Group Actions
Engagement
  • Author Author: Jan Cumps
  • Date Created: Date Created
  • Views 3858 views
  • Likes 10 likes
  • Comments 12 comments
Related
Recommended

Solar Energy Harvesting with Epishine Flexible Solar Cell - Pt 2: Circuit Analysis

Jan Cumps

I got a preview of a light harvesting kit from Epishine and element14. Check my intro here.

In this post, I try to analyse the ways of working. I don't have a schematic.It's based on google-research.

 

image

 

 

 

There are 3 main modules:

  • the photovoltaic cell to turn light into electric energy
  • energy harvesting IC with a super cap to collect and store the energy
  • a variable buck regulator for the output voltage

image

 

Solar Cell: Epishine LEH3_50x50_6

 

The largest part of the kit is covered by the flexible photovoltaic cell.

This is Epishine's core business: organic solar cells printed on plastic sheet. The company says it uses biopolymers.

For a manufacturer independent view on printed cell technology, I found the explanation of Sadok Ben Dkhil a good read.

image

You can clearly see the 6 printed elements, and the - also printed - wiring between them.

Epishine suggests on the website that these are for ambient and artificial light conditions.

A soft suggestion. It's not explicitely stated. But all proposed use cases are for those conditions.

Jon has commented that UV light in particular may age them.

 

The flex design can be interesting. First of all, it allows the manufacturer to produce and ship these on rolls.

That's a way easier process than producing big flat surfaces. Handling in the logistics chain and storage are easier for products "on roll".

It's new technology though. And as of now, it's still difficult to master the production print process, and get stability over time.

 

Another possible advantage is that these are thin. Very thin. They take up surface, but almost no depth.The

The flex also allows to put them in a design that has a curved outline.

The material isn't hardened and can't be flexed beyond a reasonable point. Designs will have to take that into account.

 

In the evaluation kit, the cell is fixed on a flexible PCB. The PCB adds protection against over-bending - it is less compliant than the cell.

This combination is a good example of a very thin design that protects the cell against physical force.

The cell surface is exposed. It isn't protected in this kit.  Proper care is needed when handling and storing.

 

Harvesting Circuit: e-Peas AEM10941 and CAP-xx Super Capacitor

 

image

 

The harvest circuit is an e-Peas solar energy harvesting IC. It's an IC that is specifically designed for this kund of application:

  • input from a solar cell
  • additional battery sourcing option - it switches to that to extend the on-time when the super cap is discharging.

 

image

image source: AEM10941 PRODUCT BRIEF: Functional block diagram

 

You can find some of this device's parameters back in the kit's spec:

 

  • Cold start from 380 mV
  • Automatically switches to the primary battery when the secondary battery is exhausted

 

On the kit from Epishine, the PV Cell is the fex solar cell. The Primary battery is the optional one that can be connected in.

The Energy Storage is the super CAP.

The CFG [0..2] can be found back on the board too: 0 and 2 pulled down. 1 pulled up. May want to look up what that config means...

 

The super capacitor is a GA230F CAP-XX 0.40 F, 5 V component. The datasheet is available.

It's made out of two separate capacitors, with a tap in the middle called BAL. The e-Peas IC uses that middle tap.

Like all components in this kit, it's a master of miniaturisation. 400 miliFarads in a component the size of a CR-2032.

image

The circuit loads the capacitor at 4.5 V. Who remembers the formula to calculate the energy contained by  a 0.4 F capacitor at 4.5 V?

 

Output Circuit: Buck Converter with Selectable Voltage

 

image

 

The output is controlled by a Texas Instruments TPS62740.

The functionality of this step down converter, again, matches the details of the Epishine kit.

 

  • Up to 300mA Output Current
  • 16 Selectable Output Voltages in 100mV Steps between 1.8V to 3.3V

They even mention in the application: The input voltage range up to 5.5V allows also operation from a USB port and thin-film solar modules.

 

 

The three blocks above result in a nice combination.

When I first looked at the kit, I couldn't imagine that there would be a wealth of electronics to investigate.

 

 

 

Off to the next blog. First measurements ...

 

Related blog
Pt 1: intro
Pt 2: Circuit Analysis
Pt 3: Charge and Discharge
Pt 4: Battery Backup
Pt 5: test points
Parents

  • I leave it to the coulomb experts to calculate the charge.

    That's easy. Even I can do sums like that!

     

    Q = C x V

     

    so 0.4F x 4.5V = 1.8 coulombs of charge

     

    The charge is directly proportional to voltage so, unlike a battery, you can determine the state of charge simply from the terminal voltage [at low currents, that is - at higher currents the internal resistance will also play a part].

     

    The energy stored is

     

    J = 0.5 x C x V^2

     

    so, once charged to 4.5V, that will be: 0.5 x 0.4F x (4.5V x 4.5V) = 4.05 Joules

     

    1.8C isn't a massive amount of charge. Even if you could contrive to use all of it to power a circuit [not realistic], that would be

    1.8A for 1 second or
    1.8mA for 16 minutes or
    180uA for 2.7 hours

     

    So, for an extended period of darkness, that suggests either very low power operation or a sleeping processor waking occasionally for a bit of activity.

     

    But even powering from the panel, if it's scavenging internal ambient light of a few hundred lux it's not going to be producing all that much power, so the supercap probably matches that reasonably well. Anyway, you'll see when you start to do your measurements.

     

    Where is the output converter connected? Does it take its input directly from the supercap?

  • in reply to jc2048
    Even if you could contrive to use all of it to power a circuit [not realistic], that would be ...

    The usable energy is depending on the output, but always relates to the capacitor's voltage for when it starts up and shuts down.

    The circuit starts to work when Vcap reaches 3.9 V - regardless the requested output voltage.

    image

    At discharge, the circuit keeps on working until Vcap drops to Vout.

     

    There's a also a graph that shows how a 1.2 V battery can extend the on-time by letting it charge the cap a bit when it's almost discharged down to Vout level.

    I'll save that for later as it doesn't help for the analysis of basic circuit behaviour.

Comment
  • in reply to jc2048
    Even if you could contrive to use all of it to power a circuit [not realistic], that would be ...

    The usable energy is depending on the output, but always relates to the capacitor's voltage for when it starts up and shuts down.

    The circuit starts to work when Vcap reaches 3.9 V - regardless the requested output voltage.

    image

    At discharge, the circuit keeps on working until Vcap drops to Vout.

     

    There's a also a graph that shows how a 1.2 V battery can extend the on-time by letting it charge the cap a bit when it's almost discharged down to Vout level.

    I'll save that for later as it doesn't help for the analysis of basic circuit behaviour.

Children
  • in reply to Jan Cumps

    I've done some measurements:

     

    image

    (edit: corrected the drawing. It said 590K, should be 590R)

     

    37000 samples. The first charge was done without load. I added a 590 Ohm load at the point where it then starts the first discharge.

    The green line is the theoretical output. Not what's actually happening on the output side. Because I only logged the Vcap. It's representative though.

    You have to think along here:

    - the output switches off at every low dip of the charge

    - then switches back on at the highest point and lasts till the next low point.

    Total horizontal timeline: 7633 seconds, 37770 datapoints.

     

    Output current before I add the load is nothing. When the load is added, 3 mA.

    On time is approx 513 seconds when the load is permanent and I take output as soon as Vout is available.

    The first cycle, where I gave the supercap time to charge to the highest point, gave 2190 858 seconds of 3 mA output at 1.8 V.

    Smart brains can calculate the energy consumed ...

     

    edit, now fixed: I need to recheck the spreadsheet. the timings seem incorrect, when visually checking on the chart. 513 looks like half of 2190. Not normal for a linear axis image

  • in reply to Jan Cumps

    I did a full charge yesterday evening, 19:30 , Vcap to 4.50 V

    Then, with no load, covered the device so that no light could get in (black box + towel over that)

    Today at 12:00 Vcap is 4.13. A self-discharge of 0.37 V in 16 hours and a half.

  • in reply to Jan Cumps

    Here is the zoom in of the first discharge ramp, where Vcap starts from 4.5 V, Vout = 1.8 V, Iout = 3 mA

     

    image

    datarange in sample spreadsheet: $supercap1.$A$19001:$A$23245

    Start time: 3839.588863681 s

    End time: 4697.276276045 s

    duration: 857.687412364 s

  • in reply to Jan Cumps

    Same exercise for a discharge where we immediately start to take power when Vcap reached it's minimum: 3.9 V

     

    image

    datarange in sample spreadsheet: $supercap1.$A$27185:$A$29750

    Start time: 5493.527778257 s

    End time: 6011.899451787 s

    duration: 518.371673529999