A solar supercap powered Moteino (15Farad charged by BQ25504)

[perky]

All good work WhiteHare.
Or possibly:

Code Select Expand

solar_panel-> LDO +-> schottky_diode-> 470uF ->  ideal_diode  +-> MCU (BOD enabled)
                  |                                           |
                  +-> schottky_diode->   5F  ->  ideal_diode -+

Edit: You'd need resistors obviously to charge up the two caps to different voltages.
Mark.

[perky]

Nice PCBs (never new OSH PARK was that cheap, need to look at UK delivery )

Mark.

[ChemE]

Not sure of their quality yet, but PCBs.IO is even cheaper.

[WhiteHare]

Not sure of their quality yet, but PCBs.IO is even cheaper.

Matter of fact, the PCB for the one black colored breakout board in the earlier photo was manufactured by PCBs.IO. 

I have no complaints about the PCB quality.

[WhiteHare]

All good work WhiteHare.
Or possibly:

Code Select Expand

solar_panel-> LDO +-> schottky_diode-> 470uF ->  ideal_diode  +-> MCU (BOD enabled)
                  |                                           |
                  +-> schottky_diode->   5F  ->  ideal_diode -+

Edit: You'd need resistors obviously to charge up the two caps to different voltages.
Mark.

Thanks for the idea!  I started to play around with it, with an eye toward keeping the part count low.  The attached schematic, or something similar, might actually work as a "good enough" compromise, and it has only 4 more parts than the "super simple" version.

[Edit: I haven't dialed in the exact resistor value yet, but the two dollar 10F supercap together with a 470 ohm resistor (instead of 270 ohm) seems actually to work pretty well. ]

[perky]

Yeah, disadvantage is there is two diode drops from LDO voltage to the Moteino via the 5F cap. What you could do is put another schottky in series with the one on 470uF cap and dial up the LDO voltage by 2 diode drops to compensate. But you have to watch their forward voltages at very low currents, we've been over that one before

Mark.

Edit: Strike that, the 2.7V limit is on the supercap not the Moteino :-O

[WhiteHare]

Interestingly, the resistor on the supercap means that the supercap's charge current is a lot less than the solar panel can supply, which by itself isn't good, but I believe it also means that a much smaller solar panel of the same voltage would charge just as well using the same circuit.

Anyhow, I like the "instant-on" Moteino capability that the small cap in this circuit allows, which makes the longer charge time on the supercap not such a palpable annoyance.

I haven't thought it through, but maybe this is a case where using the earlier "5v Perky Supercap", together with a higher regulated voltage of around 3.6v or so, would be an advantage, because that way one wouldn't run into the very long taper charge tail caused by the Schottky diode until later, after the voltage on the supercap was already decently high.  I can't say off the top of my head if it would actually make a difference versus just using a higher Farad 2.7v supercap, but maybe. 

[WhiteHare]

So, rethinking it from that angle, the attached new schematic might work pretty well. 

Of course, depending on personal taste and how long you're willing to wait, one could also relax the "instant-on" requirement to merely "faster-on," which would allow lower ohms on the resistor and higher current charging on the supercap.

[perky]

I wonder if you could use a constant current source for the 5F charge resistor (made with a bipolar transitor, FET and 2 resistors). It'll have about 0.65V drop, but it would change it's resitance from low when the voltage difference is low to high when the voltage difference is higher.

Here's one I use.

[WhiteHare]

Sounds promising.  I don't know enough to properly evaluate how that design would perform over the entire voltage range, but I wouldn't mind ordering the parts and building it.  Then we'd know for sure.

[WhiteHare]

@Perky
Since the tsm2313cx p-channel mosfet isn't stocked by either Digikey or Mouser, does it matter at all which particular substitute p-channel mosfet I use in your constant current circuit?  i.e. any parameters in particular that I should look for?

[Edit1:  I put together the notional v002 schematic on a breadboard and then hooked up a mini 5.5v solar panel and a Moteino. Honestly, using just that, the Perky cap charges fast enough as-is from even indirect sunlight that it's already of practical use, and I confirmed that I do get instant-on with the Moteino from the small cap.  So, I'm pretty happy.     However, if using 2.7v supercaps, the constant current circuit may still be worth looking into, at least as an option.]

[Edit2: Afrotechmods has a constant current circuit that uses a LM317T plus two capacitors and a resistor:
https://youtu.be/iuMngik0GR8

Plugging it into our our circuit, though, I rather suspect that at least one of those extra capacitors could be eliminated.  However, again, I'm not sure how well or poorly it might perform over the entire voltage range.  On the other hand, maybe it doesn't need to.  I already have all the parts for the Afrotechmods' circuit, so I may try putting that together also. ]

[WhiteHare]

Actually the DMG2305UX appears as you say to be available with two part numbers, with a -7 or -13 in it. They point to the same datsheet, and the Diodes website doesn't have these -7 or -13 references. The Q parts are just automotive qualified parts but appear also to have the same specs. Curious.

Anyway, there seems to be a better version with a lower RDSon of 113mR at Vgs of 1.8V called the DMP2305U-7 which is also widely available, so I'll use that instead:

DMP2305U: https://www.diodes.com/assets/Datasheets/ds31737.pdf
DMG2305UX: https://www.diodes.com/assets/Datasheets/DMG2305UX.pdf

Mark.

Since the tsm2313cx is the same out-of-stock part that we had previously discussed, I'll just assume the same replacement part you picked in this quotation (DMP2305U-7) is an equally good substitute in your constant current circuit:  https://www.digikey.com/product-detail/en/diodes-incorporated/DMP2305U-7/DMP2305UDICT-ND/2052810

I didn't pick the 1.8v LDO used in the notional schematic v002 very carefully, but I ordered this one:  http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail&itemSeq=225694980&uq=636281315387995756
if only because it has a mere 40mv voltage drop.  No mention of the quiescent current in the datasheet (usually not a good sign), so I may have to revisit that.  If anyone reading this has any favorite 1.8v LDO's that they like, please do name names by posting.

Order placed!  There's a good chance I'll receive it by this weekend.

[Edit:  Also, the solar panel probably shouldn't be more than about 5v, not 6.5v as I had in the notional schematic v002.  That's because the Perky Cap is rated at 5v, and the voltage drop across the Schottky Diode is pretty small near the end of the taper charge. ]

[perky]

I hope you meant 3V LDO, or 2.7V, not 1.8V (or was that the minimum working voltage, you're Digikey link is broken)

Mark.

[WhiteHare]

Hopefully this link works better:  https://www.digikey.com/product-detail/en/texas-instruments/TPS73218DBVR/296-41770-1-ND/5224553
I really did mean 1.8v though.  Do you think it may be cutting it too close?  Perhaps not enough margin when the radio transmits?  How about 1.85v or 1.9v?  What number would be safe?

The reason I went with the 1.8v number is that a at least some LDO's have quite high quiescent currents if their input voltage is less than their rated output voltage.  I seem to recall that the Moteino's LDO might be one of them (?).  Not sure if it applies to all LDO's or not.  Since a Moteino (with regulator removed) can allegedly operate at 1.8v, I went with a low output voltage to avoid that problem as much as possible. If the Moteino still has its original LDO installed, then that's the LDO that should be used, not the 1.8v LDO shown in the notional schematic v003 (attached).  i.e. in that scenario, the 1.8v LDO would be removed from the schematic entirely.

[perky]

Er, 1.8V LDO means the regulation voltage is 1.8V and the output won't ever go above 1.8V, so you'll never charge anything up! What you need is a 2.7V, or possibly 3V (2.7V plus a diode drop) output LDO i.e.one whose graph shows the output voltage follows the input voltage until it reaches regulation, then if the input increases above that the output won't go the regulation voltage. The working voltage is the minimum that the input has to be before the output will start following it.

Mark.