One of the DS18Bx’s best qualities is that you can daisy chain multiple units hundreds of feet off a single digital input, which is not your particular use case.
So if you’re not actively trying to free up a spare analog input, I’d keep things simple and avoid the additional library if I were you.
Thanks again for publishing this great little project - I’ve enjoyed watching the progress and rigorous discussion!
The only challenge with the DS18Bx is the bit-banged library can be finicky. You could always use Maxim DS2482 I2C to 1Wire bridge with @rickkas7 code for reliability. I do agree with @epbwhiz that having a 1Wire bus allows for easy expansion of sensors and GPIO if you need it.
The intent of the switch is for long term shipment / storage. Normal operations when installed is for the switch to be turned on and left on. This is why I simplified the fancy push on / push off approach from before, I did not think it was worth the complexity.
Hmm, is this different on the Argon / Xenon than the Boron? I do not have experience with these in solar uses cases. I know the Boron can handle the 6V panel, can you expand on why Xenon / Argon would be different?
Thanks,
Chip
@chipmc, the Boron has a PMIC onboard whereas the Argon and the Xenon only have basic charge controllers.
Ugh, ok looks like a 6V Zener diode is required then. Will waste some perfectly good electrons for the Boron but don’t want to fry the other devices.
Thanks, Chip
I use 6.2V reverse biased Zener Diode (5 Watt) for Xenons. The 6V Panels that I’ve tested rarely reach 6.0 V, but this was only a 1W panel and the Xenon never Sleeps.
But Larger panels easily hit the 6.2V during full sunshine with the small load from a Xenon.
On a Xenon - Once the Li-Po voltage is above 4.05V, any disturbance in the Solar Panel Output ( a Cloud) stops that charge Cycle. A new recharge Cycle wont begin until the Li-Po is less than 4.05V.
Can you route the Zener to a Jumper that allows the user to remove for Borons ?
Sure, routing to a jumper sounds like a good idea. Couple thoughts, questions:
Did you put a resistor in series when you added the Zener across the DC-In and Ground? I understand this is typically required to limit the current passing through the diode but, is this required for a solar panel in which the voltage will collapse quickly?
A Jumper would be easy enough, do you think a solder jumper would be acceptable or perhaps a switch?
If we can accept the losses from the Zener protection, it becomes possible to support the EN pin shutdown functionality for applications where the device is run without a battery. I would do this using a simple TI SPDT electronic switch. Valuable?
Thanks,
Chip
#1
No Sir, just the Large 6.2V Zener from Solar V+ (which is Vusb) to Ground.
It's only active when the Voltage approaches 6.2V.
As you suspect, it takes very little current through the Zener to reduce the Panel voltage below 6.2V
For instance, below are my results from a Voltaic Panel (6V, 2W), full sun.
As you can see, the actual max was 1.2W during these particular Solar Conditions for the 2W panel.
But we only need to waste ~ 50 mA to bring the Operating Voltage to below 6.2V.
Please note that I can't promise this is the correct method (without the current limiting resistor).
I assumed since the Goal is to clamp the Max Voltage no matter what, that the resistor defeats the purpose.
#2
I personally think a solder jumper is fine. We wouldn't be swapping a particular carrier board between Borons and Xenons very often in real life.
#3
I don't completely follow. If that's a parasitic load, how much are we talking ?
If someone has a project that was solar powered and the environment was too hot or cold for a Li-Po, a pair of series supercaps on the Li-Po connector could be used and your existing design would operate as normal for this very limited Use-Case ...... correct?
Thank you for sharing your data. So, if there is no series resistor, the Zener has little / no effect on the circuit until DC-IN goes over 6.1V. If you agree, we don’t even need the jumper - agree? If we can do without the series, resistor it opens up an interesting solution.
But first, this safety note: If we do away with the series resistor and someone connects a DC power supply that goes over 6V, it will look like a short from DC-IN to ground until something gives up the smoke (likely the zener). For this reason, I am tempted to pick a wattage for the Zener that is closer to 1.2A (I use a 3.5W 6v Voltaic panel).
OK, with that out of the way, I think we could support three different power use cases:
Battery - Only - perhaps with a big LiSOCL2 to enable an insanely long battery life (see my post on one year battery life).
Solar Panel and Battery - with the Zener diode protection in place.
5V / 1A DC Power Supply only.
Recall, the circuitry that enables the RTC to control the Enable pin is powered by the battery. To support the DC Power Supply only approach, I would need to add some circuitry to power this circuitry from either input (Li+ line in Sheet 3/3 in schematics above).
This is this a big deal? Well, I use a finite state machine approach to organize my code. There is an Error State and if the program gets there, I have a (small but growing) number of steps I will take to get things working again: soft reset, SLEEP_MODE_SOFTPOWEROFF, resetting the Particle session). I think it would be useful to add one more - remote power off. This would simply be setting an alarm for 10 seconds in the future and enabling power off sleep. This option is there in the current design but only if you have a battery connected.
Make sense?
My understanding (and testing) is when only using the 5 Watt 6.2V reverse biased Zener Diode, it has no impact on the circuit until it starts conducting current at 6.2V.
The Solder Jumper would not be bridged (default for Boron) unless the End User wants a Xenon or Argon on your Carrier Board. We'd solder the Zener's bridge for Xenon or Argon.
We want the higher voltages with Borons.
In regards to your Safety Notice...... would it make more sense to produce the Carrier Board with the Zener bridge already connected. Boron Users would cut the Trace to remove the Zener from the circuit. Could always reconnect bridge/trace for a Xenon in the future.
Make sense - one small wrinkle. If we allow for the trace to be cut and the zener protection removed, I can’t include the switching circuit as that part is only able to tolerate 6V. I guess for folks operating this device without a battery, power cycling the device is not as big a deal?
Will put some more thought into this.
Chip
Hmmmm… Looks like Scope Creep again 
Looking back, the original goal was Solar Powered Boron + Li-Po.
If you were to re-focus the Board Design for that (no Xenons, etc), then you don’t have to make compromises ? Just thinking out loud.
Just to go back to question that started the good discussion above....
I understand the need for the switch - but I don't think the lipo will charge in this configuration ?
Sorry, misread your question. When closed, the TPS22810 allows for current to flow in both directions. I am using this part in all my solar devices today with no issues around charging.
Chip
Perfect - thanks, I did not pick that up in the data sheet.
OK, going for final here. I have added the Zener diode over-voltage protection and opted for the solder joint normally closed. Take one more look and unless I have screwed something up, I am going to order 2-3 of these to be manufactured and assembled at MacroFab. Once I get them back and tested, I will see how many folks want to jump on a higher volume run. A run of 50 would cost about $40 each - minus the through hole headers.
Thanks, Chip
All,
I have ordered a few boards from OshPark in order to validate the design with swift service (5 days). I will update this thread when the boards come in and I assemble and test them. If anyone wants to order them, I have shared the project on OshPark
https://oshpark.com/shared_projects/XvOc0YfQ
Thanks,
Chip
@chipmc, are you able to create a 3D version of the board (with parts)? I use KiCad and it allows that to be done in the tool itself.
I know that Eagle does this by tying to a sister product (Fusion 360). Let me take a look. If anyone has experience with this, please chime in.
Thanks, Chip
I’ve exported from Eagle to Fusion360, and it works pretty well, assuming you have good 3D models of all your parts. If you don’t have a model for a particular part you get a nice little flat block instead sitting on your board. 
I couldn’t find more than a couple of models for the parts I use, and didn’t want to spend the time to create them. It WAS very useful though to have the 3D model in Fusion 360 when I was designing the enclosure to make sure everything fit and looked just right.
