Anyone else finding their MPPT sizing calculations completely off once you account for UK cloud-edge effect?

Been running a 400W array (2x 200W Renogy panels in series) into a Victron SmartSolar 100/20 for about 18 months now. Did all the standard VOC calculations with a temperature derating factor, left myself what I thought was a sensible headroom — array VOC at -10°C works out to roughly 89V, well within the 100V hard limit. Job done, I thought.

Then last spring we got one of those classic broken-cloud days in late March — patchy cumulus, bright sun between gaps. The Victron Connect logs showed instantaneous input spikes hitting 97–98V repeatedly. That's the cloud-edge lensing effect, where light refracts around cloud edges and you briefly get irradiance well above 1000 W/m². Caught me completely off guard. Didn't blow the MPPT, but I was running uncomfortably close to the ceiling, and the controller was clearly throttling input during those spikes.

Standard guidance says add 25% headroom above your cold-VOC figure for this reason, which would have pushed me toward a 125V-rated controller for this array. Has anyone here actually sized their system with cloud-edge factored in from the start, or did most people find out the hard way like me? Curious whether this is more pronounced in certain parts of the UK — seems like it'd be worse in areas with more convective cloud activity.

@Ozzy yes, this caught me out too on my motorhome setup. Had a 360W array going into what I thought was a comfortably sized MPPT, then started seeing it clip on bright days with broken cloud cover — panel voltage spiking noticeably higher than my standard cold-temp VOC calc suggested.

From what I've read, the cloud-edge effect can temporarily push irradiance well above standard 1000W/m² test conditions — some sources say 1400W/m²+ briefly. That's not something the typical sizing calculators account for at all.

Did you actually see the Victron hitting its input limits or are you working from the app data? Curious whether the SmartSolar logs show it throttling. Mine definitely was before I spotted it — wasted capacity I wasn't even aware of until I started monitoring properly.

SilverTrekker | Posts: 847 | Location: Scottish Borders

@Ozzy @IanStevens - absolutely, the cloud-edge effect is the one that catches everyone out. What most guides gloss over is that you can see 20-30% above rated STC output during those brief moments when direct sunlight combines with light reflected off cloud edges. On a cold winter day that's also pushing your VOC up due to low cell temps - you're essentially hitting the controller with a perfect storm of high voltage and high current simultaneously.

Worth checking your Victron MPPT logs actually - the VictronConnect app shows historical peak input power and you might be surprised what it's recorded. I was seeing 480W regularly on my nominally 400W array.

The fix isn't always upsizing the controller - sometimes repositioning or partial shading management helps, but honestly just factoring in 125% of rated array power when sizing is my rule of thumb now.

RachelLamb | Posts: 312 | Location: Shropshire

@Ozzy @SilverTrekker this is such an underappreciated issue, especially in the UK where we practically live under variable cloud cover. Worth mentioning that the cloud-edge spikes can push your array voltage AND current simultaneously above STC ratings - it's not just a VOC concern. I've seen my 450W array briefly hit 520-530W on my monitoring during those classic broken cumulus days in spring.

My rule of thumb now is to add a genuine 25-30% headroom on top of your temperature-corrected VOC calculation, rather than the 10-15% most guides suggest. Also worth checking your MPPT's input overvoltage protection spec - some of the cheaper units don't handle transient spikes graciously even if the steady-state numbers look fine. The Victron units are generally pretty robust on this front at least.

WD40Wizard78 | Posts: 1,203 | Location: Array

@Ozzy worth noting the actual multiplier varies by panel technology. My shepherd's hut array uses half-cut monocrystalline — tested 118% of rated VOC during a particularly dramatic cloud edge event last March. Standard monos tend to sit around 110-115%. The Victron MPPT firmware handles the transient current spike reasonably well, but the voltage ceiling is what'll kill you if you're already running close to the 100V input limit on that 100/20.

Practical fix: I now calculate maximum series string voltage at -10°C plus an additional 15% cloud-edge buffer before selecting controller size. Pushed me up to a 150/35 on the van build. Costs more upfront but the alternative is a fried input stage, which costs considerably more.

Victron's own tech docs acknowledge the phenomenon but bury it in footnotes.