Panel voltage.

[Beanyboy57]

I have 6 x 120w x 12v panels in a parallel- series configuration and 3 x 120w x 12v panels in a series configuration.

If I want to place the 6 panels in a series configuration I could possibly get 120 + volts. Is it more beneficial to have a higher voltage as opposed to higher amperage in a PV array?

The panels are rated at 17v max so I estimated that I should get anywhere from 36v to 51v however I quite regularly get up to 63v especially if the weather is cool. Luckily I installed a CC that can handle 150v/70amp.


My main question is why are my 12 v panels producing more than 20v?
What is the absolute maximum voltage they can produce given ideal conditions?

[sensij]

Your panels have a "Voc" rating... that is the maximum voltage that the panel can obtain at 25 deg C. Also, the panel will have a voltage temperature coefficient, typically something like -0.3%/deg C. That means that if the actual temperature is only 15 deg C, the actual Voc will be about 3% higher than the rated Voc.

[bcroe]

I have 6 x 120w x 12v panels in a parallel- series configuration and 3 x 120w x 12v panels in a series configuration.

If I want to place the 6 panels in a series configuration I could possibly get 120 + volts. Is it more beneficial to have a higher voltage as opposed to higher amperage in a PV array?

The panels are rated at 17v max so I estimated that I should get anywhere from 36v to 51v however I quite regularly get up to 63v especially if the weather is cool. Luckily I installed a CC that can handle 150v/70amp.

My main question is why are my 12 v panels producing more than 20v?
What is the absolute maximum voltage they can produce given ideal conditions?

Your panels are producing maximum voltage when no power is drawn from them, not
of much use. Loading them down to get the most power will cause a voltage drop of
around 20%. Remember the "12V battery" in fact requires charging voltages more
like 15V.

Running the panels in a higher voltage configuration will reduce the amount of copper
needed to efficiently conduct the power. Bruce Roe

[PNjunction]

Is it more beneficial to have a higher voltage as opposed to higher amperage in a PV array?

Generally yes. Going high voltage means that you typically end up using an mppt controller which does up/down conversion of the high voltage efficiently. (make sure your voltage fits with the specs of the controller!) And as bcroe mentioned, higher voltages means being able to run much longer runs of cabling, or perhaps be able to choose thinner gauge. There are charts for this.

The panels are rated at 17v max so I estimated that I should get anywhere from 36v to 51v however I quite regularly get up to 63v especially if the weather is cool. Luckily I installed a CC that can handle 150v/70amp.

Ah, there you go. Temporary high voltage spikes much higher than usual can be caused by a lensing effect, known as "cloud-edge" knife-edge, or similar. When the sun just starts to peek out from behind clouds, when the rays cross the edge, they can be much more intense than in the open.

My main question is why are my 12 v panels producing more than 20v?

That is typical under good conditions - but depending on the make, it can vary anywhere from 16 to 22v, so a "nominal 12v" panel when open-circuit / non-loaded showing 18 volts is normal. Basically designed to charge 12v batteries, you need the voltage to be several volts higher (usually at least 2) than the absorb voltage.

For instance, to charge high-end "nominal 12v batteries" like agm's that require a 14.8v absorb, the panels would have to be a minimum of about 16.8 open circuit. Charge controller limited of course!

[sensij]

Ah, there you go. Temporary high voltage spikes much higher than usual can be caused by a lensing effect, known as "cloud-edge" knife-edge, or similar. When the sun just starts to peek out from behind clouds, when the rays cross the edge, they can be much more intense than in the open.

Cloud effect irradiance bumps should primarily increase the current, not the voltage. 63 V is close to the Voc * 3... is it possible that values that high are only showing up in Float, when the CC may cut off all current flow for short periods of time?

[inetdog]

Ah, there you go. Temporary high voltage spikes much higher than usual can be caused by a lensing effect, known as "cloud-edge" knife-edge, or similar. When the sun just starts to peek out from behind clouds, when the rays cross the edge, they can be much more intense than in the open.

I think you really mean high current spikes? The Voc and Vmp of a silicon PV panel are remarkably insensitive to the incident radiation level. (In fact, the more irradiation, the higher the panel temperature and so the lower the output voltage.

[PNjunction]

Doh! yes, you guys are quite right - thanks for the correction, the amperage spikes. In my case, the edge of cloud effect is usually so fast that the panels didn't really heat up too quickly to drop voltage, but I could see that happening with no real cloud movement.

That's why at the small end of things, say with simple pwm controllers, not to cheap out and try to use a 7A controller with a 7A panel - despite the controller's ability to handle temporary spikes. In some cases it can still be too small to handle edge of cloud events. In solar, headroom is the name of the game.

[Beanyboy57]

Doh! yes, you guys are quite right - thanks for the correction, the amperage spikes. In my case, the edge of cloud effect is usually so fast that the panels didn't really heat up too quickly to drop voltage, but I could see that happening with no real cloud movement.

That's why at the small end of things, say with simple pwm controllers, not to cheap out and try to use a 7A controller with a 7A panel - despite the controller's ability to handle temporary spikes. In some cases it can still be too small to handle edge of cloud events. In solar, headroom is the name of the game.

That's always been my motto...go for larger than you think you need. It usually pays off in the long run.

[J.P.M.]

Doh! yes, you guys are quite right - thanks for the correction, the amperage spikes. In my case, the edge of cloud effect is usually so fast that the panels didn't really heat up too quickly to drop voltage, but I could see that happening with no real cloud movement.

That's why at the small end of things, say with simple pwm controllers, not to cheap out and try to use a 7A controller with a 7A panel - despite the controller's ability to handle temporary spikes. In some cases it can still be too small to handle edge of cloud events. In solar, headroom is the name of the game.

One of the perhaps somewhat subtle and perhaps secondary causes of increased array output ascribed to "lensing" or irradiance enhancement from reflection by clouds under partly cloudy conditions might be panel temps., or more correctly/completely, changes in panel temps.

If panel temps. under (temporarily) cloudy or intermittent clouds are lower, which seems reasonable, depending on the thermal time constant of the array at that time, the panel efficiencies will be a bit higher as f(cell temp), increasing the output until the array responds to the increased irradiance by increased cell temp.

I haven't estimated the thermal time constant of my array, which will vary mostly with the wind and probably a couple other things, but I'd guess it to be VERY roughly 10-20 min.

With a sudden (step) increase in irradiance, the array will initially be cooler, maybe for a min. or two, than if under the quasi steady state conditions of no clouds. If so, while most of the spike in array output to (unexpectedly ?) high levels is likely due to increased irradiance due to cloud reflection(s), some of that increased output may be due to increased array efficiency from a temporarily lower array temp.

Using the Sandia model for est. cell temp. as f(POA irradiance, wind) and holding the wind vector constant, that model estimates cell temp. increase for my stuff at something like 2.5-2.7 deg. C. per 100W increase in POA irradiance. So, if POA irr. is , say, 300 W/m^2 under cloud cover, and suddenly gets a step change to, say, 1000 W/m^ or more, the lower cell temps. might be, say, 12-15 deg. C. cooler at the start of the transient event for a short period (1-2 min. ??) than if the higher conditions at the "end" of the transient had prevailed as a "steady state" condition . If so, the array eff. may well be 3-5% higher at the beginning of the transient event, adding to the increase in output from the reflection increased irradiance.

Probably just more separating fly fraas form pepper and of little practical significance.

[bcroe]

I haven't estimated the thermal time constant of my array, which will vary mostly
with the wind and probably a couple other things, but I'd guess it to be VERY roughly 10-20 min.

With a sudden (step) increase in irradiance, the array will initially be cooler, maybe for a min. or two, than if under the quasi steady state conditions of no clouds. If so, while most of the spike in array output to (unexpectedly ?) high levels is likely due to increased irradiance due to cloud reflection(s), some of that increased output may be due to increased array efficiency from a temporarily lower array temp.

Using the Sandia model for est. cell temp. as f(POA irradiance, wind) and holding the wind vector constant, that model estimates cell temp. increase for my stuff at something like 2.5-2.7 deg. C. per 100W increase in POA irradiance. So, if POA irr. is , say, 300 W/m^2 under cloud cover, and suddenly gets a step change to, say, 1000 W/m^ or more, the lower cell temps. might be, say, 12-15 deg. C. cooler at the start of the transient event for a short period (1-2 min. ??) than if the higher conditions at the "end" of the transient had prevailed as a "steady state" condition . If so, the array eff. may well be 3-5% higher at the beginning of the transient event, adding to the increase in output from the reflection increased irradiance.

Probably just more separating fly fraas from pepper and of little practical significance.

That info might be of use in the ability to present real numbers, as opposed to the wild guesses
floating around a discussion. It was me who actually measured output under a full moon and
the ability to electrically melt snow off a panel. Based on the complete failure of the last, my
guess is the time constant is a good order of magnitude less than "10-20 min". Who wants to
run this experiment?

And if that is a real influence, guess the wind might be too. That experiment will be harder to
organize, but it makes me want to at least put a panel temp readout on my system, along with
the ambient like those indoor-outdoor sets. Lets see, 0 wind gives delta A degree rise at full
power, 10 mph wind gives delta B....... Bruce Roe

[posplayr]

That info might be of use in the ability to present real numbers, as opposed to the wild guesses
floating around a discussion. It was me who actually measured output under a full moon and
the ability to electrically melt snow off a panel. Based on the complete failure of the last, my
guess is the time constant is a good order of magnitude less than "10-20 min". Who wants to
run this experiment?

And if that is a real influence, guess the wind might be too. That experiment will be harder to
organize, but it makes me want to at least put a panel temp readout on my system, along with
the ambient like those indoor-outdoor sets. Lets see, 0 wind gives delta A degree rise at full
power, 10 mph wind gives delta B....... Bruce Roe

It is perhaps, little known( but with contemplation obvious) fact that you can block all the uncertainty that the world's entropy can generate with a measurement

Measuring the aggregate power output v.s. nominal wind speed will likely show definite trends not requiring a CFD analysis around the panels.

[J.P.M.]

That info might be of use in the ability to present real numbers, as opposed to the wild guesses
floating around a discussion. It was me who actually measured output under a full moon and
the ability to electrically melt snow off a panel. Based on the complete failure of the last, my
guess is the time constant is a good order of magnitude less than "10-20 min". Who wants to
run this experiment?

And if that is a real influence, guess the wind might be too. That experiment will be harder to
organize, but it makes me want to at least put a panel temp readout on my system, along with
the ambient like those indoor-outdoor sets. Lets see, 0 wind gives delta A degree rise at full
power, 10 mph wind gives delta B....... Bruce Roe

Part of such a discussion might start with common agreement about what is meant by "time constant" with respect to a panel or cells in a panel.

I'd suggest one way to think of it may be to assume, at least as a 1st simplifying approx., that the panel is uniform in temp. and a "lumped mass" time constant == "tau" = (M*c/h*A) , where M = panel mass, c = the gross (blended or average) panel spec. heat capacity, h = the combined effective convective and radiation heat transfer coefficient and A = the effective surface area. For simplicity, assume the effects of conduction between the panels and racking is small and therefore can be ignored, at least initially.

(h*A) can then be combined into an effective heat transfer coeff. == U., and (M*c) can be combined into an effective "thermal mass" == M,(t).

--->>> "tau" = (M(t))/U.

Using English, or customary units, the "blended" or average heat capacity for, say, a 40 lbm panel might be (.2 BTU/lbm. deg. F) * (40 lbm) = 8 BTU/deg.F. = M,(t).

"h" might be something like 3 BTU/hr. ft.^2 deg.F., and variable, depending on a lot of stuff, not the least of which is the wind.

The 2 sided surface area = A is about 35 ft.^2 or so. That would estimate as U = h*A ~ (3 BTU/hr. ft.^2 deg. F.) * (35ft.^2)~ 105 BTU/hr. deg. F.

All this (as an example) gives, "tau" = M,(t)/U ~~ (8 BTU/deg.F)/105 BTU/hr. deg. F. ~~ .08 hr. ~~ 4 or 5 min.

So, MAYBE not quite the order of magnitude less than you might suggest, but on cursory, or back of the envelope calc, probably a fair amount more than I SWAGGED.

I'd suggest that still air would drop h and thus U by about half or so, and so perhaps doubling the time constant. An educated guess from all the time I've spent measuring and recording panel temps. and other variables, as well as about half of my life designing and being around heat transfer equipment might lead me to think a 10 MPH wind would increase U by maybe half again or so with a corresponding decrease in the time constant.

This is all a crap shoot anyway, and while it may have some interest in an academic sense, it is perhaps no more than a footnote or back pocket info with respect to increase array output due to clouds or "lensing".

Take what you want of the above. Scrap the rest.

[bcroe]

Part of such a discussion might start with common agreement about what is meant by "time constant" with respect to a panel or cells in a panel.

I'd suggest one way to think of it may be to assume, at least as a 1st simplifying approx., that the panel is uniform in temp. and a "lumped mass" time constant == "tau" = (M*c/h*A) , where M = panel mass, c = the gross (blended or average) panel spec. heat capacity, h = the combined effective convective and radiation heat transfer coefficient and A = the effective surface area. For simplicity, assume the effects of conduction between the panels and racking is small and therefore can be ignored, at least initially.

(h*A) can then be combined into an effective heat transfer coeff. == U., and (M*c) can be combined into an effective "thermal mass" == M,(t).

--->>> "tau" = (M(t))/U.

Using English, or customary units, the "blended" or average heat capacity for, say, a 40 lbm panel might be (.2 BTU/lbm. deg. F) * (40 lbm) = 8 BTU/deg.F. = M,(t).

"h" might be something like 3 BTU/hr. ft.^2 deg.F., and variable, depending on a lot of stuff, not the least of which is the wind.

The 2 sided surface area = A is about 35 ft.^2 or so. That would estimate as U = h*A ~ (3 BTU/hr. ft.^2 deg. F.) * (35ft.^2)~ 105 BTU/hr. deg. F.

All this (as an example) gives, "tau" = M,(t)/U ~~ (8 BTU/deg.F)/105 BTU/hr. deg. F. ~~ .08 hr. ~~ 4 or 5 min.

So, MAYBE not quite the order of magnitude less than you might suggest, but on cursory, or back of the envelope calc, probably a fair amount more than I SWAGGED.

I'd suggest that still air would drop h and thus U by about half or so, and so perhaps doubling the time constant. An educated guess from all the time I've spent measuring and recording panel temps. and other variables, as well as about half of my life designing and being around heat transfer equipment might lead me to think a 10 MPH wind would increase U by maybe half again or so with a corresponding decrease in the time constant.

This is all a crap shoot anyway, and while it may have some interest in an academic sense, it is perhaps no more than a footnote or back pocket info with respect to increase array output due to clouds or "lensing".

Well great! If all those constants can be filled in, maybe the TC can be calculated. I prefer
to just take a measurement. Perhaps the temp rise of a panel over ambient in steady operation,
then block the sun and see how long till the panel rise drops to 37% of before. Bruce Roe

[J.P.M.]

Well great! If all those constants can be filled in, maybe the TC can be calculated. I prefer
to just take a measurement. Perhaps the temp rise of a panel over ambient in steady operation,
then block the sun and see how long till the panel rise drops to 37% of before. Bruce Roe

Constants aren't and variables won't. There's a lot of ways to carve a turkey. It all looks the same 2 days later.

The idea was to suggest that the transient effects of variable irradiance on panel temp. --->>> panel eff.. --->>> output may have some bearing on the magnitude of the observed effects of "lensing' by cloud reflections.

[Naptown]

There's a lot of ways to carve a turkey. It all looks the same 2 days later.

The exception perhaps would be corn😛