage, before purchase, but who knows for sure,.. but none of the cells are particularly close to one another in output voltage. incidentally the failed cells are all marked with an end shelf life of March 2027.
One last question for the community,… is there a way to setup an alarm,.. email etc,.. that updates are no longer taking place.??
Many tx.
As I wrote at the top of this, as far as I know there’s no mechanism for tapping into the ‘time since last update’ in emonCMS.
As an alternative, you could have a play with “If mute” and “Email” processes, along with the various “If…” tests to detect the reported voltage falling below your desired threshold. I have never investigated these functions, so I can’t say whether they will do what you need. As I see it, the problem is to not send an email every minute, which is what is likely to happen as the processes are run on receipt of data from your emonTH.
Thankyou for your comments Robert… one other avenue I will look at as well is to see if HomeAssistant has any easier solutions… ( but that is probably a case of ‘just’ battling yaml code ).
I will looking to upgrade the firmware on my emonTx’s though…
thankyou all
I know its been a while,.. but as per my last post I have since added code to my HA install to monitor and report ‘failed’ sensors…
Have been waiting for the ‘right’ conditions… and following a small home DNS outage.. I noticed my HA android App reporting the following:-
The code I cobbled together to make this happen is as follows:-
- id: '17554675448585'
alias: Alert - Temperature
description: ''
triggers:
- hours: /3
trigger: time_pattern
conditions:
- condition: or
conditions:
- condition: template
value_template: '{{ ( as_timestamp(now()) - as_timestamp(states.sensor.sensor.tasfour_bme280_temperature.last_changed)
) | int > 10800 }}'
- condition: template
value_template: '{{ ( as_timestamp(now()) - as_timestamp(states.sensor.emoncms_cb014871_ae8c_48d7_a460_ddf9c6710483_92.last_changed)
) | int > 10800 }}'
actions:
- data:
message: Alert!! Stale Temperatures!
data:
channel: Alerts
group: Alerts
action: notify.mobile_app_qw1_5g
- data:
message: Alert!! Stale Temperatures!
data:
message: there's a fire brewing man..............
title: Emergency response team
action: notify.persistent_notification
mode: single
Hope this might be useful to someone…
Rgds
2 Likes