heating water with PV

[winnie]

I'd assume that the inverter has a maximum input voltage rating and a maximum input current rating, and that is what defines the maximum DC/AC ratio.

If there is a separate maximum DC power value which is lower than 'max string voltage times max Isc', I think you'd be hard pressed to exceed that max DC power number without also exceeding one of the other numbers.

 

[ggunn]

I'd assume that the inverter has a maximum input voltage rating and a maximum input current rating, and that is what defines the maximum DC/AC ratio.

If there is a separate maximum DC power value which is lower than 'max string voltage times max Isc', I think you'd be hard pressed to exceed that max DC power number without also exceeding one of the other numbers.

Most (maybe all) the data sheets for the commercial inverters I design with (I haven't done any resi PV work in several years) show a maximum allowable connected STC DC power in addition to maximum DC current and voltage numbers.

 

[wwhitney]

I'd assume that the inverter has a maximum input voltage rating and a maximum input current rating, and that is what defines the maximum DC/AC ratio.

Here's some numbers from the spec sheet of an inverter I selected semi-randomly, the Fronius Primo 6.0-1:

Recommended PV Power 4.8-9.3 kW
Max MPPT input current 2 x 18A (2 inputs)
Max input SCC 2 x 27A
Operating voltage range 80V - 600V
MPPT voltage range 240-480V

But if we take the upper end of the MPPT voltage range (480V) and multiply by the max MPPT input current (18A x 2), we get 17.3 kW, well above the 9.3 kW upper end of the "recommended" PV Power. So the question is what the inverter will do if connected to 2 strings of panels with PTC ratings of say 440V and 16A each, for a PTC power of 14.1 kW, or a DC/AC ratio of 2.35. It clearly shouldn't blow up, but will it successfully make 6 kW during PTC conditions?

Cheers, Wayne

 

[ggunn]

Here's some numbers from the spec sheet of an inverter I selected semi-randomly, the Fronius Primo 6.0-1:

Recommended PV Power 4.8-9.3 kW
Max MPPT input current 2 x 18A (2 inputs)
Max input SCC 2 x 27A
Operating voltage range 80V - 600V
MPPT voltage range 240-480V

But if we take the upper end of the MPPT voltage range (480V) and multiply by the max MPPT input current (18A x 2), we get 17.3 kW, well above the 9.3 kW upper end of the "recommended" PV Power. So the question is what the inverter will do if connected to 2 strings of panels with PTC ratings of say 440V and 16A each, for a PTC power of 14.1 kW, or a DC/AC ratio of 2.35. It clearly shouldn't blow up, but will it successfully make 6 kW during PTC conditions?

Cheers, Wayne

The data sheets I use typically say something like this (from the SE80KUS 80kWAC data sheet):
Maximum DC Power (Module STC) Inverter / Synergy Unit 120000W / 60000W
Notice that it is 150% of 80kW / 40kW; this is typical.

I haven't investigated to see if it is possible to exceed 150% of 80kW while still conforming to their other constraints, but even if it is, I treat it as a hard stop and I will not do it.

 

[winnie]

Wow, thanks for the correction on max power vs max Voc * max Isc. I am surprised.

Now I'm curious about the internal topology of the system.

-Jonathan

 

[wwhitney]

Wow, thanks for the correction on max power vs max Voc * max Isc. I am surprised.

OK, so here's my question informed only by the I vs V graph for a string of PV cells: if the inverter is connected with an unreasonably high DC/AC ratio, like say 2.5, so that the published limit on Pmpp is violated, but the string Voc and Isc are both within spec, why can't the inverter just pick a region of the I vs V curve away from the maximum power point, where the power is within the inverter's spec? Either a lower voltage than Vmp, in the region of the curve where the current is roughly constant, or a lower current than Imp, in a region of the curve when the voltage is roughly constant, approaching Voc.

Maybe @tallgirl could explain.

Cheers, Wayne

 

[electrofelon]

OK, so here's my question informed only by the I vs V graph for a string of PV cells: if the inverter is connected with an unreasonably high DC/AC ratio, like say 2.5, so that the published limit on Pmpp is violated, but the string Voc and Isc are both within spec, why can't the inverter just pick a region of the I vs V curve away from the maximum power point, where the power is within the inverter's spec? Either a lower voltage than Vmp, in the region of the curve where the current is roughly constant, or a lower current than Imp, in a region of the curve when the voltage is roughly constant, approaching Voc.

Maybe @tallgirl could explain.

Cheers, Wayne

My understanding is that is what they do, but perhaps there is a limit to how far the inverter can move on the curve?

 

[electrofelon]

Here's some numbers from the spec sheet of an inverter I selected semi-randomly, the Fronius Primo 6.0-1:

Recommended PV Power 4.8-9.3 kW
Max MPPT input current 2 x 18A (2 inputs)
Max input SCC 2 x 27A
Operating voltage range 80V - 600V
MPPT voltage range 240-480V

But if we take the upper end of the MPPT voltage range (480V) and multiply by the max MPPT input current (18A x 2), we get 17.3 kW, well above the 9.3 kW upper end of the "recommended" PV Power. So the question is what the inverter will do if connected to 2 strings of panels with PTC ratings of say 440V and 16A each, for a PTC power of 14.1 kW, or a DC/AC ratio of 2.35. It clearly shouldn't blow up, but will it successfully make 6 kW during PTC conditions?

Cheers, Wayne

I am looking at my SMA sunny boy manual and for DC current it has the following:

Maximum input current per input - 10A
Maximum short circuit current per input - 18A

I would say the first is just the max the inverter can/will use, not that you must select a module with 10A or less of current. So the latter would be the only thing that provide a limitation. It isnt really clear to me what the purpose of declaring a maximum short circuit current is. Perhaps just to reduce further damage to the inverter if there is a DC fault that cannot be opened?

 

[ggunn]

Maximum short circuit current per input - 18A

For some (most?) inverters they mean 1.25 * Isc.

 

[pv_n00b]

Modified MICROinverters? I've never met a microinverter which didn't require a grid connection to un-island itself.

Yeah, that's what I was thinking when I saw the ad too. But they had an off grid microinverter, no idea who made it. I did not look to far into this kind of system as it was not anything I would be called to work with. I just found the advertisements and the description of operation interesting so I remember it. I'm not even sure they are in business today, a search for this type of product just returns a lot of solar water heaters.

 

[tallgirl]

My understanding is that is what they do, but perhaps there is a limit to how far the inverter can move on the curve?

OK, so here's my question informed only by the I vs V graph for a string of PV cells: if the inverter is connected with an unreasonably high DC/AC ratio, like say 2.5, so that the published limit on Pmpp is violated, but the string Voc and Isc are both within spec, why can't the inverter just pick a region of the I vs V curve away from the maximum power point, where the power is within the inverter's spec? Either a lower voltage than Vmp, in the region of the curve where the current is roughly constant, or a lower current than Imp, in a region of the curve when the voltage is roughly constant, approaching Voc.

Maybe @tallgirl could explain.

Cheers, Wayne

So, there is a limit to how much the curve can get moved, and it has a lot to do with the current carrying capabilities of some of the components, like the FETs that are used to switch the input from the panel on and off.

If the MPPT tracker tries to lower the voltage to reduce the power, the current can easily get too high. If it allows the voltage to rise well above Vmpp to reduce V*I, recalling that I decreases as V increases, something like a cloud edge could cause the current to rise faster than the tracker can react.

There also the fact that for the most part, Vmpp is 0.7 times Voc. With a ratio of 2.5, opening up the inputs, reading the voltage, then trying to get to 0.7 * Vmpp means that Impp is 2.5 times what it should be.

 

[tallgirl]

Yeah, that's what I was thinking when I saw the ad too. But they had an off grid microinverter, no idea who made it. I did not look to far into this kind of system as it was not anything I would be called to work with. I just found the advertisements and the description of operation interesting so I remember it. I'm not even sure they are in business today, a search for this type of product just returns a lot of solar water heaters.

Probably any micro could be made off-grid if the anti-islanding code was disabled. My question would be how would more than one of them synchronize in such a situation because anti-islanding is typically done by trying to change the output frequency and seeing how that works out.

 

[tallgirl]

Years ago an inverter rep confirmed to me that an internal register of input voltage in their inverters would tell them if the DC array ever exceeded the maximum voltage allowed, and that it would cause the warranty to be voided. I don't know if there is any way for the inverter to record if there had been a similar problem with DC system current (although if they were to send a rep to inspect the system, that is how they would know), but that is beside the point.

If you want to take on a situation where preservation of the warranty depends on the inverter company never finding out that you did something that they expressly prohibit, that is of course up to you, but I would never do it, especially if the system were for a customer instead of for my own use.

We know all sorts of things and have all sorts of ways to record what's going on.

One of the things I've worked on several times is calculating, and storing, the total on-time of a device. The math to make that possible gets a little weird because the memory used to store that can only be programmed and erased so many times. So, we decide how many years it has to work (I use values from 20 to 50 years), then decide how much memory I can waste on the counter. Once I know that, I can do some simple math and calculate how often I can store that number. So, if my flash memory has 100,000 program / erase cycles, and I'm spreading my counter over 4 chunks of memory, it's 400,000 into however many seconds are in 50 years, which is about 1.6 billion seconds, and 1.6 billion divided by 400,000 is around 4,000 seconds. Call it once an hour and the memory locations wear out a little faster.

If you've ever looked at a PCB and see a pad pattern with two little holes on opposite side of a block of 2 rows of 5 pins, that's a JTAG connection -- kinda like this: o :::: o

I can hook a device up to that and talk directly to the MCU and read out whatever data I might have left behind.