Capacity calculation for battery devices

[sergey]

Hi everyone!  Having mired in confusion with capacity calculation of popular 18650 batteries, i decided to ask society . On what we can expect purchasing 18650 battery on Ali and using them in Reporter devices?

my supposal is that : 

having MppReporter device that sleeps all time with consumption 30ua , and wakes up every 10 minutes for 10 sec in wifi-off mode with consumption 20ma plus 12 sec wifi-on mode with consumption 80ma we've got about 1,7ma per hour of discharging a battery.

20ma mode - 20ma*(10/3600)*6(cycles per hour)= 0.33ma\h

80ma mode- 80ma*(12/3600)*5(cycles per hour)=1.33 ma\h

0.03ma mode - 0.03ma*(1 -10/3600*6-10/3600*5)=0.0389 ma\h

As many forums says real capacity of chinese 18650 battery  that claimed 6000mah doesn't surpass 1500mah. Moreover voltage sustainability (ability to keep at least 3V) is somewhere at level of 70-75% total capacity.

Therefore my conclusion that we can expect on the Reporter work not longer than 620 hour or about 26 days w\o charging.

Does everything seems to be ok in my calculations?  Can anybody advise a good battery ?

[MikeP 4 MPP]

Looks about right.  I believe there's also a period of the wifi using 250ma when it's connecting before it hits steady state of 70/80ma, but I'm not sure how long it lasts - hopefully not until the connection is established which could be a few seconds.

Every 10 minutes does seem a bit often...  Is it possible to decide onboard and only connect to wifi when it's necessary?  The bulk of the power use is the wifi connection, if you were able to reduce that count to every 20min you'd double your battery life.

[sergey]

Yep, the wifi using 250ma lasts not longer  than 1sec or less , that i even can't see it on my testing equipment. I decided to substitute it by using 80 ma instead of 72ma in my calculation.

Sure i could set up 20 minutes instead  of 10min , but it doesn't solve the problem completely.  Moreover - increasing the sparsity among connections leads to incrementation self discharging current ...anyway. I can't reduce the measurement frequency - it doesn't make a sence...  At the same time darkness (when we don't see the Sun)  lasts long in my area  , for instance, about three month, there is no power for battery charging..

No any problem was observed in summer time, even weak sunlight could maintain the device..  I was pretty sure that chinese battery being claimed 10000mah to work in reality turns to be much fewer but i didn't expected that it might be so little.. 

I recall you mentioned that alcaline battery runs much longer life ?

[MikeP 4 MPP]

I do find alkaline batteries last longer than the rechargeable ones and tend to run at 1.7V or so rather than the rechargeable at 1.4V, but of course they can't be recharged. 

What I was suggesting was only wake for wifi if there was a change significant enough to report back as that uses so much more power.  You could continue to wake every 10m to measure without wifi.  TBH I don't really understand what could change weather-wise or soil moisture-wise in an hour let alone a 10m - I measure and report every 8hrs and it seems enough for watering.  But of course I don't fully understand your design...  Tough to fight physics but if you could find a way to reduce the wifi check to once per hour you might get 5x or 6x the battery life.

One trick I've used to reduce network traffic for power measurement devices is to report more often when there's big changes, but once the change amount slows I reduce the poll interval.

The battery devices I'm using don't wake often - MppSleepers only wake when they're used, MppAnalogSensors only wake every 8hrs - that gives pretty good battery life.  If you need faster readings you may need to switch to using zigbee (I'm not sure how much of a reduction that would be), or a separate app that accepts a group of readings maybe covering an hour or more instead of one reading at a time.

[sergey]

Anyway thanks, Mike!) 
That is a bit different project, forgot to say.  In this project we also measure the humidity,temperature and pressure , but we need a bit quicker reaction on figures changing that is way 10 minutes it is even a bit longer than we really need. Unfortunately every 10 minutes we typically observing quite significant figures changing and i can't pass the report ...

Rechargeable battery works absolutely fine , when the sun is...  What about approximate capacity of two alkaline batteries till the level of 1.5V do you think ?

Have you started any new project with Zegbee ?)))

[MikeP 4 MPP]

I find my ESP devices working pretty reliably at anything above 2.5.  My rechargeable batteries seem to drop quickly down to 1.3-1.4 so sit around 2.7v for few months. When they finally drop off it's pretty quick to 2.5.  I haven't looked at the akalines too close since I usually use rechargeables, but they do seem to last longer.

Looks like alkaline AA's have about 2500mah, half that for NiMH.  I realized recently that using two AA's doubles the voltage but not the energy (doh!) but you could use them up as 2 pairs of 2 to last longer.  I'm pretty sure batteries will lose more energy over time than is drawn but I'm not sure how much so that's a bit tricky too.

I've played a little with zigbee but decided it wasn't worth it (too expensive), the protocol was too rigid/opaque, and they don't seem to have anything like an ESP8266 yet.  But it's my understanding zigbee is often picked for multi-node environmental measuring.  I suspect it's because zigbee will connect to the hub more quickly to report - while tx power is similar to wifi, connected time is much less.  But you'd need a powered hub centrally located with the devices since they won't work in a mesh when running on batteries.

[sergey]

Unfortunately i can't works at 2.5V , because i have to use LDO for sensors stability. It works stable till 2.9 V, than send farewell letter and fall asleep till reset, preventing full discharge and  lost of data.  Ok, thanks , gonna try alkaline batteries.