diy solar

diy solar

Off-grid 48V system plan seeking feedback and advice on grounding

MilbankAustralia

New Member
Joined
Jul 13, 2021
Messages
18
Hi all,

I'm hoping to get some feedback advice on my system plan. Want to thank Svetz who is a member here who provided very useful feedback on my first plan. This is v2 and hopefully it is nearly there!

I do have some questions about grounding/earthing the system though!
  • PhonoSolar PS270P-20/U 270W panels x6, Vmpp 31.44V, Isc 9.09A
  • SuperSafe SBS 190F AGM 190Ah (10hr rate) batteries x4
  • SmartSolar 150/35
  • Victron MultiPlus ii GX 48/3000/35-32
  • ZJBenny combiner box BHS-2/1 (15-20A string fuses, SPD & 2 pole disconnect)
  • Bussman LMI Input Module (M8 stud) + 3*LMI MIDI holders + 4 way bussbar with PEC/Prolec 58V MIDI fuses: 40/45A (MPPT), 125A (Inverter), 5A (DC-DC stepdown)
  • 200A MRBF fuse and terminal fuse holder (designed for M8 terminal), M6 stainless steel washer and nut for M6 terminal on our battery
  • 6mm2 XLPE for PV connections and combiner box to MPPT
  • 16mm2 class 5 stranded XLPE for MPPT to fuse block
  • 70mm2 XLPE for batteries series cables, battery to fuse block & fuse block to inverter/charger
  • Negative bus of some variety
The panels will be in 2 strings of 3 in series on a steel rack, going to the combiner box (2 in 1 out).

Combiner box to MPPT then MPPT to fuse block (40-45A MIDI on fuse block, close to battery)

Battery bank (4 in series) to fuse block via 200A marine rated battery fuse.

Fuse block to inverter/charger via 125A MIDI fuse.

I've worked out the current carrying capacities of the cables for the installation methods and think that and the fuses should be correct.

Voltage and temperature sensor will be connected to the MultiPlus II GX, and it will connect to the SmartSolar via VE.Direct to utilise distributed voltage and current control (which should share the temp and voltage info with the MPPT). I'll monitor using the MP2GX's wifi, the SmartSolar's BlueTooth and VictronConnect on my phone, or via the Mk3 VE.Bus to USB on PC (and for firmware updates).

My outstanding questions which I'd really appreciate advice on:
  1. Can I earth of the SPD in the combiner box on one of the PV frame earthing rods? The combiner box will be on one of the solar racks as it is.
  2. If the panels are touching, and mounted to a galvanised steel frame, does each panel frame need to be in contact with the earth conductor or can I earth from either the rack or just one of the panel frames?
  3. Can I use 6mm2 XLPE cable to connect the panel frames/racks to earth rods? Can I use the same for SPD to earth?
  4. Can I use the 70mm2 XLPE cable for the common earth for the MPPT and inverter chassis and the battery negative? Can these cables be joined before they get to the earth rod and if so what is a good way to do this without increasing resistance much? Can I attach them to a common stud and then run a single cable to the earth rod (<3m away)?
  5. Assuming correctly installed residual current circuit breakers on the AC circuits, is any further earth/grounding required on the AC side or is this all handled through the inverter/charger (i.e to the battery earth when not running from a generator and to the generator earth when running that way)?
  6. When the system is running on the generator it is earthed through the generator. The generator I am using is portable so I will need to connect it to an earth rod when using it to run the inverter/charger. Can I use the common earth rod the battery negative, MPPT and inverter chassis are connected to? Or does it need its own earth rod? And what type and size cable should I use to attach the generator chassis to this earth?
  7. Should I have a disconnect switch between the battery and the fuse block? I've had a hard time finding any that are rated for the 58V the batteries will be getting charged at so would appreciate suggestions if this is necessary.
  8. What does a shunt add? The low voltage cutoff will be on and I can monitor battery state through Victron Connect. I expect this system is oversized for our immediate needs and the loads to begin with will be pretty simple. I'm not sure it is worth the extra money but would be interested to hear opinions.
Most of my questions relate to grounding/earthing of the system. But, if anything is clearly wrong with the rest of the setup please do let me know!

Have learned a lot from the forums (I hope), and thank you all in advance for your valuable advice.

Cheers,

Simon
 
Last edited:
3kW inverter, 48V battery? The 125A fuse mentioned seems appropriate for continuous load. So long as inverter surge power above that is brief, like a second or so to start motors.
The 200A fuse(s) are between battery and busbar? Would support larger loads and take care of short circuits.

Grounding - We (U.S.) use a single point where neutral wire ties to ground wire. Usually at box where utility power goes to house. There is a ground rod there, maybe multiple rods or other earth ground configuration. Foundation steel and metal pipes might be tied together too. In a few cases (separate building), an additional rod is required.

For boat or RV, neutral and ground are connected somewhere in the utility electrical system feeding the power pedestal they plug into. When unplugged, a relay instead ties neutral to ground at the inverter. Your system is on a building, so shouldn't require switching of the ground bond.

(6) Generator provides earth connection? Does it bond neutral to ground (most don't)? Even if generator has a ground rod tied to ground wire, I would keep ground rod connection at cabin. But not multiple neutral/ground bonds. I would ground the generator through the cord connecting it to cabin.

(5) Residual current or ground-fault devices for AC just compare current through Line and Neutral, trip if different. That shouldn't be affected by where ground rod(s) are. Neutral/ground bond has to be on supply side of the device.

PV connections may use ground-fault. It can be as simple as a 1A fuse between PV negative and ground; if charge controller detects voltage across the fuse it stops drawing current. Follow whatever instructions in the manual.

Arc-fault is another thing for PV; bad series connections can cause a fire, especially with mixed brand MC connectors and off-brand products. Our code now requires AFCI for rooftop mount panels, but not for ground-mount.

(2) PV panel grounding - If PV system carries hazardous voltages, metal frames should be grounded. This could be lugs bolted to frames with star washers (to puncture aluminum oxide) so individual panels can be connected to ground wire or removed without interrupting ground wire running past all of them. Or, "WEEB" connections which bite into aluminum mounting rails so they serve as ground conductor. Steel wouldn't be used as the conductor.

(7) If switches/breakers can shut off current from all charge controllers and inverters, that should be sufficient for installing fuses or disconnecting battery cables. Like working on a car, good to break the circuit at one connection (typically negative cable) so a wrench shorting to ground doesn't complete the circuit.

Especially for people with lithium batteries which deliver 5x higher short-circuit current, a precharge resistor can be used to charge any capacitors prior to completing the circuit. Could be a resistor and loose wire, or could be a 2-position switch off-precharge-on.

Few switches rated for max charge voltage of a 48V battery as you noted. Probably even at rated voltage, can't interrupt full current without drawing a damaging arc. I think those switches are rated at lower voltages for reasons of safety requirements to prevent access to (slightly) higher voltages considered hazardous. If switched "cold" with no current flow I think they would function. I've seen some rotary AC switches which also had DC ratings, but for extremely low currents compared to AC rating, like one percent as high a current. But, some were rated for 100% current at DC so long as not switched with current flowing.

My inverters have breakers built in (and software controlled by pushbuttons to shut off and stop current flow), so I have no switches just fuses in the wires.

You can find DC rated breakers (with suitably high AIC to interrupt short circuit current from battery) easier than finding switches.
I note the following are "remote trip" which should mean it would be possible to have some safety/control device (e.g. lithium BMS) shut them off.




(8) Shunt serves to measure current. Your charge controller and inverter would each measure their own current. Depends on system and software, who keeps track of battery state of charge and charge current. Some systems determine low battery disconnect not just by voltage (which varies with current flow and temperature), but also by counting coloumbs.

I think your PV can charge battery at about 0.18C which may be fine. If you had more PV capable of delivering too much current, then a battery shunt plus a monitor Victron makes (don't know name or model) makes it possible to program an exact charge current such as 0.15C, and charge controller is told to deliver more or less as inverter consumes some of its current, while battery current remains constant.

For someone with lithium, a higher charge current like 0.5C may be fine. But only when reasonably close to 25 degrees C; if charged down toward the zero degree C minimum, a lower charge current would be necessary. Being able to program a particular (maximum) charge current plus a low-temperature charge disconnect in the BMS would be good.
 
Thanks for the reply Hedges.

First up: 48V just seemed like the most efficient way to go. If we expand the battery bank to 380Ah @ 48V we can add another set of panels and another charge controller to continue charging at 0.1-0.15C. It also meant I could use a single, lower current charge controller which we couldn't do if we'd set up a 380Ah 24V bank.

The 125A fuse is what the inverter/charger recommends.

The 200A fuse is on the battery terminal, designed to protect the 70mm2 cable between battery and fuse block, and the MRBF is chosen for its high interrupt rating.

6) I've attached the circuit diagram for the inverter/charger. When it is inverter mode (no AC input), the AC input relay is open and the earth relay is closed. Under this condition I believe the system is earthed via the inverter chassis?

And when it is in charger/feed through mode (AC input from generator), the system is earthed through the generator. The generator manual says the chassis of the generator can be earthed which I can only assume means the neutral is bonded to the chassis for grounding?

So are you suggesting that when the generator is being used, I bond the chassis to the common system earth rod (close to the battery negative and to which the battery negative, MPPT chassis and inverter chassis are attached), then disconnect it from the earth rod to move it when not in use? If so, what type and size of cable should be used to attach the ground rod clamp to the generator chassis? Any risk associated with touching the earth rod while the system is energised e.g if the generator gets turned off and the inverter turns on, can I safely disconnect the generator chassis from the earth rod?

PV connections may use ground-fault. It can be as simple as a 1A fuse between PV negative and ground; if charge controller detects voltage across the fuse it stops drawing current. Follow whatever instructions in the manual.
This is from the MPPT manual:

"Battery grounding
The solar charger can be installed in a positive or in a negative grounded system. Apply a single ground connection, preferably close to the battery, to prevent system issues or ground loops.

Chassis grounding
A separate ground path for the chassis ground is permitted because the chassis is isolated from the positive and the negative terminals.

PV array grounding
The positive and negative of the PV array should not be grounded. Ground the frame of the PV panels to reduce the impact of lightning. Do not connect the solar charger to a grounded PV array. Only one ground connection is allowed, and this should be near the battery.

Ground fault detection

The solar charger does not have internal ground fault protection. The USA National Electrical Code (NEC) requires the use of an external ground fault protection device (GFPD). The system electrical negative should be bonded through a GFPD to earth ground at one (and only one) location"

My interpretation of this is that the PV shouldn't be earthed (except the frames), just the common earth at the battery negative terminal. I'm a bit confused about the GFPD part here though. Is this just referring to the need to have a GFPD on each of the AC circuits or is it saying a GFPD needs to be installed somewhere on the DC side too?

Arc-fault is another thing for PV; bad series connections can cause a fire, especially with mixed brand MC connectors and off-brand products. Our code now requires AFCI for rooftop mount panels, but not for ground-mount.
These aren't required in Australia and there are only 1 or 2 very expensive options on the market here. The system is ground mounted.

(7) If switches/breakers can shut off current from all charge controllers and inverters, that should be sufficient for installing fuses or disconnecting battery cables. Like working on a car, good to break the circuit at one connection (typically negative cable) so a wrench shorting to ground doesn't complete the circuit.
I can disconnect the PV at the combiner box, and turn off the inverter/charger when needing to do something in the circuits. As such I'm not sure I need a breaker or switch to isolate the batteries? I should be able to disconnect the negative terminal from the battery safely with the PV disconnected and the inverter turned off? Appropriately rated breakers are pretty costly so if it isn't a necessary safety feature I would prefer to go without. Can I rely on the fuses to go if something major goes wrong, and just use the PV disconnect switch and turn off the inverter/charger for maintenance or circuit work?
(2) PV panel grounding - If PV system carries hazardous voltages, metal frames should be grounded. This could be lugs bolted to frames with star washers (to puncture aluminum oxide) so individual panels can be connected to ground wire or removed without interrupting ground wire running past all of them. Or, "WEEB" connections which bite into aluminum mounting rails so they serve as ground conductor. Steel wouldn't be used as the conductor.
Thanks. I will connect each panels frame with a suitable lug. The manual recommends 12AWG cable for grounding the frames so I presume I can use the 6mm2 XLPE cable I will be using for the PV circuitry too.

(8) Shunt serves to measure current. Your charge controller and inverter would each measure their own current. Depends on system and software, who keeps track of battery state of charge and charge current. Some systems determine low battery disconnect not just by voltage (which varies with current flow and temperature), but also by counting coloumbs.

I think your PV can charge battery at about 0.18C which may be fine. If you had more PV capable of delivering too much current, then a battery shunt plus a monitor Victron makes (don't know name or model) makes it possible to program an exact charge current such as 0.15C, and charge controller is told to deliver more or less as inverter consumes some of its current, while battery current remains constant.
I had forgotten that the shunt would allow the controller to automatically adjust its charge current to maintain at least 0.1C. But as you say, given the max possible rate is 0.18C I think once we have a fridge connected we may just be able to let it run at full amperage, or limit the controller to 30A initially and keep an eye on how much load we have plugged in to ensure we keep the charge rate at 0.1-0.15C. Perhaps if we find we are varying the load a lot during the daylight hours I will explore the shunt.

Grounding - We (U.S.) use a single point where neutral wire ties to ground wire. Usually at box where utility power goes to house. There is a ground rod there, maybe multiple rods or other earth ground configuration. Foundation steel and metal pipes might be tied together too. In a few cases (separate building), an additional rod is required.
So the only point in the entire system (DC and AC side) is the ground rod close to battery negative (our soil is clay that holds water well so we will only use a single grounding rod), and the MPPT chassis, inverter chassis and battery negative are all clamped to this rod? As well as the generator chassis when the inverter/charger is being supplied with generator AC input?

On that, what size and type of cable should be used for these connections to the earth rod please? And can some or all of these cables be combined before joining the earthing rod? E.g could I combine the cables from the MPPT chassis and inverter/charger chassis +/- the battery negative before they reach the earth rod? Perhaps at a panel mounted stud?

Can I use the 16mm2 class 5 stranded XLPE cable that connects MPPT to batteries for the MPPT chassis earth connections?
Can I use the 70mm2 XLPE cable that connects battery to fuse block and fuse block to inverter/charger for the battery and inverter/charger chassis earth connections?


I really appreciate your response Hedges, very useful!

I know I've asked a bunch more questions and hope they aren't too obvious! They still mostly relate to grounding (no secret I'm not an electrician then)!

Many thanks,

Simon
 

Attachments

  • Feed through mode earth relay circuit.png
    Feed through mode earth relay circuit.png
    47 KB · Views: 15
  • Inverter mode earth relay circuit.png
    Inverter mode earth relay circuit.png
    47 KB · Views: 13
I would have a ground rod at the inverter connected to ground wire and chassis regardless of generator.

"disconnect it from the earth rod to move it when not in use"
If you move the generator to store it out of the weather, then you would disconnect it. If stationary, the ground wires would just remain connected.

"which I can only assume means the neutral is bonded to the chassis for grounding"
Don't assume neutral bonds to ground in the generator, check with a meter. If isolated, single location bonding in your breaker panel or inverter. If generator does provide bonding, then if disconnected you should bond elsewhere. From your attachments, looks like your inverter isolated neutral from generator and grounds neutral internally, which is what you want. But now you have to make sure neutral & ground are bonded in the inverter, and if not then make that connection in the wiring between inverter and generator.

Here's a link with NEC size of ground wire. It is smaller than for normal current-carrying conductors because it only has to carry long enough to trip breaker.


Once current flow is turned off you can disconnect battery cable. When reconnecting you may get a spark - I just had one, didn't see before with just inverter (switched off) but this time I had charge controller connected (PV off, battery connection on.) Next time I have it in that state I should measure capacitance/resistance, maybe use a resistor to precharge.

PV GFCI, if not included in charge controller might be an external gizmo. That would be on the PV DC wires. Since SCC says not to ground either positive or negative wires, I would expect them to recommend what GFCI to use because the ones I know of do ground PV through a fuse.

"Disconnect PV at combiner" - do you have a switch or breaker? Neither fuses or MC cables should be disconnected with current flowing. Want to shut off charge controller somehow before disconnecting. I have DC switches either as an accessory on my GT inverters or separate. Preferred to turn off their AC breaker first. Once off I can disconnect wires and fuses. For an SCC I have breakers on PV and battery side.

Charge controller, you can overpanel to about 140% without any clipping if multiple PV arrays oriented differently. My estimate is the area presented to the sun, but second order effects probably change the number a bit. Assuming nothing casting shadows on the panels, more hours of good charging.
 
I would have a ground rod at the inverter connected to ground wire and chassis regardless of generator.
Yes inverter handles grounding by its chassis when the generator is not connected, and has a relay built in to handle switching between chassis grounding and passing the earth through to the generator

I will test whether the generator does bond earth to its chassis, thanks for the suggestion. I probably would have assumed since it has a labelled point on the chassis to earth that this was the case, and hopefully it is because I do not want to run the system with a floating generator when the inverter passes the earth through to the generator while receiving an AC input.
Here's a link with NEC size of ground wire. It is smaller than for normal current-carrying conductors because it only has to carry long enough to trip breaker.
Thank you for this.
Once current flow is turned off you can disconnect battery cable. When reconnecting you may get a spark - I just had one, didn't see before with just inverter (switched off) but this time I had charge controller connected (PV off, battery connection on.) Next time I have it in that state I should measure capacitance/resistance, maybe use a resistor to precharge.
What is the issue with the spark when connecting it?
PV GFCI, if not included in charge controller might be an external gizmo. That would be on the PV DC wires. Since SCC says not to ground either positive or negative wires, I would expect them to recommend what GFCI to use because the ones I know of do ground PV through a fuse.
Yes, I was confused but this. The charge controller specifically says not to ground the PV circuit. The way that it mentions the GFCI suggest to me that maybe they are mentioning it because it is in the NEC. They also specify that the system should only be earthed at one point, close to the battery.

"The USA National Electrical Code (NEC) requires the use of an external ground fault protection device (GFPD). The system electrical negative should be bonded through a GFPD to earth ground at one (and only one) location"
"Disconnect PV at combiner" - do you have a switch or breaker? Neither fuses or MC cables should be disconnected with current flowing. Want to shut off charge controller somehow before disconnecting. I have DC switches either as an accessory on my GT inverters or separate. Preferred to turn off their AC breaker first. Once off I can disconnect wires and fuses. For an SCC I have breakers on PV and battery side.
There is a two pole isolator switch in the combiner box that will allow me to isolate the PV. Should also be able to turn off the charge controller via the Victron app. I was not planning to install a breaker on the battery side, but rather to disconnect the PV and turn off the inverter prior to doing any maintenance or adjustment on the battery side wiring.

Charge controller, you can overpanel to about 140% without any clipping if multiple PV arrays oriented differently. My estimate is the area presented to the sun, but second order effects probably change the number a bit. Assuming nothing casting shadows on the panels, more hours of good charging.
Good to know. We were given six panels for free so this is what we are starting with. If we want to upgrade the battery bank we will need more PV capacity.

I really appreciate your input. I finally felt confident enough to start ordering components and are now waiting on the fuses, combiner box, charge controller and inverter. I think I now also understand how the RCBO's work as well, so I am hopeful I can wire up the rest of the system without too much difficulty now.

Can I earth the SPD in the combiner box on the same rod as the PV frames?

The storm last night made me think we may need to look into lightning rods as well.

Cheers
 
What is the issue with the spark when connecting it?

A spark could burn whatever is making connection (switch/breaker/threaded terminal), causing damage.
Capacitors being charged might be damaged by current surge. The product might have been engineered around lead/acid which appears to have 5x the series resistance, 1/5th the short-circuit current.
Fuses can be blown by this current. I did some math on this to see if a 200A fuse could see 1000A long enough (around 8 milliseconds) charging caps in an inverter to blow, and that seemed within reach.

Many commercial products have a device (maybe NTC thermistor?) to slow the initial precharge when connected to AC line. Front end of a switcher is often just diodes & capacitor, so without soft-start that would instantly drive capacitor to peak line voltage. You might notice a pop when plugging in some devices which don't.
 
Thanks for the heads up.

Would the risk be addressed by ensuring the inverter is switched off prior to ever connecting the battery?

Or do I need to explore the resistor option described in that thread?

I gather a single pole disconnect switch wouldn't do anything to address that concern.

Cheers
 
Switch off would contain the spark to inside the switch, if it is a physical disconnect switch. Maybe it is designed for it.
Concern could be switch was engineered/tested with the lower surge current of lead-acid, but lithium does more.

I've had a disagreement with another forum member using Sunny Island whether the disconnect isolates capacitors or not.

Obviously if FLA is involved, want the spark away from any gasses.

Some people just touch a resistor, or add a precharge pushbutton, to bypass a single disconnect.
If you're planning to connect it once and leave it connected, you could use something temporary.

I don't have anything yet, but I'm using AGM.
I should take a measurement, see what's going on with that spark I got. I guess that means clamp a CT around the wire and trigger scope. That's only rated for 100A. I can use my new clamp meter with inrush as well. But these spikes are faster than motor inrush.
 
Back
Top