💬 The Harvester: ultimate power supply for the Raybeacon DK
openhardware.io last edited by openhardware.io
If you haven't already, you may want to consider spv1040 instead of spv1050. From what I read previously, spv1050 is aimed more for outdoor light harvesting, whereas spv1040 can still thrive with less available energy indoors than the spv1050 can.
Anyhow, either way, I look forward to hearing how your design pans out, as well as whatever you might learn along the way! Sounds like you'll be having a lot of fun with this, so, please do keep us posted as you start making progress.
Hi @neverdie, thanks for your comment! Well, from what I've found in datasheets the SPV1050 was designed for ultra-low power apps and is very efficient in low light conditions, when SPV1040 is capable to gain up to 3W and its step-up converter is more suitable for outdoors. I've also considered bq25504 and LTC3103, but found that they seem isolate battery from the output where the load is expected to be. In contrast, the SPV1050 implements very simple approach by connecting the store and the battery via a FET - this allowed to connect main load parallel to the battery (where by fact it is) and disable integrated LDOs. Another advantage of the SPV1050 is that it's omnivorous - can accept input up to 18V (although I've configured it for 12V) and officially supports TEGs.
But to be honest, never tried anything from the list so will eventually post my findings So far the Raybeacon was finally assembled and sent to Nordic for tuning - you may have noticed updated pics on its page. So hope to advance the project status from preview to beta by the end of the year.
@mishka I was going from memory, so I may have mixed the two of them up. Sounds like you're on top of it, which is great.
@Mishka Have you been able to test the spv1050 yet, and if so, how did it go?
One of the things I like about the spv1050 is that on paper it has a buck-boost configuration with a very wide input range of 0.15v to 18v. It seems that most chips that can work with as little as 0.15v as input are boost only, which makes them not as versatile as the spv1050.
The cold boot minimum voltage for the buck-boost configuration is 2.6v, however, so you'd want it to default to boost-only mode if it runs out of juice. In boost-only mode, the cold-boot voltage would be 0.5v.
Which leads to a quandary: how does one know in a cold-boot scenario what the input voltage would be? In theory it could be quite high (if, say, you walked into a dark room and flipped on a light switch), but if you set it conservatively for that kind of scenario, you risk missing start-up opportunities where lower voltages are available and would have been enough had boot-only been the default.
@NeverDie From my understanding it's a game of the harvesting source. In boost configuration V_in must be lower than the end of charge V_eoc value which limits possible input to low voltage sources only, such as a TEG or some PV panels.
On the other hand for buck-boost configuration there is the cold-boot voltage indeed. It's also somewhat unclear how it will maintain V_store when V_in later drops below 2.6V.
I've ordered number of KXOB25_02X8F panels to give it a try with one, two, and three "solar bits" in series. The Harvester board is configured to V_oc = 12V so it's also interesting to check how well this will work in a single panel configuration. The panels and boards are somewhere on the way (China Post), hope to build everything in early February.
@Mishka Please do let us know how your testing goes once you get your parts together.
I've tried "solar bits" of the type you linked. FWIW, I found they perform quite well with sunlight but surprisingly poorly on indoor LED lighting (where, say, one of the keychain scavenged solar cells seems to perform relatively better). Just FYI, depending on what light source you are planning to use. In that regard, perhaps there's a further dependency on the type of indoor LED one has. I have no data on that, as I've only tried what I currently have installed, which are from CREE.
@NeverDie Thanks for the info! I've never used IXYS cells before - chose them mostly for the size. The harvesting IC and the SoC are very low power so I hope there's will be a chance to reach a usable configuration indoors. Anyhow, will share my measurements, of course.
@NeverDie Just received the KXOB25-02X8F panels. In short - you was right, and I was too optimistic about indoor lights.
Generally speaking, the pannel works really not bad and looks extremely handy. However, in order to achieve the desired 2.6V it has to be illuminated with greatly above 1000 lux which roughly equals bright supermarket or factory lights. And yes, it seem differentiates LED lamps from the sun.
Accordingly to my (totally uncalibrated) luxmeter:
- Living room room: 50 lx - 0.4 V
- Overcast daylight, indoors: 250 lx - 1.4 V
- Supermarket lights: 1300 - 2.0V
- Overcast daylight, street: 2000 lx - 2.7V
Therefore, it's safe to say that two of 02X8F would make it acceptably work in buck-boost configuration. By acceptable I mean it could charge an ML2032 battery, but batteryless operation is quite limited indeed.
This pushes me to review use-cases for the Harvester. It has two input sources - first one is spv1050, and another one is MIC5205 3.2V LDO. I'm thinking about switching the spv1050 to boost configuration and cover only PV and TEG sources from 75 mV to 3.6V. At the same time LDO could handle sources capable to deliver 2.0V and above (2.0-3.2V with battery removed, and 3.2V to 16V with or without a battery). The point here is that at higher voltages there is likely a more capable source so MPPT might not be required.
Such, although less flexible, one or two KXOB25-05X3F, 30mW each, and a "supercap" should be enough to power the nRF52 (and even an extra module) without the need of a battery. For more demanding applications, up to five KXOB25-14X1F could be used delivering 150mW in total. With 3.2V, 70mA limit on the spv1050 battery pin it sounds like a good match.
Another option is to make the Harvester configurable via solder jumpers. This will put four more jumpers on top of it (and there are five on the back side already). Well, although possible, sounds too complicated as for one inch board.
P.S. I'm going to assembly a buck/boost board anyway so we can compare later.
@Mishka Well, since you ask, I think the EM8500 would be interesting to try: https://www.emmicroelectronic.com/sites/default/files/products/datasheets/8500-ds.pdf
I only recently discovered it, but the datasheet says it can self-start with an input voltage of 300mv and a mere 3 microwatts. Once started, the datasheet says it can continue functioning on as little as 100mv and 1 microwatt. It also claims to include MPPT.
The other one that would be interesting to try would be the EM8900, which can self start with an input voltage as low as 5mv, which is, AFAIK, the lowest of any available commercial chip.
However, for a solar cell as small as your solar bit, it would require extra circuitry to operate it in burst mode, because at 5mv the micro-ampere requirements would likely exceed what your solar bit could deliver on a continuous basis. It's the same issue as with the LTC3108: https://www.openhardware.io/view/732/Extreme-Energy-Harvester
I haven't yet tried a solar cell scavenged from a solar calculator, but that might be interesting to try as well, since presumably those are well designed to work with indoor lighting and have been perfected over decades for that use since the 1970's. Unfortunately, I don't know of any that can be purchased outright instead of as part of a solar calculator, like say the FX-260 or similar. Maybe someone reading this knows of a source?
Though it's cheating, the last option would be to use a long-lived button battery purely to avoid cold-boot scenarios and to manage the collection of real solar power during those times. I imagine that such a battery could be quite small if limited to that type of use. Some such batteries might last as long as 40 years, such as Tadiran. That said, if it were deprived of light for long enough, it would probably run the battery down sooner than desired, so I don't like this option.
@NeverDie Thanks for the hint! At first glance, EM8500 looks like a highly integrated solution for embedded devices.
What I like: very efficient for tiny, highly embedded applications, built-in USB charger (cool), can complement primary battery (super cool), programmable (including MPPT and battery thresholds) with EPROM support.
Unfortunately, for the same reasons it will be very hard to use it as an addon module. First of all, it requires the battery to be exclusively attached to the harvester IC. In contrast, spv1050 has much simpler circuit where STORE on the module can be conned to battery on the main board.
Next, ultra low-power capabilities of the EM8500 come at cost of its upper limits - 1.8V max for harvesting source, 20mA for LDO. Choosing PV for such device would be even harder than satisfy 3.2V requirement for boost-only spv1050. Since nRF52 may consume about 20mA at 3V, it might be tricky to run it batteryless, especially when some other devices are sitting on the line.
Also, I2C bus will occupy two of six I/O. Well, fair enough. The spv1050 also has status lines for battery charge and battery connection. But they are completely optional and are off by default. Broadly speaking, the spv1050 is completely transparent to its load.
So my impression is that the IC is very cool indeed, but it just doesn't fit this particular project. At least in its current form.
@NeverDie Use additional battery is an interesting idea. My only concern is that the harvester IC will drain it if connected to input. On the other hand it looks like the IC will work just fine as long the STORE has some voltage. And the STORE is connected to battery, which is charged by the harvester IC - i.e. the rechargeable battery is exactly the battery used to prevent IC from shutting down. But perhaps I missed something - still have no boards for test.
Regarding solar cells from a calculator - yeah, that was my first intent, but have no solar calculator out there. Taking in account that such a calculator costs three times less than this SolarBIT thing I have no excuse for not to buy one from a local store - going to fix that tomorrow
Okay, grabbed cheapest calculator from a store. It turned out to be EATES DC-837, priced below $4. Surprisingly well made, it uses monocrystalline PV cell and one AAA battery, connected via a dual diode. This means the calculator should be able to work from as low as 1V, and the PV panel should not cause overvoltage at the same time.
The panel size is 12.6x37.8 mm. It looks really inexpensive and is made of glass, protected with paint on the back.
Despite that, the panel works damn good. Such, I was able to get 1.8V from it in a room with dim lights (about 50 lux). At the same time, short circuit current was about 5.9 µA. Exposed to bright sun it resulted in 3.2V open circuit, and 3 mA at short circuit. I reached limit with 1200 lumens torch at 5 cm distance - Voc=3.25V and Isc=7.2 mA.
I did similar measurements for KXOB25-02X8F. This time my torch forced the panel to Voc=5V and Isc=11mA which somewhat differs from datasheet. Results are on the charts below:
@Mishka Looks very promising!
Thanks for posting the charts and the photo. I tried googling the "SC-1338-4" but found no datasheet, no sellers, nothing. I'm guessing the SC probably stands for "solar cell". Is it by any chance 13mm x 38mm in size? I'm only guessing, but perhaps the 4 is a reference to 4v, which is perhaps the voltage at maximum solar illumination. In that case, what appears to be a part number ("SC-1338-4") is actually just a generic description of the part, not a true manufacturer part number.
The only other clue as to which solar cell is might be is perhaps a barely visible logo in your photo of the cell. The logo looks like the mirror image of "CAT" with a kind of triangle pushing up the bottom of the "A" letter.
By the way, it looks like an interesting source for truly tiny scavenged solar cells might be solar watches:
That solar cell looks tinier than any that I've ever seen for sale.
@NeverDie You're absolutely right. The cell size is 12.6x37.8 mm which explains 1338. I also agree with your guess on 4V. The CAT is just reflection of my phone.
Well, I'm impressed that a noname manufacturer managed to make such a useful cell. BTW calculator consumption is roughly 5 µW - pretty neat too
Regarding solar watches - they should be fine to charge battery, but I'm in doubt they have enough capacity to power on anything so advanced like nRF52 - I hope should be feasible with the SCNE.
Grr, the boards are out of radar - last track is 2019 December, 29.
That solar cell looks tinier than any that I've ever seen for sale.
Not sure where to go with this though
Maybe the reason solar cells like the SC-1338-4 are so elusive to find in the usual parts market is that the manufacturer is basically just buying a relatively cheap 5x5 or 6x6 inch slice of monocrystal that probably looks like:
and then carefully breaking it into bits of such a size that when wired together they exactly fit whatever net size the buyer wants. Though I've never tried it, I presume that you or I could be doing very small scale "manufacturing" ourselves if we were so inclined. Obviously some people make their own DIY 100w panels this way, but from what I've seen those people do so by soldering foil tapes across the front face of the monocrystalline slices, which presumably these cells would be too small to do. Hmmmm...., this now has me curious as to exactly how the bits get wired together. Anyway, provided they're not too hard to make, the upside to DIY'ding small solar cell planels would be getting any size/voltage combination we want plus freedom from having to scavenge pre-made cells from pricier consumer products.
@Mishka Looks as though an array of BPW34 can actually be packaged together fairly tightly.
It still wastes some real estate, but not as badly as what I had imagined.
I suspect they wouldn't do well under indoor LED lighting though, as they seem to have peak sensitivity at around 900nm, which is infrared.