Adding a third aspect to a two MPPT inverter?

Zenith74 · 10-09-2020 8:33pm #1

I had a 14 panel 5kW East/West system put in this year including a 5kWh battery, loving it so far. Dual MPPT 5kW hybrid inverter. Peak generation is somewhere around 3.5kW but this is rare, so plenty of headroom in the inverter. Generating 6AM-10PM on nice days.

We went East/West because the South facing roof is fairly small. But with the battery, two EVs and both of us working from home we have no problem sell consuming all we generate, even on the best of summer says. So I’ve been wondering about getting some panels up on that South roof.

I’ve been reading various articles stating that in the right circumstances, there is very little loss connecting two aspects to a single MPPT. Anybody any experience with this kind of thing? Or should I just look at trying to squeeze some more onto the E/W?

10-09-2020 10:07pm

Yes if there's no hard shadows dual MPPT is completely pointless.

unkel · 11-09-2020 12:51am

This is a very interesting thread. I have most of my panels facing south on one string on my inverter, but I also have 3 panels facing west, on a separate string on my dual MPPT inverter. I found this west facing string a bit weak and the voltage is below the minimum for the inverter string at times (although it does produce all day), so I connected my single east facing panel to this string as well (with an optimizer - it does get shadow later in the day). This improved the performance of the string throughout the day. So:

Sir Liamalot wrote: »

Yes if there's no hard shadows dual MPPT is completely pointless.

I would tend to believe that

Why are there dual MPPT inverters then? Is the only reason for their existence that it doesn't completely kill production when one of the strings is shadowed at times?

11-09-2020 12:57am

unkel wrote: »

I would tend to believe that

As with most things I say Unkel I've measured it

unkel wrote: »

Why are there dual MPPT inverters then? Is the only reason for their existence that it doesn't completely kill production when one of the strings is shadowed at times?

Steep rooves with chimneys. Multi aspect with partial tree shadows. Marketing asking for more features to make the product attractive.

11-09-2020 1:00am

Panels producing Vmp at zero current do not pull the array down. You have to reverse bias the cells with a hard shadow to justify isolation.
Panels will produce Vmp in almost all light conditions.

unkel · 11-09-2020 1:18am

Sir Liamalot wrote: »

Panels producing Vmp at zero current

Eh? I thought Vmp was the voltage at maximum current by definition? Maybe I got that wrong

Anyway I have measured something this time too. The voltage of my west facing string on its own (not shadowed at any time) varies significantly depending on the situation. At a minimum I get about 84V (about 27-28V per panel) and as a maximum (when the panels produce their full rated wattage) I get about 108V (about 36V per panel).

My inverter has a minimum of 90V per string, and my west facing string was under that regularly (but the string still produced amps so my inverter was nowhere near as fussy as the specs said). I added the east facing panel and obviously the voltage was well over the minimum 90V at all times

11-09-2020 1:26am

Voltage max power

power at zero current = 0W but not negative

The array voltage will vary when connect to an inverter it's adjusting the load to suit max power output; that's what MPPT is.

If you're too borderline on the input specs the inverter will not perform very well. It's like playing 25 with 4 cards. There's no voltage to play with to boost the current.

11-09-2020 1:44am

unkel wrote: »

Eh? I thought Vmp was the voltage at maximum current by definition? Maybe I got that wrong

I think you are thinking Vmp @ STC (standard test conditions) 1000W per meter and 25°C cell temperature.

Vmp normal every day is the under load voltage (varies with input and load).
Voc is the no load voltage.

Zenith74 · 11-09-2020 7:05am

When does a shadow become a ‘hard’ shadow, a panel completely shaded out?

I noticed there’s a bit of shading on the East array early in the morning from the top of a tree, and some on the West late in the evening from a neighbours house. There would be no shading on the smaller South facing roof.

So is it realistic to just feed my East and West strings into the one input on the inverter and run it for a few days, see how the generation compares? Any gotchas to this? I’ll have to look at how things are connected, see what is involved junction box wise.

If you did end up with say two 7 panels strings (East and West) and maybe a 4-5 panel string (South), which two strings would you connect together on the one MPPT?

Also wasn’t aware that optimisers could be added to a small number of panels without other changes to the array/inverter etc. Should I consider some for those panels that catch a bit of shade, if the guys were back up on the roof anyway? Any downside (other than cost)?

Really interesting discussion thanks!

11-09-2020 9:11am

A hard shadow is one which has a defined edge. Sundialesque. A soft shadow is uniform.

Sure connect them all together for science. Stay within the input threshold parameters of the inverter most importantly Voc -20%.

I've a dual string I'm using one input, my array is east & south west.

I'd put 9 panel string one 9 panels string 2 in your scenario.

Optimisers are a bittova gimic. Don't put panels in silly places and you don't need them. When they fail they are very hard to get to or flag they are not working.

unkel · 11-09-2020 10:03am

Sir Liamalot wrote: »

If you're too borderline on the input specs the inverter will not perform very well.

I must be on the ok side of borderline so. System is performing well and I do get current from the west facing string even early in the morning. I'd say the specs of the inverter might be a bit cautious?

Maybe you could glance your eye over them?

Here's the datasheet:

https://static.solartricity.ie/wp-content/uploads/2020/06/Solis-3.6.pdf

MPPT voltage (I presume this is per string?) 90V-520V and I didn't always get the 90V on my west facing 3 panel string (until I connected the east facing single panel)

11-09-2020 5:01pm

I expect it'll perform better with a 410Voc input.

unkel · 11-09-2020 5:42pm

It has about 360V on the south facing string too

11-09-2020 7:05pm

You can't really put them all together as there's no overhead for cold temp uprating. 50:50 split might work but you'll probably have cabling issues.

Arah sher knowing you you'll have more panels before I finish typing. Bung them in the east or don't worry about it. The more voltage the more of a range you give the MPPT to play with the easier it is for it to tune the array to conditions....it'll also help with bird poo instigated inverter dropping out.

When I say it'll perform better with a higher voltage...well of course it will that takes more panels. What I meant was it'll perfrom better per watt installed.

unkel · 11-09-2020 7:42pm

Better if I put 10 * south + 3 * west + 1 * east all on the one string?

The east panel gets some shading, but I have an optimizer attached to it. Or 10 * south + 3 * west on the one string? In that case I can use my east panel for some other projects

And yes, I wish I had the space you have. I'd probably have at least 30 panels up

11-09-2020 7:49pm

You know what!? I'm confusing hard shadows with nightime! (I must be getting old) :pac:

I remembered my experiment about 2 posts ago.

Nah don't put them all together you'll cook the inverter in cold weather if theres a cloud lense.
Parallels in arrays are not so straight forward especially when they're different breeds of panels. It's difficult to implement right.

unkel · 11-09-2020 7:54pm

I wasn't planning on going parallel. Just series. You have me confused now. Is it, or is it not a good idea to have all my panels (that don't get shadow) into one series string? Or would it make little difference?

11-09-2020 8:06pm

There's two things at play Unkel.
Shadows will not pull down the array anymore than the loss of the panel(s) output that are under shade and MPPT cannot help with this.
Dual string inverters...I don't know what their function is.

Yes you can in theory put them all together but in your particular case you cannot because you do not have enough overhead on the inverter tolerance.
Yes they will be more efficient with a higher voltage but 13 panels over one strings will not be as effective as 14 panels.

Have a read of the clicky. School of I have tried this myself is unanimous.

Zenith74 · 12-09-2020 1:03pm

Sir Liamalot wrote: »

A hard shadow is one which has a defined edge. Sundialesque. A soft shadow is uniform.

Sure connect them all together for science. Stay within the input threshold parameters of the inverter most importantly Voc -20%.

I've a dual string I'm using one input, my array is east & south west.

I'd put 9 panel string one 9 panels string 2 in your scenario.

Optimisers are a bittova gimic. Don't put panels in silly places and you don't need them. When they fail they are very hard to get to or flag they are not working.

Thanks!

I’m a bit limited on roof space and the shading is only at certain times of year/day, and of course the panels are already up now :-). Guess I was wondering if I was getting somebody back on the roof to add more panels, should I have them toddle over there and add optimisers to the couple of panels that sometimes see shade?

On the 9+9 panel setup, you’re suggesting 7East panels plus 2South panels on one string, 7West plus 2South ok the other right? Are those 2South panels in each case going in series or parallel with the 7 existing panels?

In terms of sticking the two strings on one inverter input for science :-), presume I can get a junction box of some sort that takes the four DC connectors from the two strings and gives me a pair of outputs? Before or after the fireman switch I guess makes little a difference? And yeah before I go mad, check the voltages.

I’ll have a chat to the installers (NextGenPower) and see what it would cost to get a few more panels up there anyway.

Thanks again!

unkel · 12-09-2020 1:23pm

Zenith74 wrote: »

Are those 2South panels in each case going in series or parallel with the 7 existing panels?

You shouldn't parallel connect a series string of 2 panels to a series string of 7 panels, they have very different voltages

12-09-2020 4:40pm

All series legs ought to have identical Imp...unless like me you don't really care because it was close enough and your array has 7 breeds of panels because they were such a good deal.
The lowest Imp will current limit the array.

Forget about the optimisers they're not going to do anything significant except add complexity, expense and a failure point.
Probably all you need Zenith is a pair of branch connectors at the inverter.

unkel · 18-09-2020 3:41pm

Sir Liamalot wrote: »

Dual string inverters...I don't know what their function is.

Sir Liamalot wrote: »

When I say it'll perform better with a higher voltage...well of course it will that takes more panels. What I meant was it'll perfrom better per watt installed.

Tried that and nope, performs way worse when I combine both my south and my west strings into the one string in series.

Producing 1600W at 10:30 this morning in hazy sunshine as separate strings and producing 700W as one string. None of the panels had shadow

And before anybody mentions it: I know it's not a great idea to connect and disconnect panels to your inverter when they are under load. I risked my life for science here :pac:

So now we know that indeed you should put different orientation strings of panels on separate inputs on your inverter...

27-09-2020 8:22pm

Me too. New tree shadow is pulling down the array 50% for the sake of two panels outtov 10.

27-09-2020 8:25pm

unkel wrote: »

And before anybody mentions it: I know it's not a great idea to connect and disconnect panels to your inverter when they are under load. I risked my life for science here :pac:

:rolleyes: Get a PV disconnect switch dude! You're not supposed to pull the MC4 apart because they arc like bejaysus....err...you could not touch it if you start one...that works too.

27-09-2020 8:25pm

Oh and while I'm eating my own words. My latest new genset is 79% round trip efficiency.

unkel · 27-09-2020 9:02pm

Sir Liamalot wrote: »

:rolleyes: Get a PV disconnect switch dude! You're not supposed to pull the MC4 apart because they arc like bejaysus....err...you could not touch it if you start one...that works too.

I've connected many MC4s, never seen an arc. But indeed I should get my finger out and get a few DC disconnect switches between the strings and the inverter. Rather than learning this lesson the hard way

27-09-2020 9:17pm

Put a blanket over the sunny side first.

It's a highly youtoobable scenario.

unkel · 27-09-2020 9:23pm

I tend to only connect / disconnect them in the evening / when it's (almost) dark and when the amps are (near) zero anyway

air · 03-10-2020 3:59am

Zenith74 wrote: »

If you did end up with say two 7 panels strings (East and West) and maybe a 4-5 panel string (South), which two strings would you connect together on the one MPPT?

Put the 2 strings of 7 panels in parallel into 1 input, put the 4-5 panel South array into the other input.
I'd recommend instaling a diode in series with the East & West strings before the inverter, you can buy them pre installed with MC4 connectors so plug & play.

air · 03-10-2020 4:13am

Sir Liamalot wrote: »

There's two things at play Unkel.
Shadows will not pull down the array anymore than the loss of the panel(s) output that are under shade and MPPT cannot help with this.

This is incorrect, the shaded panel voltage will not be pulled down, but it's current output will be.

If Unkel puts 2 panels from a different orientation in series with the Southern array, the current of the entire series string will be clipped to that of the panel with the lowest output current.
This is the whole reason that optimisers exist.

Sir Liamalot wrote: »

Dual string inverters...I don't know what their function is.
.

Their function is fairly clear, they allow independent MPPT tracking of 2 completely independent arrays that do not have to have anything in common in terms of orientation, panel type, count or characteristics.

They reduce the need for a second discrete inverter in very many scenarios.

bobbyy gee · 03-10-2020 4:21am

change the tilt of panals and add panals on south roof

https://www.energysage.com/solar/101/impact-of-roof-orientation/

bobbyy gee · 03-10-2020 4:29am

mppt inverters

https://forum.solar-electric.com/discussion/352411/sharing-pv-input-with-two-different-mppt-inverters

unkel · 03-10-2020 10:40am

air wrote: »

I'd recommend instaling a diode in series with the East & West strings before the inverter, you can buy them pre installed with MC4 connectors so plug & play.

Any link to one?

air · 03-10-2020 4:09pm

unkel wrote: »

Any link to one?

https://m.aliexpress.com/item/4001189061502.html

unkel · 03-10-2020 4:58pm

air wrote: »

Put the 2 strings of 7 panels in parallel into 1 input, put the 4-5 panel South array into the other input.
I'd recommend instaling a diode in series with the East & West strings before the inverter, you can buy them pre installed with MC4 connectors so plug & play.

Bought a couple of those. Currently don't have any parallel strings but might come in handy. Thanks for the link!

So if you want to parallel 2 strings (different orientation or not), do you recommend only to do that with strings of the same number of identical panels or would slightly different panels (with slightly different max output) still be ok?

03-10-2020 8:51pm

air wrote: »

This is incorrect, the shaded panel voltage will not be pulled down, but it's current output will be.

thanks air. i was confusing issues. so we could benefit from optimisers [i'm thinking of contactors myself...or move the array...hrrmm]

...but we don't need diodes because the voltage will not be pulled down?

air · 03-10-2020 10:41pm

Sir Liamalot wrote: »

...but we don't need diodes because the voltage will not be pulled down?

I think we've been through this before.
There's virtually no downside to using diodes, barring the very tiny efficiency hit.

They protect against one string reverse biasing the other & potentially damaging it in a freak circumstance like a sheet flying off a line onto the array (totally shading it) or something equally unlikely.

You could very well run for decades without ever encountering an issue, up to the user whether they think it's worthwhile or not.

air · 03-10-2020 10:54pm

unkel wrote: »

So if you want to parallel 2 strings (different orientation or not), do you recommend only to do that with strings of the same number of identical panels or would slightly different panels (with slightly different max output) still be ok?

You really want to match the same number and type of panels in parallel strings to get the best tracking performance.
If you're mixing panel types in different strings you ideally want the total VMP of each string to be as close as possible.
Diffferent current ratings between parallel strings is not an issue.
On the contrary mixing panel types in a single series string rarely makes sense.
The max current of the string will be clipped to that of the lowest rated panel.

03-10-2020 11:23pm

unkel · 03-10-2020 11:46pm

air wrote: »

On the contrary mixing panel types in a single series string rarely makes sense.
The max current of the string will be clipped to that of the lowest rated panel.

Damn, didn't think of that. That makes a significant difference now I have upgraded some of my panels to 355W that have a current at max power of 10.6A vs my older 290W panels that have a current at max power of just 9.2A

Thanks for your comment. I have choices to consider and changes to make now...

air · 04-10-2020 12:41am

Sir Liamalot wrote: »

It's generally considered poor form to link a video without any commentary or explanation of relevance.

To save anyone else the bother, the guy is making the point that optimisers are pointless for series strings with common orientation & shading where they can only effectively compensate for differences in panel output due to manufacturing tolerances.

I don't think this is really news to anyone & the reason that string inverters rightly dominate the market.

He makes a further point that optimisers are not much better than panel diodes for landscape panels (a minority of installations) where the shade moves from bottom to top.
I would expect that optimisers are far more likely to be installed to deal with shading from chimneys & other irregular objects which don't create this type of shading.
Overall not much useful information presented in the video.

04-10-2020 8:52pm

air wrote: »

I think we've been through this before.
There's virtually no downside to using diodes, barring the very tiny efficiency hit.

why are they needed when the effect they protect against has an opposing voltage at zero current and that current is traveling through conducting bypass diodes? how do you reverse bias an opposing conducting diode?

optimisers kindov make sense in retrospect in regards to moving shadows in that they reduce string loss and accommodate multi-aspect mounting. i dunno...i think i'll find a simplier solution myself.

unkel just derate the array in accordance to applying the lowest panel lmp to the array lmp. l reckon you're still ahead on price per watt.

i have two 60v 5a lmp panels in parallel in my array making a 60v 10a in series with a loada 8.5ish lmp 30v - 40v jobbers.

my array is oversized to the inverter. l still hit the peg anytime it's considerably sunny.

air · 04-10-2020 9:17pm

I don't think you fully understand how bypass diodes work, they don't provide any blocking protection when a panel is reverse biased.

Sir Liamalot wrote: »

why are they needed when the effect they protect against has an opposing voltage at zero current and that current is traveling through conducting bypass diodes?

The unlikely scenario I proposed (total coverage of an individual panel) means that the shaded panel would produce zero (or very low) voltage and the unshaded string would then reverse bias it.

Sir Liamalot wrote: »

how do you reverse bias an opposing conducting diode?

Not sure what you mean, bypass diodes are fitted in the same direction as the cells, they only conduct when some or all of a panel section is shaded - thus "bypassing" the current past the shaded section.
They don't provide any alternative conduction path to protect a panel when a panel is reverse biased as they are reverse biased themselves in this scenario!

Sir Liamalot wrote: »

optimisers kindov make sense in retrospect in regards to moving shadows in that they reduce string loss and accommodate multi-aspect mounting. i dunno...i think i'll find a simplier solution myself.

They should be a last resort ideally but in certain circumstances are the best option. A bit like having an additional MPPT inverter input for every optimiser that is added.

04-10-2020 9:42pm

air wrote: »

I don't think you fully understand how bypass diodes work,

:pac: probably not

air wrote: »

they don't provide any blocking protection when a panel is reverse biased.

conceivable...i haven't thought about it to be honest

air wrote: »

The unlikely scenario I proposed (total coverage of an individual panel) means that the shaded panel would produce zero (or very low) voltage

i have measured this myself and shaded panels produced vmp give or take at zero current give or take

air wrote: »

and the unshaded string would then reverse bias it.

if it had no voltage but i've not measured the phenomenon despite trying to contrive it

air wrote: »

Not sure what you mean, bypass diodes are fitted in the same direction as the cells, they only conduct when some or all of a panel section is shaded - thus "bypassing" the current past the shaded section.

so they are conducting a forward voltage when the panel is shaded, the panel next to it is forward biased, with a forward voltage through a diode feeding it, where and how can it be reverse biased?

air wrote: »

They don't provide any alternative conduction path to protect a panel when a panel is reverse biased as they are reverse biased themselves in this scenario!

as i say i've never seen a reverse biased panel. cells yes, cell strings with bypass diodes no.
i'd love to see a demonstration of them helping.
all well and good having small losses on a 400v circuit and a low risk added failure point.
in elv land diodes are a very significant waste of power.

air wrote: »

They should be a last resort ideally but in certain circumstances are the best option. A bit like having an additional MPPT inverter input for every optimiser that is added.

i've half an idea involving a chainsaw, some vehicle mounted panel bartering, the second mppt inverter input and a spdt contactor maybe with a photocell.

i'll worry about it later. i've another 12kwh i pulled outtov solar storage and 6kva in the post to be getting on with.

air · 04-10-2020 10:30pm

Sir Liamalot wrote: »

i have measured this myself and shaded panels produced vmp give or take at zero current give or take

Covered I said, not shaded.
Dark sheet blows off the line, can of paint spilled while painting the chimney etc.
All highly unlikely, but not impossible.

Sir Liamalot wrote: »

if it had no voltage but i've not measured the phenomenon despite trying to contrive it

As above.

Sir Liamalot wrote: »

so they are conducting a forward voltage when the panel is shaded, the panel next to it is forward biased, with a forward voltage through a diode feeding it, where and how can it be reverse biased?

as i say i've never seen a reverse biased panel. cells yes, cell strings with bypass diodes no.
i'd love to see a demonstration of them helping.
all well and good having small losses on a 400v circuit and a low risk added failure point.
in elv land diodes are a very significant waste of power.

Take 4 panels, 2 x 2 series strings.
Cover one panel completely, compare output voltage of both strings. The one with the covered panel should be about half the other one (bypass diodes conducting)
Now (if you dare risk it) connect the two strings in parallel. The output from the unshaded string will flow backwards through the two with the shaded panel.
* caveat, I can't be sure sure of the reverse breakdown voltage required for reverse flow to start, it might take more than 2 panels in series depending on the panels / cell type, light intensity etc.

Sir Liamalot wrote: »

i have two 60v 5a lmIp panels

Have you an Irish supplier for these by any chance? I have 1 but need a few more.

Sir Liamalot wrote: »

in elv land diodes are a very significant waste of power.

Nobody is talking ELV here though and at the low current levels typical in Irish conditions the forward volt drop is going to be <0.5V nearly all the time in practice.

You should really be going at least 24V and preferably 48V for your vehicles at this stage mind you.

Running 32V and 48V in lithium myself, 12V/24V down converters for loads that need it, I would never use lead in anything mobile.
Drifting way OT sorry!

04-10-2020 11:53pm

air wrote: »

Covered I said, not shaded.
Now (if you dare risk it) connect the two strings in parallel. The output from the unshaded string will flow backwards through the two with the shaded panel.

oh I dare. in the name of science. could I be bothered is another more poignant issue.
so if there' no parallels with partial pitch blackouts then it's not an issue?

I'm gonna document pwm kicking mppt bottom with panel to battery voltage suitablity first...or not...

air wrote: »

Have you an Irish supplier for these by any chance? I have 1 but need a few more.

nah mine are oooold. 90 cells iirc.
250w rated 275w performing [until i derated them with my install]
i'm happy to trade for 30 cells all black whatevers.

air wrote: »

Nobody is talking ELV here though and at the low current levels typical in Irish conditions the forward volt drop is going to be <0.5V nearly all the time in practice.

i'm doing 80amp solar installs these days. no diodes.

air wrote: »

You should really be going at least 24V and preferably 48V for your vehicles at this stage mind you.

i go whatever is a good deal. 12v, 24v, 48v and 60v. it's hard get away from 12v even when it's bonkers not to. the clients don't like having to change alternators.

i'm grid tying a truck; 4kva hybrid with 2.5kva synchronised backup and 1.5kw solar rough spec atm...trying to convince myself it doesn't need 3.5kva 3 phase.

air wrote: »

Running 32V and 48V in lithium myself, 12V/24V down converters for loads that need it, I would never use lead in anything mobile.

easy lose a quarter ton in a truck [or a barge]. i've 15kwh of freecycled lead li-ion can't compete with that.
meh, i use both. li-ion is good at being tenacious and light. i use it to torque fill the marathon running lead.

air wrote: »

Drifting way OT sorry!

i'm sure they'll get over it...someone will striaighten us out before long.

Adding a third aspect to a two MPPT inverter?

Comments