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

[WhiteHare]

I connected the NCP301 to a load switch (see attached photo), and voilà!  It prevents the Moteino from being booted until the voltage on the supercap reaches the threshold voltage (in this case, 1.88v).

Another nice thing about the circuit is that it would appear to also function as a brown-out detector, which is nice because I'm running with the atmega328p's BOD disabled.  Moreover, it's a BOD with a favorable twist: rather than just issuing a reset command, it powers down the Moteino and won't let it boot again until there's enough voltage.

[perky]

Good to have some hysteresis on voltage level triggered resets, the voltage can jump up when the current stops on reset due to impedances in the power supply (like battery internal resistance).
Mark.

[pjeran]

Check out this part from TI - TPS3806J20.  Two separate voltage monitors one where you can set the High and Low with three resistors.  I might build up a board with this and a simple LDO + Si Diode as solar power supply. 

Paul

[WhiteHare]

Check out this part from TI - TPS3806J20.  Two separate voltage monitors one where you can set the High and Low with three resistors.  I might build up a board with this and a simple LDO + Si Diode as solar power supply. 

Paul

Good find.  In its favor, you'd need only one TI chip to accomplish what might otherwise take two separate ncp301's to do.  Possible negatives: its price on digikey appears to be about 3x that of one ncp301, and its typical current drain appears to be about 3ua versus about 500na for the ncp301.  The price difference might be close to a wash though because two nc301's might require more PCB real estate, which isn't free.  Also, the TPS3806J20's lower part count implies easier and faster assembly.

For a solar app, I'm not sure that the difference in current consumption is significant.

I'm tempted to say it's six of one or a half dozen of the other.  [Edit: But in reality, it would be six of one or a dozen of the other.   ]

[WhiteHare]

Regarding Perky's schematic for an ideal diode, if someone can please verify that the following two schematics are equivalent (the first is Perky's, and the second is my attempt at Diptracing it), I'd be happy to post the Gerber's for a break-out board for it.  The main difference should only be the choice of P-Fet's.  I couldn't find a supplier for the given TSM2313CX, so I'm hoping (?) that substituting SSM3J338R will work just as well.

[Edit: replacing .jpg of diptrace schematic with a PDF, which should be easier to read.]

[perky]

Damn it, I designed that in last year when they were generally available and now they've already gone obsolete

I originally chose that because it had a relatively low RDSon quoted at 1.8V Vgs, low gate leakage of 100nA at 25deg C, and 1uA zero gate drain leakage again at 25 deg C. Drain leakage wasn't too much of a problem in my application but I had to make sure gate leakage wasn't too high. Leakage currents rise exponetially with temperature so at 55 deg C they can be an order of magnitude more.

Your SSM3J338R should fine, it has low RDSon at a very low 1.8V Vgs and that's what you need in that circuit to get the response, you want the opamp to drive the FET to the on state as quickly as possible. A lot of these FETs also show a rapid large rise in RDSon well before the minimum specified Vgsth is reached, that compromises operation down at 1.8V levels.

In my projects I'm going to replace the TSM2313CX with DMG2305UX I think, it's a little worse for RDSon at 1.8V but does have low gate leakage.

BTW I can't read the jpg of your equivalent circuit, could you repost?

Mark.

[WhiteHare]

BTW I can't read the jpg of your equivalent circuit, could you repost?

Done.  I just now edited the prior post and replaced the .jpg of the schematic with a PDF.  That should be much easier to read.

Attached are the Gerbers and drilling code in a zip file for the breakout boards.  Oshpark.com price is $1.55 for quantity 3.

Regarding the BOM, both resistors and the 100nf cap are 0805's.  The 47uF cap is a 1210.

[perky]

Nice to see you doing this. It looks good, connectivity is correct.

BTW I based this circuit on the link below, that uses a TS1001 opamp that has a max working voltage of 2.5V and I needed it to work up to 3.2V for 2 AA cells, so I replaced the opamp with a higher voltage type, and the BSH205 FET (which has a max RDSon of 600mR at 1.8V and too high) for a lower RDSon type. Note this application data is for solar cell rectification

http://www.edn.com/design/analog/4368525/Use-a-self-powered-op-amp-to-create-a-low-leakage-rectifier

Mark.

[WhiteHare]

In my projects I'm going to replace the TSM2313CX with DMG2305UX I think, it's a little worse for RDSon at 1.8V but does have low gate leakage.

In case it's of interest to you, I notice that Mouser stocks a newer version of of that chip with a lower RDSon ( http://www.mouser.com/ProductDetail/Diodes-Incorporated/DMG2305UXQ-7/?qs=sGAEpiMZZMshyDBzk1%2fWi7D7EaJfF%252bz4cXrr%2fkc9njzVEA2%2fjI28CA%3d%3d ) than the older versions.  So far, Digikey only stocks the older versions.

[WhiteHare]

We all get more potential leverage if we're using the same platform, so I re-did the breakout board to specifically use the DMG2305UX.  Attached.

[perky]

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.

[WhiteHare]

Today, I prototyped the attached charge termination circuit,  hooked it up to a 5v mini solar panel, and tested it.  Not shown is the Moteino, which was powered by the supercap through the minimum voltage boot circuit discussed earlier in the thread above.  Even though it was an overcast day, the supercap charged-up from zero volts and then oscillated between the high and low of the hysteresis.  The Moteino reported the loaded voltages about every 5 minutes once it cleared the startup voltage threshold.  I haven't yet dialed in the high-end of the hysteresis setting, but for a first-attempt I'm happy with the outcome.   

As you can see, the circuit is quite simple.

[perky]

Good work WhiteHare! Are you going to prototype the ideal diode as well?

[WhiteHare]

For the ideal diode, I ordered the breakout board that I posted above from OshPark.  I'll put it together after it arrives, which should be in about 2 weeks.

[WhiteHare]

Unfortunately, after the sun went down, the leakage current on the prototype circuit proved to be severe.  Tomorrow I'll try a higher value for the pull-up resistor and see whether that makes any difference.