I haven’t done the math for your particular battery pack design, but if you are designing a system that is only powered by an external battery it does need to be able to supply the 1600 mA to 1800 mA to power the peak load of the modem. This can be a problem with, for example, with some LiSOCl2 battery packs.
It is also possible to disconnect the LiPo by disabling the BATFET. It should not discharge in that case. You then have an external circuit that applies power to VIN. This could be a another battery (switched), mains power, whatever. The system should boot up in SEMI_AUTOMATIC mode so the modem will not be immediately enabled. Your firmware then reenables the BATFET, then the LiPo can supply the instantaneous current needs for the modem.
Your code then monitors the power good value periodically. If the power is no longer good, the external power supply has been turned off, at which point your code does whatever cleanup is necessary, then disables the BATFET again. The system will now be completely turned off and disconnected from the LiPo.
Quick Question for clarification: Is the intent to not power the device when the door is closed?
I'm assuming a person would open the enclosure door to initiate the Alarm Sequence & Electron ?
If my assumption is correct, what's wrong with just using 4xAA Energizer L91 Primaries and the inline switch to power the Asset Tracker Input Terminals ?
The L91 Primaries have an OCV of 1.8V, but this immediately drops to 1.65V at load.
When the voltage drops to 1.4V per cell, it's time to start publishing a Battery Warning.
When using (4) Energizer L91's, your "design" operating range is 5.6V to 6.6V to the Asset Tracker Input (5V-12V Spec).
With fresh batteries @ 6.6V, you would be "slightly" Out-of-Spec for the 6V neopixel rating, but you'll see a voltage drop below 6.6V when all the gadgets are powered during an Emergency Event.
The 4xAA L91 pack's rating would be in excess of 4.5 Wh, and can provide the 2 Amp current for the Electron's spikes.
Again, my assumption is that you only want to power your device when the door is open. If not, then obviously this wont help.
Yes, I only want to power the device when the door is open. and yes I’m connecting them to the asset tracker input terminals with an in-line switch (see my last schematic above)
The 3 batteries are Energizer Ultimate Lithium AA
(Max Discharge Current: 3.0A (Continuous), 5.0A (Pulse))
4 of these at 1.8v is 7.2v, too high for the Neopixel strip, that’s why I’ve gone to a 3-pack.
I’m trying to separate the 5+v stuff from the electron’s 3v standard. The schematic above uses power from the 3-pack for the LED’s and servo, with just the digital signals coming from the electron.
I’m sorry guys, the earlier schematic had a mistake from what I intended, I still had the LEDs going to the electron for ground. My intention was to go directly to the asset tracker battery terminals.
Those are the L91 batteries I'm talking about.
1.8V is the Open Circuit Voltage for those batteries.
This will immediately drop to 1.65V per cell the instant the switch is closed for your system.
I'm simply suggesting to re-consider a 4xAA pack verses the 3xAA.
Here's my graph from the same L91 Energizer Ultimate Lithium AA, in a 3 Series pack:
As you can see, the usable capacity terminates at 4.2V for 3xAA L91's (or 1.4V per cell).
But 4.2V is already way lower than the minimum input for the Asset Tracker Terminal (5V).
So you will leave much of the battery capacity on the table if using 3xAA.
Thus the reason to consider moving up to a 4xAA pack, which gives you an actual operating range of 5.6V to 6.6V. The pack still has life at 5.6V, but it's time to start making arrangements for a battery swap. You will never measure a voltage over 6.6V with a fresh set of batteries and the power switch closed with these devices connected.
The advantage of 4xAA is a pack with a much longer useful life, the disadvantage is the possibility of running the neopixel out-of-spec with new batteries. I would think it's worth a test
I guess the main difference is I wouldn't be concerned about the 1.8V open circuit voltage of these L91's since you will never have a condition that doesn't represent a load when the door opens (switch closes), so the max Voltage becomes 1.65V per cell.
I was running 4 cells, but bought the 3 cell holders when I thought the Neopixels were at risk.
I’ll just have to un-solder the 3 pack and re-solder the 4 pack.
Other than the 3-4 cell pack, do you see anything else in my schematic I need to worry about?
I changed back to the 4-cell battery pack and though it’s working, there is an unusual behavior I’ve not seen before.
I’m running one of the included effects in WS2812FX.h , (FX_MODE_CIRCUS_COMBUSTUS)
What’s weird is I’m getting periodic bright flashes that concern me. They are very brief like a camera flash and possibly at random intervals. I’ve set the brightness at 10-20 and the flashes seem to be full brightness.
After noticing the above, I verified my pixel count by subtracting 1 and having one pixel that wouldn’t light up.
I then started cycling my servo back and forth, rotating to 25 deg after 5 seconds, then back to 0 at 10 seconds and resetting the timer. With the servo moving the flashing wasn’t happening.
Do you think the flashes could have been excess stored voltage getting released that the servo is now draining?
They were working fine when I was powering them from the Electron itself, but had mistakenly assumed the VBAT pin was 3.3v from the LiPo.
I suspect that’s what damaged my previous Electron board (top of this thread)
I then decided to separate the LEDs from the Electron. I started by converting to a 3 cell battery holder and posted my plan and schematic. Based on the feedback above from Rftop I converted the system back to the 4-cell battery setup.
Just tested the LED strip powering from the 3v3 and gnd on the Electron board. Seemed to work fine (brightness at 20 roughly 10%?).
There was a concern above about max current if the LEDs are on full brightness
You can power the Neopixels from the battery voltage and not via the 3.3v pin on the MCU since it can not handle the max Neopixel current at full brightness.
That’s how I ran a 100 neopixel strip by powering it directly from the battery that was also powering the MCU.
I would Power the Electron from the Li+ pin and the Neopixels via the + side of the battery. That way all the current for the Neopixels come from the battery and does not run through the Electron terminals.
Then testing showed the neopixel doesn't like the Voltage.
Did you measure the voltage during the unusual behavior, by chance?
It's probably best to use the 3-cell pack, like you planned previously.
The Asset Tracker V2 lists the min Input voltage at 5V, but in reality, it may not be much different than the Electron ( Electron Datasheet = input voltage range on VIN pin is 3.9VDC to 12VDC).
If the Asset Tracker will function down to 4V on the screw Terminal or Vin , then a 3-cell is the better choice.
That would be your latest drawing/schematic.
3xAA L91 Energizer Ultimate Lithium
5.4 V Open Circuit Voltage - fresh batteries
5.0 V Voltage under load - fresh batteries
4.2 V Publish Battery Warning - batteries near EOL
4.0 V Batteries Dead
The last thing I need is to eliminate the Electron’s Lipo and use the shield’s auxiliary power terminals only so when the cover switch is open the system has no power at all for long term remote deployment. Once the switch is made the system needs to power up and have full functionality (cellular, GPS and cloud)
I tried this but had no functionality at all after disabling the charging, even with the LiPo still attached: Topic
Will this work and also allow me to get the system power status once powered up?
( see the green circle below)
Nope, no reason to connect the Asset Tracker Terminals to the Electron's Li-Po connector as shown in the Green Circle above.
I would think this should work fine:
You are sending 5.4V to the Asset Tracker screw terminals only.... no Li-Po involved.
I don't know of any reason to mess with the PMIC either, since no Li-Po will be connected/used.
Maybe I'm missing something?
How & What did you have connected ?
BTW for folks reading this in the future, this isn't a LiOn pack (3xAA L91 Energizer Ultimate Lithium)..... They are primary batteries, not a rechargeable chemistry.
The problem is, the Electron will not operate without the LiPo attached even if there is adequate power supplied to the shield power terminals. I believe this is an intentional behavior (to protect the battery from draining too low maybe?)
According to the Asset Tracker V2 datasheet, it “should” run on aux power without the LiPo, but this doesn’t seem to be the case.
With only the external battery connected to the shield power terminals the Electron will not run, it boots up but will not run.
