How best to find the "best" small solar panel of a particular size?
-
Charge pumps work on this principle. It's just that they dump their higher voltage from the series configuration into another capacitor, whose voltage eventually rises to the series voltage. I guess that could work too, but you need switches which switch at very low voltages to pull it off if you're starting from a cold-start and, say, a solar panel is only providing you with extremely low voltage. I'm not sure how far below 0.6v you can find switches that still function. Same for the oscillator that typically drives a charge pump. That's why Ceech's board (that I linked earlier above) that works with a startup voltage of as low as 20mv (?) is so impressive. Or is that its lowest operating voltage, and it still requires a higher start-up voltage? I guess I should read the chip spec sheet.
-
@NeverDie 20mV is stat-up voltage.
-
@ceech very impressive Roman! Keep us posted when the board is ready for orders
-
@ceech Is there a particular brand/source of supercaps that you recommend to fit your PCB? Or do they come pre-installed?
-
@NeverDie I like those:
http://www.mouser.com/ds/2/40/AVX-SCC-1018831.pdf
Super capacitors will be installed on boards.
-
I must admit, I got totally lost in the middle of this thread with so many links.
Did you manage to find a holy grail? I'm keep to change a 400mA LiPO for supercapacitors on one of my nodes
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
Did you manage to find a holy grail?
I haven't received the supercaps from China, so I can't comment on them as yet. I expect they should arrive by next weekend, or possibly before. However, there are lots of other supercaps that one might use.
I still think that for most people a simple 6v solar panel for around $1 on Aliexpress, together with https://www.openhardware.io/view/382/Tiny-Solar-Charger-for-27v-Mote-Supercap
and a sufficiently large supercap is all you're going to need for most things. If you have a sensor that needs 3.3v, then you will want to add a pass-through boost charger such as:
https://www.openhardware.io/view/285/33v-Boost-Converter-with-Pass-Through
or
https://www.openhardware.io/view/279/Adjustable-Boost-Converter-with-Pass-Through
that you enable when your node wakes up and disable before going back to sleep.One can get more elaborate to handle more challenging low-light cases, but for 80%+ of the cases, I would expect the above to be enough.
Fortunately, all this stuff is fairly inexpensive and easy to assemble.
-
I have this Solar Powered IoT Device Kit from Cypress....but have not started using it yet. though I would post it in case you might find the BOM useful.
-
@NeverDie Thanks for your suggestion. Here is a thing - I cannot find MAX8887EZK27 anywhere.
Meantime, I did manage to find this interesting article - https://www.ti5.tu-harburg.de/publications/2009/fgsn09_lifetime.pdf
And these boost converters on the e-bay:
A bit expensive, but based on a very advanced chip which may be used with supercaps.
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
@NeverDie Thanks for your suggestion. Here is a thing - I cannot find MAX8887EZK27 anywhere.
https://www.digikey.com/product-detail/en/maxim-integrated/MAX8887EZK27-/MAX8887EZK27--ND/6227255
-
@NeverDie it is a no go for me at USD 18 delivery charge to the UK. Unfortunately, there are no other sources for this chip.
-
You'll need to substitute a different LDO for it then.
-
@alexsh1 in case you don't already know, octopart is great for finding sources. https://octopart.com/search?q=MAX8887EZK27
There seems to be only american sources for this component though
-
@alexsh1 in case you don't already know, octopart is great for finding sources. https://octopart.com/search?q=MAX8887EZK27
There seems to be only american sources for this component thoughI did not know about this source
Well, that's the problem. The postage is golden and then if the value is greater than 20 bucks (including postage), I have to pay 20% VAT.
:(((((
-
@alexsh1 @NeverDie You can substitute the MAX8887 with MIC5365.
This whole thing got me so intrigued that I'm going to make a board myself. With charger, super capacitors and DC-DC converter. I'll use MIC5365 and for DC-DC conversion TPS610986. It has loads of options. And the whole thing will cost less than 10 bucks. I've made a schematic
-
@ceech said in How best to find the "best" small solar panel of a particular size?:
TPS610986
Excellent choice - TPS610986 - The TPS61098x
suits for low power systems very well, especially for the system which
spends the most of time in sleep mode and wakes up periodically to sense or transmit signals.
-
@NeverDie I am sure you saw this brilliant write up by Nick Gammon:
http://www.gammon.com.au/forum/?id=12821
He experimented with 0.47F capacitor. It was enough to run a mote (<10uA in a sleep mode) overnight
-
I hadn't seen it, but a mote in sleep mode draws practically nothing (<300na for combined draw from both atmega328p and RFM69). It's the self discharge rate of the capacitor more than anything which seems to govern.
-
@NeverDie This is what he experienced! Have a look at the link. Within a few charges though, self discharge stabilised
-
Regarding the choice of replacement LDO. If the LDO you pick turns out to have unacceptably large reverse current flow, just place the blocking diode after it instead of before it. That particular diode will leak only about 20na, and the entire issue doesn't even come up until the last ~0.3v of charging, where you'll then have a very long charge tail. However, if that were to be an issue, you could just use a larger Farad supercap, and then it wouldn't matter because you'd have enough charge already before encountering the long tail of the charge process.
-
@NeverDie Thanks for heads-up. I'll keep it in mind
I found an interesting cap -
https://www.digikey.co.uk/product-detail/en/vishay-bc-components/MAL219691204E3/4705PHBK-ND/5015889Its voltage is very close to the solar panel so I think it may be connected directly via diode (0.7v drop = 6 - 0.7 = 5.3V vs 5.6V max rating)
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
@NeverDie Thanks for heads-up. I'll keep it in mind
I found an interesting cap -
https://www.digikey.co.uk/product-detail/en/vishay-bc-components/MAL219691204E3/4705PHBK-ND/5015889Its voltage is very close to the solar panel so I think it may be connected directly via diode (0.7v drop = 6 - 0.7 = 5.3V vs 5.6V max rating)
I've tried that cap already, and I wouldn't recommend it. It has a fairly high ESR.
-
@NeverDie Given that you have been tinkering with it for a while, what caps can you recommend? With what ESR?
This cap has got 2.4 Ohm @ 1kHz - I think it is low?!
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
@NeverDie Given that you have been tinkering with it for a while, what caps can you recommend? With what ESR?
This cap has got 2.4 Ohm @ 1kHz - I think it is low?!
Perky explains it better than I can: https://lowpowerlab.com/forum/projects/a-solar-supercap-powered-moteino/msg18120/#msg18120
-
@ceech said in How best to find the "best" small solar panel of a particular size?:
TPS610986
1.5 mm x 1.5 mm WSON Package
Not sure how easy that will be to solder?
-
@NeverDie You need to get SMD rework station or oven.
This is extremely difficult to get done by even an iron with a fine tip.
-
-
@ceech
Well, I may end up buying it from ceech then. Even properly aligning it onto the solder pads becomes difficult. I really struggled with the BQ25504. Not saying it can't be done, but just that it requires a lot more skill.
-
@NeverDie and a strong magnifying glass
-
Those are the boards BTW
-
@ceech it is very small. Excellent. Let us know when it is ready
-
And if it is too tall, still, only two capacitors can be installed in horizontal position, or even just one.
-
@ceech How about a flat supercap like this:
https://www.digikey.co.uk/product-detail/en/vishay-bc-components/MAL219691252E3/4699PHBK-ND/5015883
There is no need to have two in my view unless you want to raise voltage. One flat to make the board compact may be enough.
-
Flat ones are nice. Nice package. High internal resistance, therefore low current. The double cell 15F flat capacitor can only supply 70mA.
-
This is great.
Maybe more people will start playing with this stuff if they can buy inexpensive pre-made boards from ceech. The more, the merrier.
-
@ceech said iThe double cell 15F flat capacitor can only supply 70mA.
It should be enough if not using nrf24l01+ pa lna or rfm69HW. In this case perhaps, it would be better to have 150mA at least to cover those transmitters?
I like the size - It is really flat.
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
@ceech said iThe double cell 15F flat capacitor can only supply 70mA.
It should be enough if not using nrf24l01+ pa lna or rfm69HW. In this case perhaps, it would be better to have 150mA at least to cover those transmitters?
I like the size - It is really flat.If you start to get choosy, that's when it's time to start making your own boards. It's easier than you might think. Otherwise, you just have to accept that the off-the-shelf stuff will never be exactly what you want.
It's really too bad there isn't more of a universal PCB format that's easily edited, so you can simply start with a board that's close to what you want and just customize it a bit (the way one might with, say, software). For instance, even just changing LDO's--unless the new LDO has exactly the same land pattern--can necessitate redoing the board. No problem if you're the original designer, but it's a PITA if you aren't.
-
I received the Chinese supercaps today. My "el cheapo" component tester is, I guess, too cheap, because it can't identify them:
So, if anyone has suggestions for a good capacitor tester, let me know.However, I did plug the first one into my solar node after charging it ,and, so far it's not holding it's charge very well at all. I'll try the green one next.
-
@NeverDie I just watched this video:
Measuring the Capacitance of a SuperCapacitor – 11:54
— Julian IlettIs this what you have been looking?
-
@alexsh1
I like Julien's videos, butI'd find something that's faster and more automated to be preferable, as well as something that can measure ESR too.
-
Well, so far the green supercap isn't holding its charge any better. I'll leave it in-circuit for two or three days, though, and see if that improves. It's encouraging that the rate of voltage decrease is decreasing. Some supercaps apparently improve their ability to retain charge if they are forced to hold a charge long enough, though I have no idea why that is. For instance, some capacitor leakage test protocols call for holding a capacitor at its rated voltage for 72 hours before running the leakage test.
-
@NeverDie you are right - Nick Gammon states exactly the same that overtime leakage "improves". A new cap has to be broken in before leakage is measured.
ESR measurement - I have no idea if it is possible to measure it and how. I think right now we are going down deep into the physics in order to do that
-
For more normal capacitors (though apparently not for 10F supercaps), the little gizmo I showed in the pictures above does measure and display the ESR. That's one of the main reasons I purchased it. It's probably not hard to measure either. Measure the unloaded voltage. Then measure the voltage under, say, 1 amp of current. The second voltage measurement will be less, and from that you can calculate the ESR. At least, that's how I'm guessing it's done. Except you'd near to do a near instantaneous measurement, because the longer you drain the current, the more the voltage goes down for that reason too, not just ESR.
-
Incredibly, I may have to resort to the Julien Ilett method of measuring capacitance. None of the meters or testers I've seen thus far handle supercapacitor capacitance (e.g. 10F). So, thanks for posting the video! Very timely.
-
@NeverDie the old fashioned way!
Glad we are doing or reading about the same things at the same time.How long does it take for your device to measure ESR? Did you try a normal capacitor, I.e. 100uF?
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
How long does it take for your device to measure ESR? Did you try a normal capacitor, I.e. 100uF?
Yes. About 5 seconds.
[Edit: that includes the time it also takes to measure capacitance. It does both measurements (including part type identification) at the push of the button. ]
-
Maybe it is just outside of the measurable range of the meter.
-
Is there a way to do a full disconnect of the solar panel (so as to do an unloaded voltage measurement) without resorting to a relay? I've tried using a PFET, and although it can turn off the current flowing from the solar panel, it still somehow sucks the measured voltage down to about 0.3v above the supercap's voltage. On the other hand, if I physically disconnect the solar panel, then I'm able to then read it's unloaded voltaqe.
-
I played around with it a bit more, and I found that if I used some external voltage source that was otherwise not part of the circuit to switch the P-FET "off", then it effectively accomplishes a full disconnect.
-
Found a very good on-point thread which addresses the issue I'm having: https://forum.arduino.cc/index.php?topic=115233.0 So, it may be mosfet related, not just circuit related. If worse comes to worst, I'll order the Hexfet it references that supposedly does achieve a full disconnect the way you'd expect.
-
@NeverDie said in How best to find the "best" small solar panel of a particular size?:
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
@ceech said iThe double cell 15F flat capacitor can only supply 70mA.
It should be enough if not using nrf24l01+ pa lna or rfm69HW. In this case perhaps, it would be better to have 150mA at least to cover those transmitters?
I like the size - It is really flat.If you start to get choosy, that's when it's time to start making your own boards. It's easier than you might think. Otherwise, you just have to accept that the off-the-shelf stuff will never be exactly what you want.
It's really too bad there isn't more of a universal PCB format that's easily edited, so you can simply start with a board that's close to what you want and just customize it a bit (the way one might with, say, software). For instance, even just changing LDO's--unless the new LDO has exactly the same land pattern--can necessitate redoing the board. No problem if you're the original designer, but it's a PITA if you aren't.
I do not disagree with you. I have started looking at Eagle software. Obviously, the idea is that I'm going to start with improving somebody's design first and then designing something myself
-
@ceech said in How best to find the "best" small solar panel of a particular size?:
MIC5365
I'm going to switch from the MAX8887EZK27 LDO over to: http://datasheet.sii-ic.com/en/voltage_regulator/S1313_E.pdf
which also is the same size SOT-23-5 package.Why? Current consumption. The S1313 datasheet says it should consume a maximum of 1.35ua of current, whereas the MAX8887EZK27 may be consuming as much as 85ua (though I haven't actually tried measuring it). When the overall charging current from the solar panel is only around 50ua indoors far away from the windows, I'm betting it will make a difference. Likewise the MIC5365 also appears to have a rather high quiescent current of around 29ua, at least according to its datasheet that is.
-
@NeverDie Solid choice.
-
I did the experiment of charging one of the green Chinese cheapo capacitors for 3 days at 2.7v, and then I removed it from the circuit and measured its voltage after 3 days with no charging and no load at all. It came out at 2.54v. So, I guess I would deem that use-able.
-
@NeverDie I suppose you broke it in. Try to do a few more cycles (charge / discharge) and see if self discharge improves
-
It should get a little better after some cycles
-
@NeverDie said in How best to find the "best" small solar panel of a particular size?:
I went ahead and made the switch on the node that I'm experimenting with: I'm now using a LTC3235D 3.3v passthrough boost converter. I switch it on when the node wakes up, do various measurements, then switch it off before transmitting the results if the supercap voltage is above threshold. Everything working smoothly.
I think I may next try putting a buck converter on the solar panel so I can get more juice out of it when its voltage is above 3.3v. I'll be trying this one:
https://www.openhardware.io/view/276/33v-Buck-Boost-DC-DC-Converter
because it's the only one I have on hand. If it seems to help, then I'll look for something more on-point (and less expensive!).How does LTC3235D compare to BQ25570?
BQ25570 has got buck converter with Vout from 3.3V down to 2V (Vbat is from 5.1 down to 2V), which is just enough to run nrf24l01+ with AVR.
-
@alexsh1
Well, I'm sure it's easier to solder than the BQ25570. I'm no longer really considering anything that's difficult to solder unless there's just absolutely no alternative.It looks as though @ceech sells them though already soldered to a breakout board, so maybe he can answer your question.
-
I'm playing around with a coulomb counter, actually a clone of:
https://www.sparkfun.com/products/12052The only downside to it is that it requires a minimum of 2.7v on the input current that it's measuring, which isn't really ideal for a solar panel that might nonetheless be charging at a lower voltage. It also includes its own current consumption in the measurement, and it appears to be about 75ua. In addition to counting the coulombs, you can also measure the average current between interrupts by doing a little math (which is illustrated in the sample sketche).
So, kinda cool. I changed out the sense resistor from 0.05ohm to 10ohm so that I get more frequent interrupts.
-
I found this interesting fella
https://www.tindie.com/products/closedcube/tps63030-high-efficient-buck-boost-up-to-96/
Anyone tinkered with tps63030?
-
It's kind of pricey
-
@alexsh1 I use it. On this board:
http://www.ebay.com/itm/ATmega328p-board-w-NRF24l01-socket-LTC4079-battery-charger-in-Arduino-PRO-form-/331800929918?ssPageName=STRK:MESE:ITIt works.
-
@alexsh1
Is it any better than:
https://www.openhardware.io/view/276/33v-Buck-Boost-DC-DC-Converter
?
-
@NeverDie said in How best to find the "best" small solar panel of a particular size?:
@alexsh1
Is it any better than:
https://www.openhardware.io/view/276/33v-Buck-Boost-DC-DC-Converter
?Interesting buck-boost converter, but Input voltage is from 1.8v to 5.5v.
I'm looking at the one with a better range of 0.9-5.5V
I need to use two 2.7V super caps in series with a maximum efficiency
-
@gohan said in How best to find the "best" small solar panel of a particular size?:
It's kind of pricey
Not cheap, I agree, but ideally I'm looking to use the whole range from 0 to 5.5V in two 2.7V super caps in series. If you have an idea of a good buck-boost converter outputing 3.3V, I'm all ears
-
@ceech no wonder I like your board. These buck-boost converter works really well with a solar panel
-
I already have one boost buck converter, I only need the supercap charge controller for 2 supercap or use the 5.5v 4F I have
-
@gohan what's the input voltage for your buck-boost converter?
-
-
@gohan I think it looked at it. I liked the range, but I think it is a noisy one. Did you test it with oscilloscope?
-
Nope, no oscilloscope available
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
Interesting buck-boost converter, but Input voltage is from 1.8v to 5.5v.
I'm looking at the one with a better range of 0.9-5.5V
I need to use two 2.7V super caps in series with a maximum efficiencyAccording to its datasheet (http://www.ti.com/product/TPS63030), the TPS63030 has exactly the same Input Voltage Range: 1.8 V to 5.5 V
-
There is a Maxim integrated buck-boost converter, that goes down to 0.7V
http://www.mouser.com/ds/2/256/MAX1584-MAX1585-43040.pdf
It's quite complex though
-
@NeverDie you are correct - TPS63030 has got 1.8-5.5V input voltage.
I wonder if there is a buck-boost converter with lower than 1V lower voltage?
-
@ceech you bet - personally I'd stick to something much more simple
-
@alexsh1 said in How best to find the "best" small solar panel of a particular size?:
@NeverDie you are correct - TPS63030 has got 1.8-5.5V input voltage.
I wonder if there is a buck-boost converter with lower than 1V lower voltage?
You'd probably get the biggest range by combining a separate boost converter with a separate buck converter, rather than using a single integrated buck-boost unit.
-
@NeverDie said in How best to find the "best" small solar panel of a particular size?:
Well, so far the green supercap isn't holding its charge any better. I'll leave it in-circuit for two or three days, though, and see if that improves. It's encouraging that the rate of voltage decrease is decreasing. Some supercaps apparently improve their ability to retain charge if they are forced to hold a charge long enough, though I have no idea why that is. For instance, some capacitor leakage test protocols call for holding a capacitor at its rated voltage for 72 hours before running the leakage test.
Well, try as I might, I just can't recommend these Chinese supercaps. I've held them at rated voltage for days on end, and I've power cycled them too. For whatever reason, they just don't perform as well as the branded supercaps that I've ordered from Digikey. I'll use-up the few I have on some throw-away prototype experiments, and I won't be ordering any more of them. Presently I think the 10F supercap on Digikey for $2 from Illinois Capacitor continues to be the best of the "value" capacitors. For best performance, I've been pleased with the AVX brand of supercaps.
However, if someone finds something better, either by value or by performance, please do post and let me know.
-
@NeverDie I have a few AVX caps and they are really good
-
A few boards arrived and a couple are assembled
The board charges four 1.5F 2.7V super capacitors. It also provides a boost circuit with 3.45V output on main and sub outputs. It works down to 180mV and provides 3.5mA for 30 minutes. I'll report on leakage.
There are two outputs main and sub. Main is on all the time and sub is switched with high signal on the mode connector.
-
@ceech is the output clean enough to power a nrf24?
-
@gohan if not you could add.... wait for it.... a capacitor. (Weak, I know, but I couldnt resist)
-
Unfortunately it's not that easy, it depends on the type of "noise" created by the booster so one type of capacitor is better than another, a low pass filter may be needed.
-
Noise isn't bad at all. Surprisingly. It is no bigger than the background noise when no input is connected.
-
@ceech very impressive! I really like:
It works down to 180mVthe boost to 3.45V works until the board voltage (between pin 5 and GND) goes down to 180mV?
-
@alexsh1 Yes, the boost converter works down to 180mV. In other words, it can almost completely discharge capacitors.
-
@ceech Which brand of caps you are using?
I may be interested in this board too.I am currently running your other board with two 10F caps in a test mode with a sensor. Do not have much time for it now, most likely will report in September about my findings.
-
@alexsh1 Normally I use AVX brand.
-
Here is a link to openhardware page.
-
@ceech They are very good - I ordered a few myself.
Have you looked at using a flat super caps?
http://sigma.octopart.com/17606633/image/Maxwell-Technologies-PC10.jpg
-
@gohan
If it became an issue, you could turn off the boost converter while doing the radio receptions, and then turn it back on. Unless you were in some sort of continuous receive mode, which would drain things pretty quickly anyway.
-
@alexsh1 I think they are being discontinued.
-
@ceech all of them? I mean you get the idea. As flat as possible. Size, size and size ...
-
@alexsh1
See if you can find something you like on Digikey. They categorize by active, obsolete, and discontinued.
-
@NeverDie Digikey is good, but too expensive for me. I have to order minimum £33 to have a free delivery.
I still may use them as the search engine only.
Thanks for heads up
-
@alexsh1
yeah, that's what I meant. You needn't buy from them.
-
@alexsh1 Manufacturers are probably not very keen on making them, since the production costs are higher.
-
I've been following this thread and have bought a variety of little panels to experiment with.
These 5.5v 4F 'supercaps' have arrived in the post and I'm not sure how I should test them.
For now I would just like to see how well they hold a charge. What's the best procedure to test them?
-
@xgarb
I wish there were standard tests that were widely used, but there don't seem to be. You might try holding it at its rated voltage for 72 hours, then remove power and check its voltage after 24 hours with no charging to get some idea as to what its leakage is. These tests only mean something if you compare them against a capacitor that you like though. Then maybe it will tell you whether it's better or worse.The only test that really matters is how it performs in your use-case.
-
@xgarb I have them too, I tried and there is definitely leakage but it should not matter much as the sensor will not have to stay for days without charging
-
https://www.tindie.com/products/MadElectronEngineering/infinity-solar-bank-v401/
Just another option. Quite expensive though.
mCSTFZtUluPLYwuCR7RYaeQ
