Segmented DC Conductor Sizes for Voltage Drop

[pvgreeze]

Hello again everyone,

Thank you all for your help with my previous question regarding transformer & OCPD coordination for PV systems...the feedback was really appreciated!

I have run into another issue which I don't really have an experienced reference to ask. One of our DC installers asked me if it would be permissible to install what is basically a 'segmented DC conductor' to accommodate for voltage drop. To keep it simple, our plans for a particular string call for #8 the entire run (about 400 feet), but field conditions make #10 far easier to install between individual panels. The question he proposed to me is if he could string the panels together with #10, and splice that #10 circuit to a #8 wire at a junction box to complete the run back to the inverter.

Generally speaking, I figured it would be bad practice to use multiple conductor sizes in a single circuit, but other forums & articles I've read seem to indicate it is fine to actually do this if necessary. I made a small 'segmented voltage drop calculator' in excel, and found that about 90 feet of #10 spliced with 310 feet of #8 could keep the voltage drop under the target value of 1.5% voltage drop. Without worrying too much about the physical execution of this proposed solution, has anyone designed something like this before (multiple conductor sizes on a single circuit for voltage drop)? For clarity, the #10s would be used at the PV end of the circuit to string the panels together, and then the #8 would be used to carry that string to the inverter. At the end of the day, I'm just trying to avoid the ire of the inspector, and it seems that numbers wise that this solution should be perfectly acceptable.

Any thoughts or feedback would be appreciated.

Thanks!
-pvgreeze

 

[tom baker]

Using smaller size wire as the load drops is common in street lighting.

 

[retirede]

Changing conductor size is not uncommon and perfectly acceptable. Another example is when the wire size selected to minimize VD is too large for the terminals of the device at the end of the run. You splice to a wire size that the terminals can accommodate.

 

[jaggedben]

It's fine, just add some notes to your plans so the inspector doesn't think #8 is code required everywhere.

Btw there was a paper in Solar Pro magazine at one point that argued that avoiding DC voltage drop wasn't worth the money in wire cost. Might be even more true right now given current copper prices.

 

[ggunn]

Hello again everyone,

Thank you all for your help with my previous question regarding transformer & OCPD coordination for PV systems...the feedback was really appreciated!

I have run into another issue which I don't really have an experienced reference to ask. One of our DC installers asked me if it would be permissible to install what is basically a 'segmented DC conductor' to accommodate for voltage drop. To keep it simple, our plans for a particular string call for #8 the entire run (about 400 feet), but field conditions make #10 far easier to install between individual panels. The question he proposed to me is if he could string the panels together with #10, and splice that #10 circuit to a #8 wire at a junction box to complete the run back to the inverter.

Generally speaking, I figured it would be bad practice to use multiple conductor sizes in a single circuit, but other forums & articles I've read seem to indicate it is fine to actually do this if necessary. I made a small 'segmented voltage drop calculator' in excel, and found that about 90 feet of #10 spliced with 310 feet of #8 could keep the voltage drop under the target value of 1.5% voltage drop. Without worrying too much about the physical execution of this proposed solution, has anyone designed something like this before (multiple conductor sizes on a single circuit for voltage drop)? For clarity, the #10s would be used at the PV end of the circuit to string the panels together, and then the #8 would be used to carry that string to the inverter. At the end of the day, I'm just trying to avoid the ire of the inspector, and it seems that numbers wise that this solution should be perfectly acceptable.

Any thoughts or feedback would be appreciated.

Thanks!
-pvgreeze

It's really no big deal; you just consider each length of wire of a particular gauge in a series run independently, get the Vd in volts, add them up, and compare the sum to the voltage to get the %Vd. Keep in mind that you may need to upsize the EGC in some segments.

Also, as jaggedben suggested, look at the economics to see if sizing up the conductors for Vd is worth the cost of the wire.