diy solar

diy solar

About to spend a lot of money tomorrow on this system (am I making a mistake?)

I'm only concerned now with the PV location in the sunny spot, and our cabin 500 feet away. I don't want to deal with 12 wires for 14 split phases in 8 dimensions for 500 feet of trenching to the cabin.

One DC wire carrying the PV to the cabin/inverter/MPPT/etc and then I'll hire an electrician to pick whatever gold plated, zinc infused, Auto Transforming Flux Capacitor he wants to make the GSHP and the cabin circuits work. Since it will be a small distance.

The PV is 22KW.
Its good to know the watts but I asked about the voltage.
I sense that you are over this for now.
 
Last edited:
The wife tells me we're at 5700 Watts of her in the dead of winter using an oven, two induction burners, water pump pumping water, drying a load of clothes, and having a rice cooker going all at the same time. Of course I'll be using the computer and watching TV and charging lithium batteries for power tools and getting further depressed on my lack of basic stuff I should have learned had I paid attention in school and that will take some power. Oh, and the GSHP will be taking another 2KW.

So I think 8KW will work, but it's close. And we hope to get electric ATVs in the future which we will need to charge, so that will break the electrical bank. And we may have water filters that need power in the rainwater collection system we haven't yet designed.

But instead of trenching AC wires for 500 feet after gazing into a snow covered crystal ball, I think I'll just get a fat DC wire that can take all of the 22KW of power to the cabin and deal with the rest later. I don't want to have to mess with that trench after I backfill it in the future.

Anyway, my monkey brain is full so I'll hand over this account to the wife since she's smarter than me and let her take it over from here. She's paying for all this stuff anyway and I don't want to be blamed!
 
The wife is waiting on an admin to approve her account so she can take over where I have failed. Although I did just watch a YT video on three phase electricity so I'm probably an expert now.
 
This is MrsAlaskanNoob (own account pending approval). I'm working on figuring out the "AC option" for running power 500 ft without going bankrupt in this thread: https://diysolarforum.com/threads/best-way-to-run-ac-power-500-ft-from-array-to-cabin.34376/

Re the DC option for doing so, in the other thread someone suggested that might be better, if we picked an "inverter with highest PV V input". He said "22kw @700v dc is only 35amps."

My question(s) about that approach are:
1. How should I figure out the max voltage that would be pushed from the solar array. VOC * # of panels? With temp corrected VOC, that would be 56.58 VOC x 50 panels = 2829 volts. Presumably, we'd have to organize in strings b/c I doubt there's an inverter (or a reasonably priced/sized/available one, anyway) that could take that high voltage.
2. Plus, it seems like charge controllers can't take anything close to that voltage, so again, the result is multiple strings. E.g., if charge controller can accept 450 volts, that's a string of 7 panels in series. For 49 panels, that's 7 strings. Each string requires two wires.
3. With multiple strings, running DC means running a large number of wires (e.g. 14 of them) 500 feet. So even if each wire is smaller, it ends up being very costly, plus cosmic to figure out how to address all of those wires being run together.

The only way the super high number of strings could be avoided is if there are charge controllers that take much higher voltage, no?

Edited to fix a typo (should have been 22kw, accidentally typed .22kw) and also to make clear it was someone else who suggested the "inverter" with highest PV V input. Presumably meaning an inverter/controller combo.
 
Last edited:
Re the DC option for doing so, in the other thread someone suggested that might be better, if we picked an inverter with highest PV V input. He said .22kw @700v dc is only 35amps.
Its the solar charge controller that connects to the pv array.
The higher the voltage and the lower the current the smaller the wires that can be used.
 
Its the solar charge controller that connects to the pv array.
The higher the voltage and the lower the current the smaller the wires that can be used.
MrsAlaskanNoob again. Understood re higher voltage/lower current = smaller wires. My question, in deciding whether we use AC or DC to run the 500 feet is this:

AC, if we step it up, we can do it with two wires, I think (single phase, live wire and neutral wire). If we are able to step it up high enough, this wire can be a reasonable size. Or so I think.

Folks have said DC would be better. But for DC, if we put all the equipment by the cabin, we have to run 500 feet between the panels and the charge controllers. The charge controllers' capacity limit us in terms of voltage, i.e., we can't do one massive string of all the panels - too many volts. So doesn't running DC mean running many wires 500 feet, even if they're smaller, compared to the two-wire AC option? My question is - does anyone know of a way to run just two wires for all of the DC power that our array can produce, i.e., is there any charge controller that can take that much voltage? (2829 volts, by my calculation ... or say half that, 1415 volts, given I think PV wire is rated at 2000 volts).
 
MrsAlaskanNoob again. Understood re higher voltage/lower current = smaller wires. My question, in deciding whether we use AC or DC to run the 500 feet is this:

AC, if we step it up, we can do it with two wires, I think (single phase, live wire and neutral wire). If we are able to step it up high enough, this wire can be a reasonable size. Or so I think.
This option would mean the batteries, inverter, generator and the balance of system would need to be in a heated structure at the site of the panels.
The inverter can only deliver 8000 watts in inverter mode and the generator can only do 7000 max so lets size for 8000 watts as a gesture to surge capacity.


For simple single phase 230VAC@60hz
1 un-grounded conductor(live)
1 grounded conductor(neutral)
1 grounding conductor(ground)
8000 ac watts / 230VAC = 34.782608696 service amps
34.782608696 service amps / .8 fuse headroom = 43.47826087 fault amps
The means the live and neutral have to be copper 1 awg or larger, so better than 2000 feet of 1 awg wire
The ground wire can be smaller but it still has to be larger enough to clear a ground fault. so lets say 2 awg. So 1000 feet of 2 awg.


For split phase 120/240VAC@60hz
2 un-grounded conductors(live)
1 grounded conductor(neutral)
1 grounding conductor(ground)
8000 ac watts / 240VAC = 33.333333333 service amps
33.333333333 service amps / .8 fuse headroom = 41.666666667 fault amps
The means the both lives and neutral have to be copper 1 awg or larger, so better than 3000 feet of 1 awg wire
The ground wire can be smaller but it still has to be larger enough to clear a ground fault. so lets say 2 awg. So 1000 feet of 2 awg.

I think it comes down to is 3000 feet of 1 awg worth more than a auto-transformer and the additional complexity and and another point of failure.
Folks have said DC would be better. But for DC, if we put all the equipment by the cabin, we have to run 500 feet between the panels and the charge controllers. The charge controllers' capacity limit us in terms of voltage, i.e., we can't do one massive string of all the panels - too many volts.
I suggest that you have 2 solar charge controllers for redundancy.
Have the voltage as high as possible which also means the current can be lower.
Lets take 24000 watts.
Each run requires 2 wires, 12000 watts@600 volts = 20 amps service amps
20 amps / .8 fuse headroom = 25 fault amps.
That requires a little over 2000 feet of 6 awg thhn in conduit in a trench.

So doesn't running DC mean running many wires 500 feet, even if they're smaller, compared to the two-wire AC option?
See previous.
The ac is not a 3 or 4 wire option.
My question is - does anyone know of a way to run just two wires for all of the DC power that our array can produce, i.e., is there any charge controller that can take that much voltage? (2829 volts, by my calculation ... or say half that, 1415 volts, given I think PV wire is rated at 2000 volts).
Not that I know of.

Direct bury aluminum is the cheapest option.
Much cheaper than copper even though you need larger conductors.
The direct bury part means no conduit which is another expense and a pita to install.
So you can't just pull another wire.
Once the trench is back-filled its a done deal.
I championed it a couple of times but people seem to not like aluminum although I've not heard any clear objections.
 
The inverter can only deliver 8000 watts in inverter mode and the generator can only do 7000 max so lets size for 8000 watts as a gesture to surge capacity.
For the "AC option," which I'm trying to sort out on another thread, I'm not assuming we're using the Victron equipment that AlaskanNoob picked out. In fact, adding up the power draw for all the stuff I would like to have running in a perfect world, I think it is undersized because I think our power draw is more like 15kw. So which inverter is an open question. But contemplating higher amps going through the wire, so needing to step it up above 240 to keep wire size reasonable. Trying to flesh that out on the other thread.

Thank you for the info re AC meaning 3 wires (4 if split phase) - I hadn't accounted for the ground wire (but glad it can be smaller!).

We will definitely have everything in a heated enclosure.
 
I championed it a couple of times but people seem to not like aluminum although I've not heard any clear objections.

Glad to have my account back...

I don't know much about aluminum wire other than we used to have a house with it and every house inspection that was noted as not the best for whatever reason. Some of the challenges we'll face is potential frost heave (I'll trench it ten feet down but it's possible the frost could get down there) and perhaps a tree root. I don't know if it can be a fire hazzard at ten feet underground, but if it can we'll go with whatever is safer since we live in a tinderbox.
 
Have the voltage as high as possible which also means the current can be lower.
Lets take 24000 watts.
Each run requires 2 wires, 12000 watts@600 volts = 20 amps service amps
20 amps / .8 fuse headroom = 25 fault amps.
That requires a little over 2000 feet of 6 awg thhn in conduit in a trench.
1. Thank you for this. To make sure I understand, it's 2000 feet because it's 1000 feet for each run, made up of 2 500 ft wires for each run?

2. This would mean we'd need two charge controllers that could accept 600 volts each, yes?

3. I'm confused how we get to 600 volts. Each of our panel's Voc is 46. If we're doing 1/2 the panels in series to make one of the runs, wouldn't that add up to 1150 volts (not accounting for temp correction, more with temp correction b/c cold), not 600? But amp rating on each panel is 10.23, so I would think that would mean each run at 1150 volts/10 amps(-ish), and I don't know how we get to each run at 600 volts/20 amps instead. (Is it possible to split the panels into 4 strings in series (approx 600 volts/10 amps) then tie two series strings in parallel (600 volts/20 amps) before the run to the charge controller?)

(Edited to fix a typo)
 
1. Thank you for this. To make sure I understand, it's 2000 feet because it's 1000 feet for each run, made up of 2 500 ft wires for each run?

2. This would mean we'd need two charge controllers that could accept 600 volts each, yes?

3. I'm confused how we get to 600 volts. Each of our panel's Voc is 46. If we're doing 1/2 the panels in series to make one of the runs, wouldn't that add up to 1150 volts (not accounting for temp correction, more with temp correction b/c cold), not 600? But amp rating on each panel is 10.23, so I would think that would mean each run at 1150 volts/10 amps(-ish), and I don't know how we get to each run at 600 volts/20 amps instead. (Is it possible to split the panels into 4 strings in series (approx 600 volts/10 amps) then tie two series strings in parallel (600 volts/20 amps) before the run to the charge controller?)

(Edited to fix a typo)
Here is a charge controller that takes 600 volts. Would have to see how many amps it can take in, and how much it can charge the batteries. But this one can only charge the batteries with 60 amp so it doesn't solve the problem cuz we'd have to have a bunch of them (unless you can wire solar charge controllers together and they all charge the batteries even though only some of them are getting PV plugged into them which is not how *I think* they work).

What we need is a charge controller that takes a ton of volts, at least 21 amps, but can also send a lot of amps to charge the batteries. Something like a 600/200 charge controller that can take in at least 20 amps.
 
1. Thank you for this. To make sure I understand, it's 2000 feet because it's 1000 feet for each run, made up of 2 500 ft wires for each run?
Yes.
2. This would mean we'd need two charge controllers that could accept 600 volts each, yes?
Yes.
3. I'm confused how we get to 600 volts. Each of our panel's Voc is 46. If we're doing 1/2 the panels in series to make one of the runs, wouldn't that add up to 1150 volts (not accounting for temp correction, more with temp correction b/c cold), not 600? But amp rating on each panel is 10.23, so I would think that would mean each run at 1150 volts/10 amps(-ish), and I don't know how we get to each run at 600 volts/20 amps instead. (Is it possible to split the panels into 4 strings in series (approx 600 volts/10 amps) then tie two series strings in parallel (600 volts/20 amps) before the run to the charge controller?)

(Edited to fix a typo)
The solar charge controllers, wire and tools we use are typically limited to 600 volts.
Beyond 600 volts is uncharted territory for me personally.
I've 0 experience with anything beyond 240 volts.
The actually voltage for the PV trunk is going to be something below 600 volts.
It will be achieved by seeing how many panels can be run in series and how many of those series strings can be combined in parallel.
There are plenty of folks here more experienced with pv than I am on this forum and I bet they will lend their expertise.
 
Here is a charge controller that takes 600 volts. Would have to see how many amps it can take in, and how much it can charge the batteries. But this one can only charge the batteries with 60 amp so it doesn't solve the problem cuz we'd have to have a bunch of them (unless you can wire solar charge controllers together and they all charge the batteries even though only some of them are getting PV plugged into them which is not how *I think* they work).

What we need is a charge controller that takes a ton of volts, at least 21 amps, but can also send a lot of amps to charge the batteries. Something like a 600/200 charge controller that can take in at least 20 amps.
Solar charge controllers can be paralleled on the system side but not the pv side.

Their are high end integrated solutions from these companies that probably deserve a look

There may be others.
 
Last edited:
There is some PV equipment for 1000V. That may only be grid-tie inverters for commercial or maybe European markets, not US residential.

AC can be stepped up to hundreds, even thousands of volts to transmit over the distance with smaller gauge wire. Just takes good enough insulation.

An alternative to situating batteries and inverters at the PV panel location is AC coupling. For instance:
3x Sunny Boy 7.7kW inverters located at PV array.
Step up to 960V center tap grounded. A couple 12kW, 480/240V transformers could do that.
After the long wire run, step 960V back down to 240V (I think leave center tap ungrounded).
4x Sunny Island 6048-US.

The catch with this setup is the Sunny Island battery inverter delivers max continuous charge current 110A, so four needed.
If you only had to charge at 13kW rate, two would have been sufficient.
MSRP $5700, typical retail $4700.

Typical wire is rated for 600V. Need to double check if that is OK for 480Vrms AC, or if peak voltage must be under 600V.
Haven't confirmed, but I see references to the insulation being good for 600Vrms.
 
Just for posterity I've been thinking of ways to get some redundancy for minimal extra cost.

instead of 4x 500 feet of wire for 2x pv trunks.
Use 3x 500 feet... 2x positive and an oversized common negative.
If one of the positives fails you can limp along with a single charge controller.
If the negative fails you can re-purpose on of the positives to become a negative and limp along on one controller.
Remember to leave appropriate service loops on the wires.
This gives N+1 redundancy.

Probably easier to just run 4 cables but its an idea.
 
Some devices want both PV wires isolated.
4 runs of wire makes that possible (and gives you N+2)

If we allow any significant IR drop in the long runs, then even if SCC don't officially care about common ground, one's exploration of PV I/V curve will show up as a perturbation of the other's input.
 
This is MrsAlaskanNoob (own account pending approval). I'm working on figuring out the "AC option" for running power 500 ft without going bankrupt in this thread: https://diysolarforum.com/threads/best-way-to-run-ac-power-500-ft-from-array-to-cabin.34376/

Re the DC option for doing so, in the other thread someone suggested that might be better, if we picked an "inverter with highest PV V input". He said "22kw @700v dc is only 35amps."

My question(s) about that approach are:
1. How should I figure out the max voltage that would be pushed from the solar array. VOC * # of panels? With temp corrected VOC, that would be 56.58 VOC x 50 panels = 2829 volts. Presumably, we'd have to organize in strings b/c I doubt there's an inverter (or a reasonably priced/sized/available one, anyway) that could take that high voltage.
2. Plus, it seems like charge controllers can't take anything close to that voltage, so again, the result is multiple strings. E.g., if charge controller can accept 450 volts, that's a string of 7 panels in series. For 49 panels, that's 7 strings. Each string requires two wires.
3. With multiple strings, running DC means running a large number of wires (e.g. 14 of them) 500 feet. So even if each wire is smaller, it ends up being very costly, plus cosmic to figure out how to address all of those wires being run together.

The only way the super high number of strings could be avoided is if there are charge controllers that take much higher voltage, no?

Edited to fix a typo (should have been 22kw, accidentally typed .22kw) and also to make clear it was someone else who suggested the "inverter" with highest PV V input. Presumably meaning an inverter/controller combo.
You folks have so many concurrent threads running that I can't keep track. At the risk of being spanked for talking about DC when the topic is AC, I'll answer this question. You would not connect all 50 panels in series. You would connect enough to suit the solar controller you buy (not the inverter--the inverter input voltage is whatever voltage you have decided your battery bank should be). There are a number of MPPT controllers that can handle in the neighborhood of 400-600 volts for the input voltage, and they output 48 or 24 or 12V. It's much harder to find MPPT controllers that can suit a high voltage (380V battery) but they do indeed exist. I'd shop for one in the 600V input range since that conveniently suits the voltage spec for common wire used to extend the leads that are attached to your panels, which are probably 14, 12, 0r 10AWG. I'd use 10AWG for your transmission line to the house since it comfortably suits the current spec for that distance for a single series string and it's readily available with insulation intended for outside use. The wire size you need is a function of the current carried by the line, and to a lesser degree the distance the wire runs. In theory, a series string of solar panels needs no heavier wire than the leads that connect the panels together and to the extension lines. In practice, the voltage drop from too small a wire size gets substantial over distance, so 10AWG is reasonable even if your panel wire is 14AWG. A higher voltage requires better insulation, so you need to pay attention to the rating. You will need one pair of wires for each series string. The actual loaded voltage will probably never be 58V, but let's say it is, then your series string would be 10 panels. So you'd have 5 parallel strings of 10 panels in series, so ten wires running from the solar array to the house. As mentioned previously, 10AWG extension wire is about $350 for a 500-foot reel, so your wire would be 10 X $350 = $3500. High input voltage controllers are more expensive than the everyday stuff, but the controller I'm buying from Sigeneer that I've ordered is $1000, though the price has apparently changed, I'm enjoying a dialog with someone in China who has not quite given me the price yet, but will next week--so I'm told.
SMG-B384-80A384V Output to battery850V Max PV voltage80A Max system output

Some higher spec MPPT controllers have multiple PV input connectors, but a simple combiner box will provide this function and convert your ten input lines to however many inputs your solar controller can handle. That's all the equipment you would need to accomplish this. You will need to understand how to manage HVDC, meaning fusing and breaker requirements for circuit protection, and managing inrush current through contactors used to switch the power. Most of this will be managed by whatever equipment you buy, but it's good stuff to understand, or a good reason to get pro help. Especially if you find an electrician used to industrial systems. A residential specialist might not know what they don't know.

Supplying power to the other locations which should have much lower loads would probably just be a matter of running a pair of lines from one of the series solar arrays with 10AWG cable to a smaller MPPT controller that can handle the 600V input, and charging a battery bank to power whatever you choose to supply. The BMS of the battery would open a change-enable contactor to shut off charging when the battery is fully charged and reclose when the battery drops below whatever you choose to restart charging. It might be cheaper and simpler to add a string of lower voltage PVs at your central location, or place some close to the location of the outbuildings.

I enjoyed hearing that high voltage DC is a non-starter since that is exactly what most commercial and industrial PV installations use. The only low voltage DC used in EVs and other power applications are homebuilt stuff. There's a reason all those products use HVDC. If you don't have to worry about residential codes then it's a fairly easy solution. Residential codes worry about the difficulty in protecting circuits with HVDC--a rational worry, an off-the-shelf breaker intended for AC will become a fireball if it's used to interrupt any significant current. I'm having to get some exceptions to install the system I'm working on for my shop, but I see daylight at the end of the regulatory tunnel. Expressing the conclusion that HVDC is a non-starter without a technical reason is entertaining if not illuminating. I would draw the poster's attention to the NW to California intertie, The HV DC line carries 3100 megawatts from the Celilo station in Oregon to Sylmar, Ca. and it's not there because of synchronization issues. And of course, every Tesla ever made.
 
Last edited:
Back
Top