Two string Tee Connectors?

[Neal30]

I have a question about these two string combining tees.
My concern is having to derate, for temp, conduit greater than 24", etc.
Also even though we are using snake tray, should I have to derate for elevated temps on the
roof?

260 watt module
Isc 9.08

9.08 x 1.25 = 11.35

11.35 x 2 strings in parallel = 22.7 Amps

#10 USE-2 in "Free Air" = 50 Amps 75 deg column

With 14 strings in the system, paralleled (7 strings). I'm looking at about 14 wires in the snake
tray.

Where do I need to derate?

310.15 (b)(3)

50% due to the wires in the snake tray being bundled together?

50 x .50 = 25 amps

The (#10 use-2) still looks good.

Should there be any temp derates for the rooftop?
Is there anything that I'm missing?

 

[jaggedben]

Here's some things you're missing: 690.8(B)(2) and 310.15(B)(2)(a).

Either
1) Multiply your 22.7A by another 125%
or
2) Calculate the full conditions of use, including the temperature deration in 310.15(B)(2)(a)

"Snake tray" is not an NEC term. So I'm not clear if the the 50% deration for 14 conductors is required or not. If it's a cable tray or a metal wireway, I believe those are not required. If it's considered a raceway, then the requirement apply. I'm guessing 310.15(B)(3)(a) does not apply.

All citations are 2011 code. For the 2008 code just go with 156% of Isc.

 

[tallgirl]

Is there anything that I'm missing?

Yes, module-level mismatch.

What's the inverter or charge controller?

 

[BillK-AZ]

Question for Neal30:

How are you providing overcurrent protection with 14 strings, wired in pairs, in accordance with NEC 690.9 and the maximum allowable fuse per the UL label on the 260W PV module?

Most 260-watt PV modules have a 15-amp fuse limit, some have a 20-amp limit. Even a 20A fuse is not adequate for two strings in parallel.

 

[Neal30]

Here's some things you're missing: 690.8(B)(2) and 310.15(B)(2)(a).

Either
1) Multiply your 22.7A by another 125%
or
2) Calculate the full conditions of use, including the temperature deration in 310.15(B)(2)(a)

"Snake tray" is not an NEC term. So I'm not clear if the the 50% deration for 14 conductors is required or not. If it's a cable tray or a metal wireway, I believe those are not required. If it's considered a raceway, then the requirement apply. I'm guessing 310.15(B)(3)(a) does not apply.

All citations are 2011 code. For the 2008 code just go with 156% of Isc.

Ok, I see what you are looking at. We are actually using a slightly smaller module, I listed the
size below.

245W
Isc 8.49

8.49 x 1.25 x 1.25 = 13.27 A

13.27 x 2 = 26.54 Amps

USE-2 "free air" 90 deg. Column = 55 amps

310.15(B)(2)(a)

Ambient temp from solar abc's is 49 deg Celsius for wire 3.5" - 12" inches above roof. Near
Johnson City, Tennessee.

49 degrees = .82 correction factor

55 x .82 = 45.1 amps

If there are no other correction factors, then this looks fine either way using the #10, I just had concerns about not being able to use conduit for transitions between walkways. It also is strange to me that I can double the current I normally put on a #10 wire, without having to upsize the wire.

Thanks for the references in the code.

 

[Neal30]

Yes, module-level mismatch.

What's the inverter or charge controller?

We are using a 50 KW grid tied inverter. They do use all the derates when calculating the annualAC output for the customer. It usually works out to be 75% conservatively. We will just have
one combiner box, and one input into the inverter.

 

[Neal30]

Question for Neal30:

How are you providing overcurrent protection with 14 strings, wired in pairs, in accordance with NEC 690.9 and the maximum allowable fuse per the UL label on the 260W PV module?

Most 260-watt PV modules have a 15-amp fuse limit, some have a 20-amp limit. Even a 20A fuse is not adequate for two strings in parallel.

This is a interesting question. These modules are actually a slightly smaller, they are 245W

Isc 8.49

8.49 x 1.25 x1.25 = 13.27 amps

13.27 x 2 = 26.54 Amps

The electrical drawing is going to have these paralleled strings on 30 amp fuses, in an 8 string
combiner box.

Now since we are not going over the conductor ampacity, one would think it would be suitable.

But what your saying is the OCPD for that combined string is now a 30 amp fuse, which means
the modules are now protected by a fuse greater than 15 amps. I already know what the
electrician is going to say, something along the line of Kirchoffs law, and so forth.

What we are seeing a lot of in these megawatt designs, is the use of these string couplers.
Would there have to be additional fuses either in the coupler itself, or before we parallel these
wires?

690.9(A)(b) This is where it gets us in the highlighted text. So we would need additional fuses
before we combine these strings.

This makes a lot of sense. One of the worst case scenario is to have some type of short, that
would damage the modules and have voided there warranty.
What type of electrical or fire hazard could stem from this type of design error? Regardless it is
not going to be acceptable, for this installation.

Thanks for the input, always appreciated!

 

[GoldDigger]

What type of electrical or fire hazard could stem from this type of design error? Regardless it is
not going to be acceptable, for this installation.

Thanks for the input, always appreciated!

The primary risk is fire from excessive heat dissipation inside a defective or damaged panel caused by reverse current supplied by the other strings in the array.
If you have only two strings, the maximum reverse current that could be supplied to a panel fault will be the output current of one other string. That is within the design tolerance of the panel.
But once you parallel three or more strings, the current produced by them can exceed the fault current tolerance of the panel(s) in the damaged string. To prevent this, the manufacturer specifies the maximum allowed series fuse size.
It is not intended to block excess current produced by that string, since this is limited by the available solar energy. It is intended to protect that string from damage by current from the other panels.
In your case a 30 amp fuse after the T would limit the current to far less than the output of all of the remaining strings, but it could still be too large for the panel without blowing the fuse.

 

[tallgirl]

We are using a 50 KW grid tied inverter. They do use all the derates when calculating the annualAC output for the customer. It usually works out to be 75% conservatively. We will just have
one combiner box, and one input into the inverter.

As long as they understand they are wasting money, that's fine. I mean, there are Code issues, but paralleling modules means the module level mismatches are worse, not better.

 

[tallgirl]

Ok, I see what you are looking at. We are actually using a slightly smaller module, I listed the
size below.

245W
Isc 8.49

8.49 x 1.25 x 1.25 = 13.27 A

13.27 x 2 = 26.54 Amps

USE-2 "free air" 90 deg. Column = 55 amps

310.15(B)(2)(a)

Ambient temp from solar abc's is 49 deg Celsius for wire 3.5" - 12" inches above roof. Near
Johnson City, Tennessee.

49 degrees = .82 correction factor

55 x .82 = 45.1 amps

If there are no other correction factors, then this looks fine either way using the #10, I just had concerns about not being able to use conduit for transitions between walkways. It also is strange to me that I can double the current I normally put on a #10 wire, without having to upsize the wire.

Thanks for the references in the code.

Unless I'm missing something, you can't make this work.

As others have pointed out, the maximum allowed fuse is 15 amps. Depending on fill factor, Impp for a module having Isc of 8.49 is likely above 7.5 amps. You can't produce the power from the module without blowing the fuse.

I've looked at these Y connectors before and I don't see how they can ever be used to parallel strings made by them. To my way of thinking, if the maximum fuse size is 15 amps, and the modules have Isc of 8.49, the largest fuse that could be used in a combiner box would be 15 - 8.49 = 6.51. The paralleled string would provide up to 8.49 amps (nominally) and the current from other strings through the combiner would then be limited to 6.51.

 

[Neal30]

Unless I'm missing something, you can't make this work.

As others have pointed out, the maximum allowed fuse is 15 amps. Depending on fill factor, Impp for a module having Isc of 8.49 is likely above 7.5 amps. You can't produce the power from the module without blowing the fuse.

I've looked at these Y connectors before and I don't see how they can ever be used to parallel strings made by them. To my way of thinking, if the maximum fuse size is 15 amps, and the modules have Isc of 8.49, the largest fuse that could be used in a combiner box would be 15 - 8.49 = 6.51. The paralleled string would provide up to 8.49 amps (nominally) and the current from other strings through the combiner would then be limited to 6.51.

I am on board here, I am looking at this objectively. These Y connectors are new to me also, and I like to have a real good grasp on something before I start installing them.

What your saying is that since the modules have a 15 amp max fuse rating, that the strings
would only be able to produce 15 amps total on that paralleled string? So it would drag down the other string to keep it in under that 15 amp range?
But at the combiner box I'm paralleling 8-25 strings at times and their currents are additive and equal, is it not the same at the Y connector?

I really appreciate this great feedback, and helping me grasp this type of string configuration.

 

[tallgirl]

I am on board here, I am looking at this objectively. These Y connectors are new to me also, and I like to have a real good grasp on something before I start installing them.

What your saying is that since the modules have a 15 amp max fuse rating, that the strings
would only be able to produce 15 amps total on that paralleled string? So it would drag down the other string to keep it in under that 15 amp range?
But at the combiner box I'm paralleling 8-25 strings at times and their currents are additive and equal, is it not the same at the Y connector?

I really appreciate this great feedback, and helping me grasp this type of string configuration.

There are a number of components within the modules which have current limitations. They include both the bypass diodes and the ribbon conductors. So, the array has to be constructed so you don't exceed that current limitation, and that's the 15 amp maximum fuse limit printed on the label on the back.

Now, let's assume for a moment that you have a short somewhere in the module. Kirchoff says that current flows through all possible paths. In this case that would include the 8.49 amps from the =other= string's Isc. That's unfused because there is no fuse in the Y connector. Unless I'm missing something, the maximum current that can be allowed from any other string is the difference between 15 amps for the fuse rating and 8.49 amps, or 6.51 amps. That would be the maximum fuse rating at the combiner box for the combined string. 8.49 * 2 = 16.98 amps, which is more than 6.51 amps, so that ain't going to work at all.

Normally when you have two full strings in parallel, the maximum current the "other" string can produce is Isc. In your case, the maximum that can be sourced by the "other" string is the fuse rating, plus Isc of the string that's in parallel using the Y connectors.

If you're concern is with wire runs and getting into pipe as fast as you can, there is nothing that requires you to bring all of the conductors back to a single (or more -- depending on how many strings you have) combiner box. You can aggregate strings closer to the arrays (make sure you satisfy 690.11), then transition into pipe and combine the outputs of those combiner boxes at one larger box.

 

[jaggedben]

I am on board here, I am looking at this objectively. These Y connectors are new to me also, and I like to have a real good grasp on something before I start installing them.

I should probably have mentioned in my first reply... I've never seen these Y connectors used on a project. I've seen them at shows, but wondered who would use them and why.

What your saying is that since the modules have a 15 amp max fuse rating, that the strings
would only be able to produce 15 amps total on that paralleled string? So it would drag down the other string to keep it in under that 15 amp range?

No, they are saying that two strings in parallel would blow the 15 amp fuse, but a larger fuse would violate the NEC.

But at the combiner box I'm paralleling 8-25 strings at times and their currents are additive and equal, is it not the same at the Y connector?

The difference, presumably, is that your combiner has fuse holders for each connection, but your Y-connectors do not. The Y-connector does not allow you to properly size fuses according to the NEC.

 

[tallgirl]

I should probably have mentioned in my first reply... I've never seen these Y connectors used on a project. I've seen them at shows, but wondered who would use them and why.

I've only ever seen them advertised. I don't know anyone who's used them.

They increase the impacts of module-level mismatch. Instead of allowing the entire string to even itself out, the worst module of the pair can impact the better module.

And not to tout microinverters, but when I started looking at the power output difference between two completely identical modules, I pretty much swore off all manner of string-based power conversion, except for battery backed systems where its too complicated to AC couple a small system.

I have five modules left at the shop I need to get set up and I'll have a better idea of just how bad perfectly matched modules can be once I get all the sensors hooked up. The 240 watt panels I have right now, with the same hardware platform, are making between 205 and 228 watts AC. That's within the typical +/- 5% range for panel output, but still -- that's a lot of variation that can be avoided, and Y connectors will make it worse.

 

[GoldDigger]

They increase the impacts of module-level mismatch. Instead of allowing the entire string to even itself out, the worst module of the pair can impact the better module.

I do not see what kind of mismatch you are worried about.
If you parallel two modules (or equally likely with these connectors, two strings), you are only assuming that the Vmp of the two panels or strings will be close enough that the power loss is insignificant. A rule of thumb is that if the Vmps of the two parallel elements match within 5% the power loss will be negligible (less than 2.5%) when feeding a single MPPT input.
And although the Imp of two elements may be different because of shading, panel tolerances, etc, there will not be a noticeable difference in Vmp.

What difference would you expect to see in the performance (other than providing compliant fusing) when paralleling the two elements at a combiner versus paralleling them with a connector Y?

 

[tallgirl]

I do not see what kind of mismatch you are worried about.
If you parallel two modules (or equally likely with these connectors, two strings), you are only assuming that the Vmp of the two panels or strings will be close enough that the power loss is insignificant. A rule of thumb is that if the Vmps of the two parallel elements match within 5% the power loss will be negligible (less than 2.5%) when feeding a single MPPT input.
And although the Imp of two elements may be different because of shading, panel tolerances, etc, there will not be a noticeable difference in Vmp.

What difference would you expect to see in the performance (other than providing compliant fusing) when paralleling the two elements at a combiner versus paralleling them with a connector Y?

When a diode goes into bypass, Vmpp for that module goes down by about 1/3rd, not just 2 to 5%, as with mismatch. Because of the shape of the I-V curve, that's not trivial -- V for the pair may be high enough that the entire module is lost unless the MPPT tracker goes looking. There are a number of other scenarios by which two modules that are in parallel can cause each other no end of grief -- hot spots and bird presents being two common ones.

The biggest issue is that modules in series are limited by the voltage of the lowest performing module. If the voltage is held above Vmpp for the "bad" module, it under-produces. If the voltage is held below Vmpp for the "good" module, it under-produces. When you have strings that are just in series, that difference evens out over the length of the string.

It's just a bad idea.

 

[Neal30]

Spectacular feedback!

Spectacular feedback!

I just want to say thank you for all of this great feedback and stimulating discussion. I'm sure
this will allow our team to look at these with a more informed perspective. We face new
challenges everyday, and I am so thankful for a forum of informed individuals explaining some
the complex concepts that we face in this industry. Our goal is to design an safe, efficient, and
cost effective system that will withstand the test of time.

Thanks again and I will share some of the discussions we have in the design pool, on this
subject.

 

[ggunn]

You are going to have to fuse each string individually. If you combine strings in pairs and fuse the pairs at 30A, then there will be nearly 40A of fault current available on a string - 30A through the fuse from the rest of the array plus the Isc of the adjacent string.

 

[jaggedben]

[Y-connectors] increase the impacts of module-level mismatch. ...

What difference would you expect to see in the performance (other than providing compliant fusing) when paralleling the two elements at a combiner versus paralleling them with a connector Y?

tallgirl, I'm with ggunn here. I don't understand any mismatch difference between the y-connectors and a combiner.

I get that you're working for a micro-inverter manufacturer and want to tout the virtues of micros, but of course the typical consideration is that the lower $/watt for string and central inverters offsets the production losses due to mismatch between strings. That is, of course, unless the site has shading that drastically increases the production losses.

 

[tallgirl]

tallgirl, I'm with ggunn here. I don't understand any mismatch difference between the y-connectors and a combiner.

I get that you're working for a micro-inverter manufacturer and want to tout the virtues of micros, but of course the typical consideration is that the lower $/watt for string and central inverters offsets the production losses due to mismatch between strings. That is, of course, unless the site has shading that drastically increases the production losses.

I'm not convinced that with lower labor costs for installation and longer product lives that string inverters are cost-competitive. Perhaps the up-front costs are lower for string inverters, but not the long-term costs.

I believe the Y's are to be installed at each module, not at the very end of the string. If they are at each module, the performance of each module pair is controlled by the performance of the worst module in the pair. If they are at the end of the string, then the performance is the same as if they are going into a combiner box.