solar panel test rig - odd results ?

[sarabande]

I've just set up a solar panel test rig, prior to trundling it down to the boat.

43W PV Kyocera panel
Victron 75/15 controller
decent cabling
AGM test battery

There is power going into the battery: after an hour's running the V has gone from 12.4 to 12.6 (rounded down). But when I put the meter (not calibrated but fairly good one and previously held trustworthy) on to the input terminals into the controller from the panel, the display sweeps from about 18.7x in stages down to about 13.6x, then up to 18.6, over about a two second cycle.

The controller LEDs indicate that the battery is absorbing a bulk charge.

The sampling rate of the meter is unknown, as is the V at the output of the panel.

Is this how an MPPT controller works ? Or should I get an oscilloscope for the (android) phone and investigate further ?

 

[JumbleDuck]

Is this how an MPPT controller works ?

As I understand it, they are effectively switch-mode powers supplies, so I'd expect to see them sawtooth: lots of current into the capacitors at ~13V, less and less current as the capacitors fill up to ~18V, discharge, rinse, lather repeat. I'd have a scope on it from sheer curiosity.

 

[lw395]

The MPPT controller is varying the current drawn from the panel to find the Maximum Power Point. Having found the best load condition for the panel it should Track it, by making a smaller change in current drawn and calculating whether the delivered power has gone up or down. Once it has been connected for a while, it should spend a very high % of its time close to the optimum, unless the operating conditions keep changing rapidly, like some forumite walking in front the panel.....
Can you measure the current at all?
It may be that the controller is struggling because there isn't enough sun, so it's trapped in a start-up loop. 12.4 off load to 12.6 charging is not much charge for an hour.....

 

[Engage]

If I understand you correctly, it may be that the controller is periodically taking current from the solar panel to "top-up" the battery on a 2 second cycle. When the solar panel is open circuit i.e. no current flowing, it may well supply 18.7 volts or more, however, as soon as current is drawn from the solar panel that voltage will fall. If you can confirm that current is flowing to the battery in sync with the voltage falling then that may be the case. I have used a few charge controllers and none of them have bounced like that though.

 

[lw395]

As I understand it, they are effectively switch-mode powers supplies, so I'd expect to see them sawtooth: lots of current into the capacitors at ~13V, less and less current as the capacitors fill up to ~18V, discharge, rinse, lather repeat. I'd have a scope on it from sheer curiosity.

Yes, they are switch mode converters, but instead of regulating to keep the output voltage constant, they regulate to keep the output power to a maximum.
So they will be switching at anything from 50kHz to a few MHz, while varying the panel voltage and current on a much slower time scale.
A scope will probably show the slow tracking variation plus a bit of ripple at the switching frequency.
To see the interesting sawtooth stuff you'd have to get the lid off, the input and output should be reasonably filtered.

 

[Yngmar]

I've just set up a solar panel test rig, prior to trundling it down to the boat.

43W PV Kyocera panel
Victron 75/15 controller
decent cabling
AGM test battery

There is power going into the battery: after an hour's running the V has gone from 12.4 to 12.6 (rounded down). But when I put the meter (not calibrated but fairly good one and previously held trustworthy) on to the input terminals into the controller from the panel, the display sweeps from about 18.7x in stages down to about 13.6x, then up to 18.6, over about a two second cycle.

The controller LEDs indicate that the battery is absorbing a bulk charge.

The sampling rate of the meter is unknown, as is the V at the output of the panel.

Is this how an MPPT controller works ? Or should I get an oscilloscope for the (android) phone and investigate further ?

Did you have a peek in the manual? It actually tells you how the controller works. This bit on page 4 is especially interesting and something most people seem to neglect when they design single-panel installations:

3.2. PV configuration
● The controller will operate only if the PV voltage exceeds battery voltage (Vbat).
● PV voltage must exceed Vbat + 5V for the controller to start. Thereafter minimum PV
voltage is Vbat + 1V.
● Maximum open circuit PV voltage: 75V.

The controller can be used with any PV configuration that satisfies the three above
mentioned conditions.
For example:
12V battery and mono- or polycristalline panels
● Minimum number of cells in series: 36 (12V panel).
● Recommended number of cells for highest controller efficiency: 72
(2x 12V panel in series or 1x 24V panel).
● Maximum: 108 cells (3x 12V panel in series)

So with a single 12V panel the controller will spend a lot of time faffing about because it either doesn't get enough sun to start at all or keeps getting interrupted every time a cloud flies by. Therefore the recommendation of either 2x12V panels or one 24V panel. That's why the thing accepts up to 75V open circuit voltage!

You might have assumed you needed a 12V panel for your 12V bank, but that is actually counter productive. If you still can send it back, swap for a 24V panel or add another 12V panel in series. Your controller will suddenly work a lot better in a lot darker skies.

 

[sarabande]

thanks everyone.


It is not working now, but moonlight is hardly conducive to battery health.


I did read the manual Yngmar, and found that bit about a >5v input needed. Prior to a burst of sunlight between the cu-nims the panel was not generating enough V. but the voltage 'cycling' started when the start alogorithms were triggered.


Guess I have to wait till tomorrow to see if the system assesses the battery, then decides what the controller is going to do with the countless free coulombs emanating from the Sun.




I have always wanted one of those PC oscilloscopes anyway Now there's a reason, not an excuse.

 

[simonfraser]

so is it generally best to connect any suitable panels in paralel up to the max V of the regulator ?

 

[sarabande]

AFAIK, my controller is designated 75/15. The book says up to 75v DC, and 15 Amps.

So if the max v from one panel is 20, then I could go to three panels in series, and the controller would cope. The book says any input over the 75v will be ignored. Somehow Magic stuff...

 

[JumbleDuck]

I have always wanted one of those PC oscilloscopes anyway Now there's a reason, not an excuse.

I have used xoscope, which is a nice oscilloscope program for Linux. The downside is that you are restricted to sound card input. Bitscope do lovely hardware, but I never managed to get their software to work.

 

[Yngmar]

so is it generally best to connect any suitable panels in paralel up to the max V of the regulator ?

In series, not parallel. And open circuit voltage, not nominal voltage. For example for a panel with 12V nominal voltage, the open circuit voltage is around 18V. Victron offers a nice Excel calculator for all of this and more (e.g. voltage drop on the wire): https://www.victronenergy.com/support-and-downloads/software - it'll highlight any over/underspec of panels, wires or regulator.

 

[robertj]

Being an electrical numpty, are you saying, if setting up a panel for a 12v system it would be better going to a 24v panel or to 24v panel for that regulator?
Sorry if it sounds a stupid question.

 

[lw395]

In series, not parallel. And open circuit voltage, not nominal voltage. For example for a panel with 12V nominal voltage, the open circuit voltage is around 18V. Victron offers a nice Excel calculator for all of this and more (e.g. voltage drop on the wire): https://www.victronenergy.com/support-and-downloads/software - it'll highlight any over/underspec of panels, wires or regulator.

Series is usually more efficient. Parallel may be less prone to issues from shading. You pays yer money.....
Personally I think it is more valuable to get a bit of charge later in the day when perhaps only one panel is lit, than to get more power than you can use at midday. Which leans towards parallel, unless you know that your panels work well with one shaded. You can add a diode across each panel to reduce the shading issue, but that may still reduce the volts enough to stop the MPPT starting allegedly.

BTW, for those interested in oscilloscopes, there are LCD ones around for £15 and upwards on ebay. This avoid the lack of DC on most soundcards. I have the 4 channel DSO203 which is more like £100.

 

[Yngmar]

Being an electrical numpty, are you saying, if setting up a panel for a 12v system it would be better going to a 24v panel or to 24v panel for that regulator?
Sorry if it sounds a stupid question.

Yes, exactly as the manual says, either a single 24V panel or two 12V panels in series.

 

[cpedw]

For a cheap scope, considerthe PCB-scope for only £12. It plugs in to USB for power and display on a Windows PC.

Derek

 

[sarabande]

thanks cpedw, looks just the job.