MPPTs more efficient at higher end of input vs. lower?

[OM617YOTA]

Back story:

My system is two strings of 410w bifacial panels, each string is 2 panels in series, each string(was) feeding it's own LV2424 AIO. I'd planned to tilt the panel strings in opposite directions, hence the two two strings on separate MPPTs. Turns out I wasn't doing that, the panels stay on the same plane all the time. Frequently here in the PNW, we have enough light to kick on the MPPTs, but not enough to harvest power from, and during those times I'm just burning power to keep the MPPTs active. Each MPPT is rated at 2kw, one would handle the full output of all four panels, so I paralled both strings and fed them into one MPPT to save from burning power keeping the second MPPT active in low light conditions.

Immediately before paralleling the panels, I was bringing in ~300w per MPPT, 600w total. After doing so, that bumped up to over 800w with the same panels, and stayed there for a long time. Paralleling the strings took two minutes, and I was drawing ~2600w from the system, so battery state of charge wasn't a factor. I wish I had a light meter to take actual measurements instead of "welp, it looked the same brightness and cloud cover outside".

Question:
Are MPPTs known to be more efficient or effective when paneled at the higher end of their input, vs. the lower end? Is the 30% increase in power harvested due to a known phenomenon? I'm leaning towards eye witness evidence being the least reliable form of evidence, my eyes are not light meters, and if I got more power it's because there was more light.

Right now, light conditions are very low. Normally I'd be burning ~20w keeping the MPPTs active alone. Instead the paralleled strings are bringing in 16w. A pittance, but it's a gain instead of a loss, and I'll take it.

 

[AntronX]

It looks like typical Voltronic design with solar input being boosted to 200vdc internal bus voltage via unisolated boost dc-dc converter. Connecting panels in series for higher PV voltage and less current is more efficient.

 

[OM617YOTA]

It's the same input voltage, though. Still 2 panels in series, just two parallel strings of two panels in series.

 

[Zwy]

Back story:

My system is two strings of 410w bifacial panels, each string is 2 panels in series, each string(was) feeding it's own LV2424 AIO. I'd planned to tilt the panel strings in opposite directions, hence the two two strings on separate MPPTs. Turns out I wasn't doing that, the panels stay on the same plane all the time. Frequently here in the PNW, we have enough light to kick on the MPPTs, but not enough to harvest power from, and during those times I'm just burning power to keep the MPPTs active. Each MPPT is rated at 2kw, one would handle the full output of all four panels, so I paralled both strings and fed them into one MPPT to save from burning power keeping the second MPPT active in low light conditions.

Immediately before paralleling the panels, I was bringing in ~300w per MPPT, 600w total. After doing so, that bumped up to over 800w with the same panels, and stayed there for a long time. Paralleling the strings took two minutes, and I was drawing ~2600w from the system, so battery state of charge wasn't a factor. I wish I had a light meter to take actual measurements instead of "welp, it looked the same brightness and cloud cover outside".

Question:
Are MPPTs known to be more efficient or effective when paneled at the higher end of their input, vs. the lower end? Is the 30% increase in power harvested due to a known phenomenon? I'm leaning towards eye witness evidence being the least reliable form of evidence, my eyes are not light meters, and if I got more power it's because there was more light.

Right now, light conditions are very low. Normally I'd be burning ~20w keeping the MPPTs active alone. Instead the paralleled strings are bringing in 16w. A pittance, but it's a gain instead of a loss, and I'll take it.

My experience with the GW 3K 24V unit which is basically the same as the LV2424 is the MPPT will wake up when string voltage is high enough. Once a load is placed on the panels, if the voltage drops the MPPT will drop out, shut down and then restart. It will keep doing this until PV production is high enough. Part of this is the MPPT algorithm I believe, it starts searching for a power point and loads the panels down with the resulting drop in panel voltage.

By placing the 2 strings in parallel on one MPPT, there probably is enough power to keep the MPPT on. Voltage will not drop as drastically as you have 4 panels supplying watts and the MPPT can determine a power point. However you said you were getting 300W per string so it isn't really a shut down problem like I related but could be a power point search problem. Perhaps you had a poor connection or were both strings in agreement?

If you are marginal on PV and it you don't need the inverter on, the MPPT will charge with the inverter off. I run my truck camper with a remote switch in the galley, this allows us to turn the inverter on/off easily. The MPPT will charge the batteries when there is PV available and any production is not wasted idling the inverter. If your loads are constant, then this will not be an option.

I am confused how you could be getting 2600W from four 410W panels even if these are bifacial. Must be some pretty good gain.

 

[OM617YOTA]

By placing the 2 strings in parallel on one MPPT, there probably is enough power to keep the MPPT on. Voltage will not drop as drastically as you have 4 panels supplying watts and the MPPT can determine a power point. However you said you were getting 300W per string so it isn't really a shut down problem like I related but could be a power point search problem. Perhaps you had a poor connection or were both strings in agreement?

Yes, both strings were in agreement. I should switch all four panels to the other MPPT and see if I get similar results.

If you are marginal on PV and it you don't need the inverter on, the MPPT will charge with the inverter off. I run my truck camper with a remote switch in the galley, this allows us to turn the inverter on/off easily. The MPPT will charge the batteries when there is PV available and any production is not wasted idling the inverter. If your loads are constant, then this will not be an option.

Yes I do this through most of the winter. The 100w idle consumption of 2x LV2424s is ridiculous for my small system. The idle consumption is my single biggest complaint about the equipment I've chosen, and would be the main reason I'd choose other options if redoing the build.

I am confused how you could be getting 2600W from four 410W panels even if these are bifacial. Must be some pretty good gain.

Woops, wasn't clear there. The 2600w was inverter output, with the batteries making up the difference.

I mentioned the 2600w inverter output because if the batteries had been nearly full and the MPPTs were cutting output to 600w total to keep battery voltage in check, then the battery could be drawn down while I had solar disconnected, and when reconnected then the single MPPT could be charging at a higher rate to make up the difference. Not a factor, as I had a 2600w load going the entire time, the MPPT(s) were welcome to pull in all the power they possibly could.

 

[billvon]

Question:
Are MPPTs known to be more efficient or effective when paneled at the higher end of their input, vs. the lower end? Is the 30% increase in power harvested due to a known phenomenon?

MPPTs are most efficient when the output voltage is similar to the input voltage. For 48V battery systems that means lower voltages are more efficient for the MPPT (although of course you may see higher I2R losses in your wiring.) For 400V systems, setting your system voltage to ~360 volts will get you the best efficiency.

For pure grid tie inverters it's a bit more complicated since there is an internal link voltage, and you're not always sure what it is.

This is true of all switchmode power converters BTW, not just MPPT's.

 

[HarryN]

A good quality MPPT controller will convert the (volts ) x ( amps ) = watts from the panels into the "correct" ( volts ) x ( amps) = watts for the battery pack.

In overcast conditions, the panel Voc can hit a high enough voltage to trigger the controller "on", but then the Vmp of the panel takes over and this can drop below the required ( V panel ) - ( V battery) = delta required to operate.

What is ideal is to have the panel voltage fairly high so that when it triggers on, there is enough power to actually run and stay at a high enough panel voltage to keep running.

What fools people (including me) is that there was a scientific paper that claimed that Vmp did not vary with light intensity, but it failed to take into account the true overcast conditions that are seen, as well as the very real effect on the spectrum of light on the panels.

Short answer - get the voltage up.

 

[Bop]

MPPT charge controllers work best at the HIGH end of their PV input voltage range (NOT the low end- indeed unless you get the PV voltage under load about 3-5v above the battery charge voltage, they won't even enter the MPPT mode at all and basically become a very expensive PWM controller lol)

Although I don't recommend going right to the PVmax voltage rating- as in certain weather conditions, panels can exceed their 'label' rating by 10-15% (those ratings ONLY apply at STC (standard test conditions) when the cell temps inside the panel are 25C, the light illumination is 1000w/m^2 and the atmospheric mass is 1.5 (its written on every panel on the specs label...) if any of those is different, then the panels output will be different...)


At any other time, the outputs could be less- or more...

Personally I recommend leaving a 20% 'headroom' below the PVmax rating-- ie on a 150v PVmax rated controller, the combined Voc's of the panels should be at or under 120v (80/100 x 150v = 120v)

eg my own panels are 38.12v Voc (thats the label above lol) and I run 3 in series, so their combined Voc is 3x38.12 = 114.36v- just under the 120v limit I advise using for longterm safety of a 150v PVmax controller like my own

This will give the best 'low light' level performance with an MPPT controller, and still have a good safety margin on it in high light level situations...

Some people seem to get confused about the charge controller current limit (eg mine are 60A) this is NOT the input current limit from the panels, but the maximum OUTPUT current of the controller so it will be limited to that output no matter what the potential input is. Your panels don't 'force current out' but are capable of outputting anything from zero amps UP TO their rated limits, if the controller has hit its maximum output (60A) in my case) then the panels will output whatever current the charge controller is demanding -if the light conditions allow them to...

If the light conditions are poor, then the charge controller will make do with whatever the panels can provide... as long as the loaded panel voltage remains above the charging voltage needed (in most cases, some expensive charge controllers can drop into a buck/boost mode where even if the panel voltage is below the battery bank voltage they can step it up to whatever is required)- this is where 'overpanelling is a clear winner- in low light levels you can maximise the output of your charge controller...

'Some' people think overpanelling is bad and wasting power, but it isn't- it is simply utilising the full potential of your existing charge controller when you have prolonged poor weather conditions... getting full power in good conditions is unimportant (especially for those offgridding) and quite easy- what you need is sufficient power in poor conditions..., and overpanelling with a MPPT charge controller running near its upper end of its input voltage is the most cost effective way of doing that...

The 30% improvement thing is in comparison to a PWM charge controller (or a MPPT running down at the lowest end of its voltage range)- using my panels ratings from above on a '24v' nominal battery bank (they aren't really suited for this as they are ex gridtie)- the charge voltage on a nominal 24v L/A battery bank is usually listed as 27.6v for most types- my panels maximum Imp is 8.23A, so the maximum output you can get on a PWM controller is 27.6v x 8.23A = 227.1W- from a 250W panel (at STC)
Where a MPPT will run the panel at its Mp ratings ie Vmp and Imp ie 30.43v x 8.23A = 250.4W- an improvement of 23.3W

If the panel voltage is badly mismatched, then using a PWM controller could literally lose you half or more of the panels output!!!
Say you bought one of these 420W panels (cause 'panels are all the same aren't they' as many newbies think)

Hooked up to a PWM controller on your 24v nominal battery bank- you are going to be badly disappointed in its performance- the Imp for the 420W panel is only 5.89A (even into a direct short circuit, its maximum output is the Isc ie 6.33A) so the PWM will limit the charge voltage at the same 27.6v, so thats 27.6V x 5.89A = 162.6W....

under 163W- from a 420W panel!!!!


Hook the same panel up to a MPPT charging the same battery bank, and you get the Vmp and Imp figures ie 71.3V x 5.89A = 419.96W...


That's a 257.4W improvement- simply by changing from a PWM to a MPPT on a poorly selected panel...


Here's my own panels- at dawn, no direct light on them at all, and already up at at over 50v (they run at usually about 87-90v in good conditions, putting 40A into the 12v battery bank)- 52v from the panels and already putting in half an amp into the battery bank- literally at dawn with the sun not even completely above the horizon yet and shielded by trees...

A PWM voltage matched system won't even start to put any charge into the battery bank for a couple of hours... and by that time the MPPT is really ramping up...
Assuming you had a PWM on there instead, and the three panels in parallel- at this point, each panel is only making 17.3v- but you need 27.6v to charge the battery- hmm thats not gunna work is it????
(even a MPPT with the three panels in parallel has the same issue- it needs that 3-5v above the charging voltage to run the MPPT circuits, or else it drops back into PWM mode- and that 17.3v is just too low to do either...)

 

[OM617YOTA]

On my little bobble that I'm wondering about, the voltage was the same before and after I paralleled the strings of panels, just pulling more amps. MPPT both before and after, no PWM involved. 600w before at ~83v, 800w after at ~83v.

800w out of a 1600w array isn't particularly low light conditions, especially for here in the PNW in February.

 

[Bop]

On my little bobble that I'm wondering about, the voltage was the same before and after I paralleled the strings of panels, just pulling more amps. MPPT both before and after, no PWM involved. 600w before at ~83v, 800w after at ~83v.

800w out of a 1600w array isn't particularly low light conditions, especially for here in the PNW in February.

One thing you have to remember is that a '1600w' array is only that at STC ie panels at 25C, illumination at 1000W/m^2 and an atmospheric mass of 1.5- I'd have to look up the actual figures for that area, but I am betting that solar illumination is low there atm ie the sun is fairly low in the sky... and the A.M is waaay up...

As to why you are seeing 200w more I would think that your system is operating at the bottom end of the panels ability to supply current- as a panel is loaded, its output voltage drops and the MPPT controller will 'hunt around' looking for the 'sweet spot' that produces the most output power from the panels under those light conditions- I am surprised that after the paralleling, the output voltage from the panels didn't change slightly though
Sounds like your panels are JA Solar panels from the specs you described (410w, about 86 Vmp) if so this is their power curve for the various light levels

If so the controller isn't doing a particularly good job of keeping them on the 'knee' of the P/V curve which is down about 41v per panel ie 82v in a series pair, but at low light levels, panels can exhibit weird curves when loaded- especially right at the droppoff point of the V/I curve...

Panels outputs aren't quite clear cut especially as you were drawing a heavy load (in relationship to the panels ability to provide power- and a MPPT controller in low light conditions suddenly having effectively double the power available to it might shift both the Mpp of the panels (although I would have expected to see a change in the panel voltages, that close to the knee, even a small voltage shift can have a big effect on performance) and also on its own internal operating parameters (which only the manufacturer knows, and likely only a very small number of people there even...)
You have to remember that although to the system as a whole you haven't changed anything in terms of total panel wattage, to the controller, you have effectively doubled its available power to it, and that can put it in a whole new area of the Mpp curves...
(this is again a reason for 'overpanelling' in areas with known extended periods of poor weather- MPPTs work well in poor conditions, but not at the bottom end of their ratings as well as they do at the top end of their ratings...
I am hesitant to suggest too much overpanelling however, as their specs are pretty screwed up lol (to put it mildly)

But (ignoring the mains charger bit) the charge specs are these

Which doesn't jibe with the PV max figures at all- so what exactly they are doing- only they know...

 

[OM617YOTA]

that can put it in a whole new area of the Mpp curves...
(this is again a reason for 'overpanelling' in areas with known extended periods of poor weather- MPPTs work well in poor conditions, but not at the bottom end of their ratings as well as they do at the top end of their ratings...

I think this is putting your finger on the solution. Basically exactly what I was asking.

I am hesitant to suggest too much overpanelling however, as their specs are pretty screwed up lol (to put it mildly)
View attachment 197086
But (ignoring the mains charger bit) the charge specs are these
View attachment 197088
Which doesn't jibe with the PV max figures at all- so what exactly they are doing- only they know...

80A charging at 27v is 2160w. Seems perfectly reasonable to me. Wish I had enough panels to test that.