Dual Multiplus II 48/5000 - Series Configuation - Brainstorming

DIY
, , , , ,
1

I currently run (most of) my house in off-grid mode on a single Multiplus II 48/5000. No AC-In connected. Other components: a Cerbo, a battery, some MPPT chargers and an AC-out coupled PC system. Very stable for 18 months now.

The 4000W limit sometimes is an issue though and requires more thinking/planning to avoid overload situations than I can expect from other people (partner/guests).

After doubling battery capacity (now 32kWh), that is bound to happen more often. More available energy, free from the sun, means more freely using it, including things like topping up an EV from the home battery if it wasn’t home during the day to directly charge.

Options considered:

  1. upgrade to a 48/8000 or 48/10000 - not ideal because of:
  • quite high hardware cost
  • increase in standby load, on those rare days with energy shortage, ~0.5kWh a day are substantial
  • leaving me with an unused 48/5000
  1. paralleling a second 48/5000 - not ideal because of:
  • complicated installation (ensuring AC and DC for both are wired with exactly the same wire length)
  • doubled standby load
  • system stops delivering power if either one of the parallel MP II fails -literally doubling the risk of failure
  1. Adding a second Multiplus II 48/5000 (call it MPB) to the AC-In of the existing one (call it MPA), to be run as a generator. Only started up on demand and stopped when demand drops.
  • almost zero additional standby load
  • easy installation, wire lengths don’t matter
  • backup Multiplus II right there, ready to be swapped in, for when the first Multiplus II fails. Maybe not for another 10 or 15 years, but eventually it will.

#3 is unconventional. From what I see rarely done, but seems the best option for my situation.

Similar has been done - Andy of Off Grid Garage fame has a Phoenix inverter connected to the AC-In of his Multiplus II. I would improve on what Andy has done though.

First of all I would use a second Multiplus II instead of a Phoenix. With charger disabled. It is cheaper to buy than a Phoenix. Plus the peace of mind of having a well maintained, occasionally run and tested, ‘hot standby backup unit’.

Then: Andy is using the in-built generator assistant. Not sure if he is using the one in the Multiplus or the one from the Cerbo. I ruled out both, I think my proposed solution is much neater, see below.

What I would copy from Andy is the main principle: the relay of the existing MPA is used to start the MPB.

The main logic would be contained in a series of assistants installed on MPA that update the “Ignore AC input” in sync with triggering the relay to start MPB. Not based on a single defined wattage (the main downside of the existing assitants), but based on actual load conditions:

  • if load exceeds 4000W, wait 5 minutes before starting MPB
  • if load exceeds 5000W, wait 30 seconds, then start MPB
  • if load exceeds 6000W, instant rush to get MPB up and running
  • if temperature pre-warning is active: rush to get MPB running
  • if load drops below 3000W for 30 seconds, turn off MPB
  • minimum 10 minutes ‘generator runtime’ to avoid frequent switching

All this can be implemented in MPA itself and does not rely on any network / CAN bus / other components. It would essentially be bulletproof.

If I want to go really fancy (and why not?) I could add some Node Red code to my Cerbo in addition. Dynamically adjust the input current limit of MPA to ensure any big load is shared evenly between MPA and MPB and that MPA never charges the battery using power supplied by MPB.

Reason for this post:

  • share the idea, in case it is useful for others in the community
  • get some other human brains to double check the logic behind it all
  • full disclosure: I spent a couple hours with an AI working on the best way to set up the assistants and to see if it comes up with any flaws: it gave the tick of approval
2

Just buy a single, larger inverter. It will be more efficient overall, losses exist beyond standby. I am going to dump my 2 x5’s.
They cannot stay balanced, they wear unevenly. If one fails it is a schlep to replace.
0.5kWh off a 32kWh of battery is not worth the additional PT considering end-to-end losses.

Getting overly creative with coding is a bandaid, not a solution.
Keep it simple.
Going off piste and unsupported has it’s own challenges.

3

Thanks for confirming that running them in parallel is not a good option. :slight_smile:

0.5kWh off a 32kWh of battery is not worth the additional PT considering end-to-end losses.

Would you mind spelling out what additional losses I should consider when running a second one purely as an “on demand additional generator”?

We have had days this winter when the solar panels were generating as little as 10% of their usual seasonal output, say 2kWh. 2 days in a row. 4kWh input in total. Using an extra 1kWh (over those 2 days), just to idle a bigger inverter makes a difference in that situation. It’s not the 99% of days, although even then, since the 0.5kWh is dissipated as heat, it makes a hot garage even hotter in summer.

4

Thanks for the extra context. If your PV is down to 2kW a day, then I can understand why losing 25% of that would be a big chunk. As a percentage of your battery capacity that is still only 1.5%, so it depends how long that scenario lasts. In that off grid scenario I would throw a bit more battery at it.

There are losses everywhere, the battery isn’t 100% efficient, nor is charging, inverting definitely not. When you total all this up then the system use is a drop in the ocean and really more relevant if the system is not serving any loads.

Two 5’s will use 36W with zero load, a 10k will use 38W.
Personally the minor difference isn’t worth the schlep. The benefit by my maths is closer to 0.4kWh, assuming the second inverter never turns on, the more it runs the smaller the perceived benefit.

For on-grid systems I find parallel systems push more power back to the grid than they ever self-consume. Off-grid doesn’t have that issue at least.