0%
0%

# 1 Square Inch of Power

An MPPT solar boost converter with 3.3v Output and Supercap energy storage.

Similar projects worth following
4.4k views
A tiny power supply for connecting to tiny solar panels. To power all sorts of tiny devices.

After using these fantastic little ICs in my rock (see my other Hackaday.io project!) I thought that they would be perfect for the square inch challenge.

Taking further inspiration from "The Signal Path" vblog (http://thesignalpath.com/blogs/2013/01/03/tutorial-and-experiments-on-energy-harvesting-ics/) I thought I'd make a tiny 5v (or 3.3v) solar powered power supply, complete with a super capacitor for energy storage. Perfect for wireless sensor networks and other low power IoT devices.

The main board consists of 2 Linear Technology LTC3105 ICs. The first one is connected to the solar cell board (which will also be a square inch board, but I didn't get it sorted in time as an entry). It performs Maximum Power Point Tracking of the solar cells, to extract as much energy as possible. The other great feature of these ICs is the super low startup voltage. They'll startup at under 300mv, boosting the output to the set voltage - in this case, 2.7v - the voltage of the Supercap.

The reason for the second IC is explained in the video, but here's a summary. If I was using a lipo battery, full charge is about 4.2v, and discharged is about 3. There's not much energy left in a lipo below that level, and discharging one lower than that can cause permanent damage. The capacity of a Supercap is based on a complete discharge from its rated voltage down to zero. Now, to get 5v from a 2.7v Supercap, I need a boost converter ( which is what the LTC3105 is). And because of the very low startup voltage, I can use almost all of the energy stored. ( I am soooooo tempted to make an 800% more power available claim here.......but we all know what happens if you do that) . Plus, I wanted to make something that doesn't use lipo batteries, and could be used in a variety of temperatures, including below zero (not that I see any of that here in Australia)

The solar cell board will be able to be stacked directly on top of the power supply. It contains 3 small Ixys solar cells @ 0.5v ea. Connected in series, they should produce 1.5v @ 44mA at their maximum power point. You could attach a larger solar panel, as long as you don't exceed the IC limits and you set the MPPT voltage accordingly

• 2 × LTC3105 Power Management ICs / Switching Regulators and Controllers
• 2 × SRU3017-6R8Y Inductors, Chokes, Coils and Magnetics / Fixed Inductors, Chokes and Coils
• 1 × 470uF Tantalum Cap 6032
• 1 × 5F Supercap with 5mm lead spacing
• 2 × 10uF Ceramic Cap 1206
• ### Eureka!

OzQube01/20/2016 at 12:54 0 comments

Today I had a nice sunny day, so I plugged the solar panel in and left it in the sun, then waited 15 minutes......

Still no good!!!! Something has to be wrong. So I removed the second LTC3105 so I could check that the first one is working. Back into to sun for 15 mins....

Success! The supercap voltage was just over 2.5v! So there must be an issue with the second LTC3105. I tried resoldering using the solder left on the pad.....Ba Bong...(don't do that) so I removed all the old solder using solder wick and then got some fresh solder paste and got to work with the hot air gun. I put a bit too much on, so a couple of traces were bridges next to the DFN package. A quick cleanup with the wick and try again....

Nope! No output from second LTC3105.

Back to the schematic to see what I've done wrong. I noticed a couple of things. Because I was using a supercap between the 2 ICs, I thought I wouldn't need a big 10uf cap on the input to the LTC3105, like the first one in the chain does. I used a 1uF. Maybe if I change that.

Nope. That didn't work.

The only other "extra" part on the board is a 33pF filtering cap between the output and the feedback pin. I may as well try removing that. Can't hurt right???

Right!!! It now works!!

After leaving it in the sun for 15 mins, the supercap is full and 3.3V on the output!

Let's add a load to the circuit. I have an Arduino Pro Mini 3.3v lying around with the default Blink sketch on it. I plugged it in to the Vin pin (not Raw) and the power light came on and the second led started blinking!

In full sun, the tiny solar panel could only just continue to charge with both LEDs on, but charge it did!

Even when it was moved into the shade, the supercap and the second LTC3105 still produced 3.3v output, with the supercap voltage slowly going down.

I let it fully discharge then put it back into the sun. The supercap voltage got back up to 0.4v.....then dropped back to .3..... The output was only .2v.

So looks like if there's a load connected that draws too much power on startup, the board gets stuck in a loop.

Methinks it's time for a voltage supervisor or similar that either enables the second LTC3105 only when the supercap voltage is over a certain threshold (1v maybe?). I also only had a 4.7uF cap on the output of the second LTC3105. Looking back at the datasheet, 10uF is recommended, so I'll replace that and try again.

Regardless, the board works! It could be used to power a small micro and sensor. Maybe even BTLE or a low power RF module?

To be continued ......

• ### Suparcaps and Math

OzQube01/18/2016 at 23:11 0 comments

Just as predicted last project log, the supercaps arrived!

I soldered one of the 2.7v/5F caps onto the board, then setup the solar panel under the halogen light source ( because night time). I connected the multimeter up to the positive terminal of the supercap, which is also the output of the first LTC3105. The voltage rose to .3v..........but no higher.....

Now I know its a tiny solar panel, so I left it under the light for 5 minutes. I had previously figured out that having the solar cell really close to the light is a good approximation to sunlight.

Or is it?

No increase in voltage.

As I don't have a lab power supply, I can't provide an alternative input that is current limited. I thought I'd connect an AA up directly to the supercap to charge it up to 1.5V. And yes, this worked!

I then put the lamp back onto the solar cell, but no increase in voltage even after 5 minutes.... What's going on?

I checked the solar panel voltage again. It's outputting 1.48V. Exactly what it should be with the MPPT resistor setting. Maybe I have a short? I desoldered the second LTC3105 so I can isolate the first one. Tried again. No increase.

I went back to one of the reference web pages that are similar to this project - http://wiesel.ece.utah.edu/redmine/projects/essc/wiki

I decided to do the math to see how long it should take the solar panel to charge the supercap.

From the website:

• Charge time:
• Ic = C dv/dt
• Ic/C dt = dv
• V = int(Ic/C dt)
• V = Ic/C * t
• V/Ic/C = t

Theoretically, solar output is 1.5V @ 44.6mA. So 0.0669W. Using the LTC3105 datasheet, it shows a graph for efficiency and power loss for certain input voltages and output currents. Now I'm just using the basic Ohms law calculations for the output current.

So assuming 100% efficient, output power is .0669W. At 2.5V, the output current is 26.76mA. Efficiency at this current vs the 1.5v input is a smidge over 70%. So output power is .04683W. At 2.5V, that's 18.732mA. (

V = 2.5V C = 6F Ic=.018732

I had to check the math, so (V*C)/Ic works to get their results ( and using the math from another source)

2.5V/.018A/6F = 800 s = 13 minutes

Looks like I'll have to test during the day to make sure the solar panels are actually outputting the required current, as I also found that the halogen lamp easily makes the cell output the right voltage, but output current is questionable.

To be continued....

• ### The joys of shipping

OzQube01/16/2016 at 06:22 0 comments

Just a quick update......but nothing useful really.

The Supercaps that were ordered from AliExpress on 22nd Dec still haven't arrived, although other items ordered on the same day have arrived. So it shouldn't be too long before I can complete the module and test it properly.

Originally, I was going to solder the solar panel directly onto the power supply board using regular pin headers. The problem is that it's hard to get access to the board to switch out the MPPT resistors ( If a different panel is used).

I might get some tall headers like these:

Lets see what the postman brings next week!

• ### I Have The Power!!!!!

OzQube12/22/2015 at 15:25 0 comments

Right on the last day to enter the comp, the PCBs arrived from Oshpark! So a quick assembly and test (at night, but using a halogen light source). 3.249v output (not 3.3v) That's because I don't have the exact resistors I used in the schematic, so I got as close as I could get with the ones I had available ! The Supercaps haven't arrived yet, so I can't perform a full test. The little solar panel performs as expected. 1.5v @ approx 45mA.

• ### What is this Thanksgiving????

OzQube12/02/2015 at 13:01 0 comments

Silly me forgot that everything in USA is closed for the Thanksgiving weekend. I sent off the board designs to get made by OSHpark last week.....

Anyway, while I'm waiting, I uploaded the gerbers, BOM, schematic and added links to the OSHpark board ( That should qualify me for the next round :-) ) I created a second square inch board (although not part of the comp) to attach the solar cells. The cells I'm using are IXYS KXOB22-12x1L cells. Each one is 0.5V, and there's 3 in series. So a whole .07W (ish). The link to the project on OSHpark is there too.

The schematic has the resistors to set the MPPT voltage to 1.5V, and the output of the second DC-DC converter is set to 3.3V. These can be changed using the info from the datasheet, in case the solar cell you use is different, or if you want 5V out.

• ### Just in a nick of time

OzQube11/24/2015 at 15:21 0 comments

At the last minute, I thought I'd enter the square inch challenge. So I did a quick layout, and found that a full square inch is waaaaaaay to big. So any suggestions on what else to fit in there

Share

## Discussions

Mark Jeronimus wrote 08/08/2018 at 12:12 point

The LTC3105 seems to be deprecated ("Not recommended for new designs" according to Farnell)

Are you sure? yes | no

OzQube wrote 08/10/2018 at 03:20 point

Hi Mark. It may be that they're changing to an "Analog Devices" source, now that "Linear Technology" is part of Analog? I'm not sure. It's still listed at "Production" on the Analog website.

Are you sure? yes | no

Jasper Sikken wrote 07/15/2018 at 18:08 point

Nice project. Why don't you take part in the harvesting challenge as part of the hackaday prize?  You send in your entry in coming hours. Your project is is similar to mine. I have integrated the Ixys solar cells with the harvesting IC onto one module. https://hackaday.io/project/159139-tiny-solar-energy-module-tsem

Are you sure? yes | no

Andrew Reed wrote 05/12/2017 at 17:35 point

OzQube,

Hi, I really like your project and the direction you took it.  Any new progress?

Are you sure? yes | no

alpha_ninja wrote 12/21/2015 at 07:50 point

Please make sure to upload photographic or video proof of this project working by the next deadline (in about two days!)

Are you sure? yes | no

alpha_ninja wrote 12/07/2015 at 01:36 point

Just a friendly reminder to please upload your design source files by 23:59 UTC on Dec 8, 2015 to be in the running for #The Square Inch Project!

Are you sure? yes | no

OzQube wrote 12/07/2015 at 02:57 point

Is Dropbox OK as a source location or does it have to be Github?? I've copied the design files from Dropbox to Github just in case.

Are you sure? yes | no

alpha_ninja wrote 12/07/2015 at 03:00 point

Dropbox is fine, too.

Are you sure? yes | no

alpha_ninja wrote 12/07/2015 at 01:35 point

Just a friendly reminder to please upload your design source files by 23:59 UTC on Dec 8, 2015 to be in the running for #The Square Inch Project!

Are you sure? yes | no

alpha_ninja wrote 12/02/2015 at 00:40 point

Are you sure? yes | no

kens wrote 11/25/2015 at 17:44 point

Just a couple ideas here.

I have to use huge capacitors (100's uF) across solar cells because a boost-converter requires very high peak currents, compared to the average solar cell generated current. It's not always full sun. LTC3105 seems to be 300mA peak on 15mA output, using Cin 470uF cap. Your cap looks too small.

When there's no sun the supercap/batteries should not backfeed into the boost converter. I see a need to block reverse power flow using a MOSFET as an ideal diode.

Are you sure? yes | no

OzQube wrote 11/26/2015 at 05:02 point

Hi kens. Thanks for the tips.

The datasheet recommends a 10uF cap on the input? The example Linear uses for the industrial current loop, which is only 4-20mA, at 330mV, has a 470uF input cap. The cells I'm planning to use with it are 1.5V @ 44mA, so still tiny, but well above their example design. I'm sure it'll take ages to charge the supercap.

I have used a 1206 size footprint so I can get up to a 220uF cap if necessary.

The IC uses a few tricks to limit the peak inductor current during its startup phase, so this could be why they recommend a smaller input cap than you have suggested?

I think that if there's no input power, the device turns off its internal mosfets, limiting leakage current. Although I haven't done the math on that ( and the output voltage settings resistors.)

There is another example project using this IC ( http://wiesel.ece.utah.edu/redmine/projects/essc/wiki ) that shows only very slow leakage. Although it may still be significant enough for sensors that  need to last for days without charging. I'll keep it in mind for rev2 :-)

I'll post the schematic on the weekend, which will have both ICs on there.

Are you sure? yes | no

## Similar Projects

Project Owner Contributor

### 30W (=~240W halogen) LED balcony solar lanterns

rawe

Project Owner Contributor

### Lampion

atelierlelu

Project Owner Contributor

### Solar Monitoring System (SMS)

andrew

Project Owner Contributor

Tobias

# Does this project spark your interest?

Become a member to follow this project and never miss any updates