Mis-match V or remove panels?

[Zee]

Just added 12 panels to an old 34-panel system. New and old are different make and size panels. Also, old Poly vs new Mono.

Despite my calcs before i started job - after installation the actual measured V is quite different from expected.

I just measured 467VOC on both the original circuits. (2 @ 17 modules, 180W each) 6,120W
I also measured 438 VOC on the new circuit. (1 @ 12 modules, 310W each) 3,720W
old vs new string: 6.6% higher Voc TOTAL: 9,840W

They all parallel to an old transformer based SMA SB US inverter. Single MPPT.
Was expecting matching Voltage.
Actually i only focused on Vmp matching, not Voc, for efficient MPPT tracking and operation.
(unfortunately have not measured operating V yet, all I know is actual Voc is off more than expected)

Questions:
1. Is this V diff. workable? Is this a problem? How much efficiency loss?
2. Is it possible that the old modules will drop at a higher rate from Voc to Vmp, than the new modules? This could close the gap between V-operating of old and new a bit.
Using the values from their spec sheets, the ratio of Voc/Vmp does indicate factors of 1.26 and 1.21 respectively.
3. What would you do?
4. Perhaps remove one module each from the old strings of 17?
2 strings of 16 at 440 Voc would result. (16/17 x 467Voc)
Not happy about nixing two good panels.
Here, the question is: what produces more power? All panels intact but producing the disparate V above
vs
removing 2 modules and losing the 2 @ 180W.
Adding a 13th new module to the new string of 12 is not an option. No good room for it.

 

[GoldDigger]

A very crude rule of thumb tells us that if the Voc (and correspondingly Vmp) are within 5% on the two strings the power loss compared to the sum of the exact Vmp-Imp powers of the two strings is going to be 1% or less. The reason is that the output power from each string is changing very slowly with voltage in the vicinity of the MP point.
At a 6+% difference you will not be doing quite as well, but i think that you can be confident that the resulting total power will be greater than if you were to remove panels from the old string to make a better match.

If you do not trust this approximation, it would be easy enough to compare the total power input to the inverter as is and with two panels shorted out. There is no need to change the wiring to take them out of the string.

 

[Zee]

A very crude rule of thumb tells us that if the Voc (and correspondingly Vmp) are within 5% on the two strings the power loss compared to the sum of the exact Vmp-Imp powers of the two strings is going to be 1% or less. The reason is that the output power from each string is changing very slowly with voltage in the vicinity of the MP point.
At a 6+% difference you will not be doing quite as well, but i think that you can be confident that the resulting total power will be greater than if you were to remove panels from the old string to make a better match.

If you do not trust this approximation, it would be easy enough to compare the total power input to the inverter as is and with two panels shorted out. There is no need to change the wiring to take them out of the string.

thank you. Very useful.

I will also measure operating Voltage to see how much closer the V's get to each other once I turn it on.

 

[ggunn]

thank you. Very useful.

I will also measure operating Voltage to see how much closer the V's get to each other once I turn it on.

That won't tell you anything. Once the system is turned on all the string voltages will be the same no matter how badly mismatched they are.

 

[Zee]

That won't tell you anything. Once the system is turned on all the string voltages will be the same no matter how badly mismatched they are.

yes. excellent point. was unsure about that.

I was going to turn the strings on independently and measure.

 

[jaggedben]

Measuring current would tell you a bit more, both running the strings separately and together.

 

[BillK-AZ]

As manufacturers improve the solar cells, the fill factor (ratio of maximum power to the product of Voc x Isc) gets higher. This means that at operating voltages higher than Vmp the current (and power) drops off sharply. In this case the newer strings will likely show more of the mismatch loss than the older strings.