Can Solar Panels Harness Power from the 'Super' Moon for Data Collection?

[nsaspook]

This is last nights chart (from about 10PM to sunrise) into about a 3K load from about 500W total of panels. The hump at about 800 samples is where a separate panel of 120W started getting light from the moon. Maybe 20uW at max power.

The DAQ is a Linux based NI-DAQCard-700 12-bit diff-input card using my kernel driver with a 128X oversample to squeeze a few extra bits of resolution.

 

[jim hardy]

Gives one a perspective for dynamic range of the eye, eh ?

500 watts/20 uwatts = a ratio in incident light of twenty-five million- yet we see pretty well over the whole range.

 

[sophiecentaur]

The Sun's apparent magnitude is around 27 and the (full) Moon's is around -13. That's a ratio of around 400,000:1.
No hope of charging your phone at night then.

 

[jim hardy]

Only 400,000 ?
I'm surprised .

But then, at 3K his solar panel probably wasn't loaded for power transfer.

 

[nsaspook]

Only 400,000 ?
I'm surprised .

But then, at 3K his solar panel probably wasn't loaded for power transfer.

I'm sure it wasn't at the max power point voltage but that would still be only be a few uA of current at that voltage in the moon shine. The normal VMP is about 17.3vdc.

 

[sophiecentaur]

Only 400,000 ?
I'm surprised .

But then, at 3K his solar panel probably wasn't loaded for power transfer.

That is just the ratio of the received radiated power (using the definition of stellar magnitude)
m₁- m₂=-2.5log₁₀(I₁/I₂)

As the panel is not optimised for moonlight then you could expect anything, I guess.

 

[jim hardy]

That is just the ratio of the received radiated power (using the definition of stellar magnitude)
m₁- m₂=-2.5log₁₀(I₁/I₂)

As the panel is not optimised for moonlight then you could expect anything, I guess.

You know what ? I looked that up after your earlier post

and keep getting 10¹⁶ for change from 27 to -13...

Am i mis-handling a sign ?

 

[sophiecentaur]

The difference is 14 (I missed out the negative sign for the Sun's magnitude) and I was quoting from this Wiki entry. and I rearranged one equation to get the ratio on one side. I then quoted their answer but you are right about needing to get the sign right! I, naturally used the 'right value' when I slotted it in and got the 400000 answer. Negative magnitudes take you brighter and brighter as they increase negatively. It's not the intuitive way round but it relates to scoring the most visible stars as zero (Vega is the Zero point) and then the magnitudes increase as they get harder and harder to see. Seeing the bright objects like the Sun, obviously didn't count for the ancients.

 

[OmCheeto]

A more interesting experiment would have you measure the no load voltages generated by the moon and sun. I discovered a while back that no load voltage is directly proportional to the amount of sunlight falling on a panel. The internal resistance of the panels makes it very problematic otherwise. For me anyways.

I would do the experiment myself, but I don't have a meter that measures microvolts.

 

[jim hardy]

Ahhh mystery solved.

I looked here
http://www.astronomynotes.com/starprop/s4.htm
at their chart
and still missed a sign.

It's a lifelong frustration for me - silly arithmetic mistakes.


Thanks !

old jim

 

[sophiecentaur]

You wouldn't have got it wrong if they were taking kVA, would you? That autopilot kicks to scrutinise the first answer you get with familiar quantities. I bet that's the first sum you did about stars for (cosmological) years.

 

[jim hardy]

I bet that's the first sum you did about stars for (cosmological) years.

:rofl:

first ever. Thanks again for the illuminating insight!

 

[nsaspook]

A more interesting experiment would have you measure the no load voltages generated by the moon and sun. I discovered a while back that no load voltage is directly proportional to the amount of sunlight falling on a panel. The internal resistance of the panels makes it very problematic otherwise. For me anyways.

I would do the experiment myself, but I don't have a meter that measures microvolts.

I could get the full Voc for one Sharp 80W panel (about 21 volts) using just a Fluke 177 in full moon light but my high precision system filter network loaded it down to the recorded values.

I changed my processing algorithm to cleanup the signal for last night but it started getting cloudy so you can see the changes in intensity due to the layer during the night in Oregon until my battery charger (from running a fan on the inverter to cool down in the house) auto-kicked in and ruined the rest of the data.

I guess that's it for this cycle.

 

[sophiecentaur]

That post just clouds the issue - and moonlighting is taxable, you know.

 

[OmCheeto]

I could get the full Voc for one Sharp 80W panel (about 21 volts) using just a Fluke 177 in full moon light

I'm confused. Are you saying your panel put out 21 volts under no load illuminated just by moonlight? A Fluke 177 has a minimum resolution of 1 millivolt, so I'm having trouble interpreting that any other way.

but my high precision system filter network loaded it down to the recorded values.

I changed my processing algorithm to cleanup the signal for last night but it started getting cloudy so you can see the changes in intensity due to the layer during the night in Oregon until my battery charger (from running a fan on the inverter to cool down in the house) auto-kicked in and ruined the rest of the data.

I guess that's it for this cycle.

If those are millivolts on your y-axis, then it looks like your data is pretty much where it should be.

≈21vdc/≈400,000 ≈ 50 microvolts (your peak)

 

[RonL]

Might be a good reason to check out the latest batch of MOONSHINE

 

[nsaspook]

I'm confused. Are you saying your panel put out 21 volts under no load illuminated just by moonlight? A Fluke 177 has a minimum resolution of 1 millivolt, so I'm having trouble interpreting that any other way.

Yes. You can sometimes get full voltage from moonlight on a open circuit panel when measured with a high resistance DC meter (>10Mohm). The cooler the better to reduce leakage and the Sharp panels were some of the best made.

 

[nsaspook]

Chart from last nights 'Blood Moon' eclipse in log scale for voltage.
The chart starts at about 9pm until dawn.
Had a few clouds up high just before the eclipse.
http://www.space.com/27380-blood-moon-eclipse-photos-october-2014.html

 

[dlgoff]

Chart from last nights 'Blood Moon' eclipse in log scale for voltage.
The chart starts at about 9pm until dawn.
Had a few clouds up high just before the eclipse.
http://www.space.com/27380-blood-moon-eclipse-photos-october-2014.html

That's sooo cool. :)

 

[nsaspook]

Setting up a new rig for the Super Blood moon this month. I'll use a RPi2 this time for the data logger with another custom Linux kernel driver for the ADS1220 24bit ADC chip via the SPI port.
Just finished up the hardware and software calibration with a old Omega CL511 calibrator.

A few pics of the test setup and build.
The home workshop is a bit of a mess but it's all working now.

Using a 4.998k load resistor for the solar panel and the calibration voltage adjustments @ 2000mV from the standard. The application program 'moonlight' logs the data to a text file with a time-stamp and sequential numbers for each reading so the data can be processed later.

ADC front-end board powered by two 9 volt batteries for a bipolar analog supply.
The brass plates are ground plane/shields for the input filter and are connected to the metal box for (some) electrostatic shielding.

From the program info header:
* A special version for the TI ADS1220 SD ADC converter chip (and MCP3911 later) for low voltage sensing and
* solar panel panel light detection. +- 2.048, 1.024 and 0.512 voltage ranges @ 20 bits of usable resolution
* ADC is in single-shot conversion mode @20SPS, PGA disabled and gain from 1, 2 and 4 in differential
* signal detection mode, 50/60Hz rejection enabled. 500kHz SPI clock with direct RPi2 connection
* Analog +- 2.5VDC from Zener regulators for the bipolar input stage with external 2.5VDC Zener input
* signal protection.
Driver software: https://github.com/nsaspook/nidaq700/tree/master/supermoon

RPi2 with SPI connection to ADC with a breakout board to scope the signals while writing the driver.

SPI data stream (clocks on the bottom trace) to the RPi2 in 24 bit format after a read command. It's a Delta-Sigma converter so the last few bits are just noise but 20bits are usable.

This should be in the micro-watt range during the full moon with a 5k load resistor.

 

[nsaspook]

Running some baseline charts with the Raspberry Pi DAQ system. This is a sunset to night sky (with some cloud cover) with about a 20% moon in the sky. The first little glitch is me moving the panels (3 80W panels isolated from the rest of the panels) to line up with the moons position in the sky, the second large spike is from a cars headlights driving out of the cul-de-sac I live in.

Y Log scale for Volts in BLUE and Watts in RED, X sample number.

 

[nsaspook]

Here is the chart from most of last night as the moon moves into range and sets. It's starting to look like the weather might not be good on the Blood-moon night near Portland as the chart shows the light variations from cloud cover and ice crystals high in the sky.

Those sharp isolated spikes in the first part of the chart were driving me crazy as I thought there was a defect somewhere causing them but I matched the times to airplanes flying near the house (I'm in the glide-path for PDX) after seeing the spikes stopped after about midnight.

 

[jim hardy]

I matched the times to airplanes flying near the house

Landing lights ?

 

[nsaspook]

Landing lights ?

Yes, I live up the river from the airport so normally they do a loop from Hood to line up with the river off one side at night (I would guess to reduce noise). This places the light cone on the panels for 15 to 20 seconds per landing if the path is just right. I've lived here for so long I don't even notice them in the sky but I did notice them missing during 9/11.

 

[nsaspook]

Another plane, another small spike.

From the DSLR

Clouds moving in

1.2 uW (1.34uW at the last reading) power from a 240W array, I could charge a capacitor and use it to run my cell phone.

 

[nsaspook]

A different type of chart from last nights Voltage data:
Lots of clouds.
The 'X' time data moves counter-clockwise.

 

[nsaspook]

Measured voltage under full moon light (not from supermoon data) with open circuit, 4.7k, 1.0k and 10.0k resistors vs voltage.

A simple model of the panel characteristics during moonlight at varying power levels using the resistor/voltage data.

You can see a somewhat typical Solar PN curve even at low-light levels.

This is the log power chart from the Super Blood moon Sept 27 2015.
All eclipse voltage, power, current data and graphs are with the 4.7K load resistor.
It starts at almost full 'Blood' stage (the tall sharp spikes are car lights) and moves to left -> right to full moonlight.
The first ramp is the main panel array in moonlight (the small ramps are me moving the array to center sky and the longer spike is an open circuit voltage reading taken by removing the 4.7K load for a while for model data)
The second ramp is when the other panel array starts getting full moonlight and both arrays are in view of the full moon until early morning when the program is stopped before dawn.

Big image:

 

[anorlunda]

Really fun experiments. Thanks for sharing.

I live on a boat, so I think of morphing everything to marine applications. Using your methods, I bet I could see the rocking of the boat on the waves in the data. At the very least, a Fourier analysis of the data should give the primary period of the swells as well as the shorter periods of the chop.

There are lots of solar-powered data gathering buoys around the world. Perphaps they could be simplified or improved by harvesting data from their own solar panels. Cloud cover for example, might be estimated this way simpler than other instruments can do it. The really big data analysis challenge would be to make it work both day and night.