Can CR batteries handle Wireless Programming?

[joelucid]

I would be curious how well a CR battery can handle the burst transmission required by wireless programming though

Actually quite well. I can easily do a full wireless install using a moderately fresh CR2032 using a RFM69HW with lowest power setting (most time spent in RX anyway) in under 5 seconds. The key is to always monitor cell voltage and downgrade the protocol if it gets too low. I have two approaches: When voltage hits the first limit I wait after each kb for the voltage to recover before requesting the next. When I can't even receive 1kb before voltage breaks down I switch to packet mode (42 bytes at a time, recover voltage between requests) - which is dog slow but works until the very end.

Still somewhat open is the question if such a sustained high load will reduce long term capacity of the battery. My current hypothesis is not enough to worry. The reduced capacity of these  batteries under load seems to be primarily caused by hitting the cutoff voltage quickly under load. If you give the cell time to recover that fully reverses the voltage drop for small pulse durations according to the literature. But nobody has measured 3 second 25mA loads yet AFAIK - so experience will tell.

[JamesBryant]

I am very much a newbie on this forum, but when I use small batteries for low standby and occasional large peaks I add a supercapacitor so that the peaks come from it, rather than from the battery.

[joelucid]

Yeah, I tested supercaps for that purpose, also. Obviously they add to the size of the solution. Plus we found a couple of good papers that indicate that average current counts. A coin cell will last at least nearly as long on a super low duty cycle pulsed load as on a constant current load with the same average current. This thread has the details: https://lowpowerlab.com/forum/index.php/topic,1340.0.html

If that is actually true you'd be better off without the supercap because of the charging / charge redistribution losses with supercaps. Provided that voltage during spikes is high enough to power your sensors. And spikes are short enough that they don't draw down voltage beyond the initial drop.

Joe

[TomWS]

Yeah, I tested supercaps for that purpose, also. Obviously they add to the size of the solution. Plus we found a couple of good papers that indicate that average current counts. A coin cell will last at least nearly as long on a super low duty cycle pulsed load as on a constant current load with the same average current. This thread has the details: https://lowpowerlab.com/forum/index.php/topic,1340.0.html

If that is actually true you'd be better off without the supercap because of the charging / charge redistribution losses with supercaps. Provided that voltage during spikes is high enough to power your sensors. And spikes are short enough that they don't draw down voltage beyond the initial drop.

Joe

Adding to what Joe said, if you do use SuperCaps with coin cells (or maybe any other battery for that matter), trickle charge the supercap from the battery and rely entirely on the supercap to supply power to your radio (and other heavy loads) for the duration of the 'wake' time.  If the cap isn't big enough to sustain you during wake then you either fight a losing battle or have to have a very sophisticated charging/load switch arrangement.  DAMHIKT!

The tricky part is providing enough power at first turn on to be able to initialize the heavy load devices and put them to sleep.  It is very much a chicken and egg problem.  Example, suppose you want to trickle charge the supercap with no more than 0.5mA, if the radio powers on trying to draw 16mA, you'll never get enough charge into the supercap to get the radio to a programmable voltage.

Tom

[JamesBryant]

As batteries age they tend to have higher resistance. A new battery and a supercapacitor will probably run the initialisation and an old battery should not need to (God willing!)