Power loss and efficiency as a function of load (AC to DC / DC to AC) for MultiPlus-II 48/4K5 or 6K5

by George · 1 month ago 60 views 4 replies
George
George
Active Member
12 posts
thumb_up 8 likes
Joined Apr 2024
1 month ago
#5414

Been running a Victron MultiPlus-II 48/5000 in my shepherd's hut setup for a while now and the efficiency question at lower loads has genuinely been bugging me lately.

Most of the published specs give you peak efficiency figures which look lovely on paper, but real-world usage is a different story. When you're only pulling 200-300W to run a few lights and charge a laptop, you're nowhere near those headline numbers — and the inverter's own idle consumption starts to matter quite a lot.

Specifically curious whether anyone's seen proper efficiency curves or tables for the newer 4k5 and 6k5 variants — not just the peak figures. Something showing:

  • AC to DC (charging mode)
  • DC to AC (inverting mode)
  • Crucially, at low loads (sub-500W range)

I've been logging mine with a Cerbo GX and the numbers at light loads look... less impressive than I'd hoped. Not terrible, but not what the marketing suggests either.

This matters a lot if you're sizing a system where the inverter sits lightly loaded most of the time — which, let's be honest, is most of us. Nobody's running a full 4.5kW continuously are they 😄

Victron's documentation is usually solid but I haven't found a detailed efficiency diagram for these specific newer models anywhere. Their older units had better published data it seems.

Anyone done proper testing, or know if Victron have published anything more detailed? Would love to see some real numbers rather than guessing. Especially relevant if you're pairing with Fogstar Drift cells where every wasted watt feels personal 😅

Moor Lee
Moor Lee
Active Member
28 posts
thumb_up 27 likes
Joined Jul 2023
1 month ago
#5457

@George1975 the efficiency curve on these things is basically a cruel joke at low loads — you're essentially paying a "just existing" tax to the inverter 😄

Victron's own datasheet shows the 48/5000 sitting around 94-96% efficiency at higher loads but it nosedives when you're just powering a lamp and feeling

DuctTapeDave
DuctTapeDave
Member
7 posts
thumb_up 7 likes
Joined Aug 2024
1 month ago
#5471

@George1975 the idle draw on the 48/5000 is around 15-20W just sitting there looking pretty — I've got a Fogstar 48V bank and a clamp meter that tells no lies, so I run mine through a scheduled relay to cut it overnight rather than let it sip watts like it owns the place.

NotAnElectrician80
NotAnElectrician80
Active Member
24 posts
thumb_up 27 likes
Joined Jul 2023
1 month ago
#5490

@George1975 ran my static van on a MultiPlus-II 48/3000 for two winters and the efficiency curve below 20% load is genuinely tragic — like watching someone use a sledgehammer to crack a walnut, except you're paying for the sledgehammer's electricity bill.

Enable the "dynamic cutoff" settings in VEConfigure and set a proper search mode schedule if your loads are predictable — got mine dropping to ~6W search mode overnight and it made a noticeable dent in my Fogstar battery drain.

The 6K5 is even worse at light loads proportionally, so unless you're actually pulling meaningful wattage regularly, the 48/5000 you've already got is the lesser evil.

Tina
Tina
Member
5 posts
thumb_up 3 likes
Joined Nov 2024
1 month ago
#5511

@George1975 this is exactly why I went with a transfer switch + smaller inverter approach for my garden office — keeping the MultiPlus in standby for heavy loads and emergency backup, but routing low-draw circuits through a dedicated DC-AC unit sized more appropriately.

The published efficiency figures (Victron quote ~94% peak on the 48/5000) are essentially best-case. Real-world at 5-15% load you're looking at genuinely grim numbers.

Worth pulling the actual VRM data if you're on a Cerbo — you can plot input vs output over time and see exactly where your losses are sitting. Mine was eye-opening when I first ran it for the office setup.

Log in to join the discussion.

Log In to Reply