3
\$\begingroup\$

I am an electronics know-not-much :) I do understand the basics of current/voltage/amps etc.

I have an array of 10 x 15watt x 12v (max 17.5v) amorphous solar panels here in the UK. I want to arrange them in 5 banks of two (5 pairs) and when light levels are low (they are a lot in the UK especially during winter) I want to switch each pair from parallel operation into series operation. At the low light levels that exist during the long morning & evening periods and during heavy cloud the panels tend to produce 8-9 volts, which is not much use for charging, But if they switched to series operation, each pair then would produce a useful 16-18 volts. When the sun shines fully I want them to switch back to parallel so as not to over volt the charge controller.

My question is such a control circuit easy or difficult? I have searched & googled but cannot find such a control circuit..

\$\endgroup\$
  • \$\begingroup\$ If you have 8 or 9V and upconvert you get less current and more volts. Energy is conserved. | If you have 16-18V and down convert you get less volts and more current. Energy is conserved. | All things being equal a down converter(buck converter) is liable to be more efficient than an up converter but that's for implementation reasons. \$\endgroup\$ – Russell McMahon Oct 11 '11 at 14:33
1
\$\begingroup\$

Such a control circuit is possible, but a better idea is to leave the panels fixed in one configuration and deal with the resulting change in voltage with the right kind of switching power supply. If you want to get fancy, you could even implement maximum power point tracking, or something reasonably close.

The switching power supply can't make more power than what the panels produce, but it can convert (with a little loss) from whatever voltage and current the panels want to produce to a different controlled voltage or current (with the other limited by the available power).

If the series configuration always produces more voltage than you need even in low light, then that's how you should configure the panels. The switching power supply then always makes a lower voltage, which keeps it simple. That is called a "buck" regulator, with much information about that out there.

It would be helpful to say what voltage you ultimately want, and what current range is useful at that voltage.

Added:

You now say this is to power a 12V system, which presumably runs from lead-acid batteries although you didn't say that.

One useful feature of lead acid batteries is that they can take reasonable charge current even when full and still regulate the voltage well enough for most purposes. Given that, a really simple solution is to wire the solar panels in series to get the higher voltage you talked about, which is always a bit more than the 12V battery level even on cloudy days (when light is really low the voltage will be lower, but then there is so little power to be irrelevant), and just connect this to the 12V rail with a Schottky diode. That will not use the panels most efficiently in high illumination, but probably not so bad on cloudy days from the numbers you give.

A buck converter that runs the panels at the best efficiency for the given insolation and then dumps whatever current it can onto the 12V rail should be a bit more efficient. With a decent converter design, the extra loss in the switcher should be more than offset by running the panels at their optimum efficiency.

However, if you have more than enough power in full sunlight and the real problem is when it's cloudy, maybe the dumb series connection (with a Schottky diode to prevent reverse current when dark) will do it. I'd probably be tempted to try that first and see what you get and how efficient the whole system is on cloudy days when it really matters.

\$\endgroup\$
  • \$\begingroup\$ Hi - As in my comment to Saneesh it's 12v system. Any thing that aids charging is of course desirable. It's a total of 150watts but I am drawing about 30-35 watts continuously (24/7) for network devices. (2 x Router 1, 1 x 8 port switch) I also use it for speakers a desktop calculator & NiMh Battery charger, but not continuously. the power system has so far not got me through the winter :( So you are saying: configure for 24v operation & charge controller, Change batteries to series too. and add a DC To DC Converter to 12v. Sounds obvious now :) \$\endgroup\$ – Timbo Oct 11 '11 at 14:23
1
\$\begingroup\$

Summary:

  • It can be done easily and cheaply if desired.

The "best" way, no expense spared, is to use electronics to convert what you have to what you want. Either a boost converter tp step up low voltage panels when desired or a buck converter to step down series connected panels when required.

HOWEVER

What you describe can be achieved cheaply and simply using relays. You can also do it with electronics, but using a pair of relays is so simple and cheap that they are hard to beat.

Consider this solution with 2 panels. This can be readily extended to 2N panels as required.

See circuit diagram at end.
Note that relay can be SPDT rather than 2 x relays as described here.

Panel B = Upper panel = PVB. .
Panel A = lower panel = PVA

PVB+ connects to B+
PVA- connects to ground.

Provide two relays.
Relay 1 = single pole changeover.
Contacts are NO, NC an common.
NO = normally open
NC = Normally closed = connected to common when depowerd.

Relay 2 = normally open single pole.
Use same NO, common terminology as relay 1.

PVB- to Relay1-common
PVA+ to Relay1-NC
Relay-1_NO to ground

PVA+ also to Relay 2 common
Relay2-NO to Battery +

Provide magic circuit to operate both relays when Vpanel rises to high enough to allow parallel operation and to deoperate them when Vpanel falls too low. You could use relays themselves with hysteresis due to relay characteristics but a simple window comparator or comparator with hysteresis will do this easily.

When PV voltages are low relays are unoperated and panels are in series.
Whe PV panel voltage is high enough relays operate and panels switch to parallel.

Switching could occur due to clouds etc but with enough hysteresis this should not be too bad.

You could also use an LDR or photocell sensor to control this.

You could add a delay of seconds to minutes to minimise chatter.

ENSURE relays are break before make.

If R2 operates while R1 is not operated contacts on both relays will be entertained by current from PVB under short circuit. Adding a diode between R2-NO and battery stops this happening. HIGHLY recommended :-).

Quck sketch suggests this would work OK with 3 x MOSFETS and some head scratching. They probably need to be operated "upside down" in second quadrant due to body diodes but doable. Later.

FET1 ground PVB- when desired.
FET2 connects PVA+ to PVb- WHEN DESIRED.
FET3 connects PVA+ to battery when required.

FET 2 is mid supply referenced but should be no problem.


E&OE !!!!
Check my NC NO common etc :-)


Circuit:

Here we go, then.
Forrest mims style cct:

Relay can probably be a single DPDT or a SPDT and SPST or whatever.
As shown relay(s) is/are unactivated and PV panels are in series
When relay(s) activate(s) PV panels are in parallel.
Diode in RH NO contact lead to battery stops PVB (upper) shorting through two contacts if RH contact closes before LH contact is open.

If contacts are on a DPDT relay and contacts are break before make then diode is not needed.

enter image description here

"Controller"

R1/R2 set changeover voltage.
R3 adds hysteresis as desired.

enter image description here

0515 - bedtime.

\$\endgroup\$
  • \$\begingroup\$ Wow - 0515! Time to wake up! \$\endgroup\$ – Timbo Oct 11 '11 at 16:29
  • \$\begingroup\$ @Timbo - not if one hasn't yet gone to sleep :-). Do you like the circuit and relay idea? You can do it with MOSFETS if desired. Ask if interested. \$\endgroup\$ – Russell McMahon Oct 11 '11 at 16:31
  • \$\begingroup\$ I was answering but am at work and got timed out! Thanks for all the effort! I'm not sure my skills set is up to this - I'll would need to get my brother to build it! But it looks like (to me) what is required. Incidentally I asked Bro' and he said it was complicated too! \$\endgroup\$ – Timbo Oct 11 '11 at 16:41
  • \$\begingroup\$ @Timbo - the relay is very very easy to wire up - just follow the diagram above with contacts probably already labelled NO, NC, common on manufacturer's product diagrams. | controller can be simple - relay in series with resistor from PV panel. When there is enough voltage the relay switches. Adjust resistor to suit. Not as good as my controller. \$\endgroup\$ – Russell McMahon Oct 12 '11 at 2:39
  • \$\begingroup\$ Over two years later I know, being an old fart I completely forgot about this question, I'm here because it suddenly popped up on my stackoverflow account! But I now understand your answer thanks to Arduino, RaspPi and a lot of reading (and a few minor smoke incidents). I don't like to build/install something I don't understand. But now I do :) So I will have a go at implementing as I now know the MOST efficient conversion is NONE :) So I aim to get as close to the target voltage as possible using a micro-controller. I'm fine with programming... Thanks. \$\endgroup\$ – Timbo Jul 14 '14 at 14:32
0
\$\begingroup\$

It is better to use a solar battery charger circuit. They are available for low cost in market, and i don't think you can make a cheaper one. Basically the charger is a DC to DC converter with higher efficiency.

\$\endgroup\$
  • \$\begingroup\$ Thanks - Yes - I have a "12v solar charge controller" in the circuit which is being fed the unregulated output of all the panels in parallel at the moment. This then feeds all my network peripherals at 12V regulated via two 12v leisure Batteries. It just seemed (I know can't create moe power than is available) that the 16-18v would be more efficient than the 8 or 9v.?? Maybe thats just "wrong thinking". \$\endgroup\$ – Timbo Oct 11 '11 at 14:07
  • \$\begingroup\$ Actually i couldn't understand your last sentence in the comment.If you like you can accept this answer by clicking on the empty check mark at the left, or can do an UP vote \$\endgroup\$ – Saneesh A T Oct 11 '11 at 14:12
  • \$\begingroup\$ @Saneesh A T - see my answer. \$\endgroup\$ – Russell McMahon Oct 11 '11 at 14:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.