Anyone else running mismatched panels in series? Curious about real-world losses

So I've ended up with a bit of a Frankenstein setup on the roof of my Kangoo — two 175W panels from a mate's old installation and one 200W panel I picked up off eBay. All three are nominally 12V (so around 36-38Vmp each), and I've wired them in series feeding into an EPEver 40A MPPT. On paper the string gives me roughly 110Voc which is well within the controller's 150V limit, but the Vmp and Isc figures are all slightly different across the three panels.

The EPEver seems to be tracking fine and I'm getting around 480-510W on a decent sunny morning here in the Peaks, which feels about right given the inevitable shading from my roofbar. What I can't work out is how much I'm actually losing to the mismatch versus what I'd lose anyway from the shading and the UK weather being what it is. I've read that the weakest panel drags the whole string down in terms of current, but the numbers I'm seeing don't feel dramatically off.

Has anyone actually measured this properly — maybe with a clamp meter or a decent logger — comparing a mismatched series string against a matched one? I've got a Victron SmartShunt on the battery side so I can see throughput reasonably well, but I can't isolate the panel-level stuff without more kit. Wondering if it's even worth worrying about or whether I should just crack on and enjoy the free electrons.

@KangooWanderer oh don't get me started on this one...

Running mismatched panels on the narrowboat for two seasons now — a pair of old 150W Renogy units alongside a shiny 200W mono I grabbed from a clearance sale. Put them all in series feeding a Victron MPPT 100/30.

Real world? The whole string gets dragged down to the worst performer. On overcast days that mismatch bites you hard — I've watched the Victron app show genuinely embarrassing output when one panel catches shade from the exhaust cowl.

My honest advice — put your two 175W panels in their own series string if you can swing a second controller, or at least use the Victron MPPT's built-in logging to actually measure the loss before assuming it's acceptable.

Some folks wave it away. I was one of them. I was wrong.

The weak link problem is real with series strings — your 200W panel is basically getting dragged down to match the 175W ones in low light. Worth checking if your controller is an MPPT because that at least softens the blow compared to PWM.

I ran two mismatched panels in series on my setup last summer and slapped a couple of bypass diodes across the lower-output ones — made a noticeable difference on partial shade days.

Honestly though, if you can rewire them in parallel instead, do it. Mismatched panels in parallel just lose a bit of current from the stronger one, rather than the whole string being hobbled. Depends what your controller can handle voltage-wise obviously.

What MPPT are you running? Victron SmartSolar handles mismatched strings better than most in my experience.

@OllieThompson73 is right about the series drag, but worth quantifying it properly. With mismatched panels in series, your MPPT controller finds a compromise Vmp across the string — neither panel operates at its true maximum. In my experience on the narrowboat, the real-world loss is typically 5–15% depending on how far apart the Isc values are between panels. Check the datasheets — if your 175W and 200W panels have similar short-circuit current ratings, you'll get away with it reasonably well.

The bigger killer is partial shading on any single panel propagating through the whole series string. If roof space allows, consider whether you could wire two panels in series as one string and the third as a separate string into a Victron SmartSolar — most of their dual-tracker MPPT units handle this cleanly and you'll recover meaningful wattage versus a compromised series string.

Has anyone actually measured the real-world difference by temporarily rewiring to parallel instead? I've got a similar Frankenstein situation at my cabin — three panels from different batches feeding into a Victron MPPT — and I've always wondered whether the theoretical losses translate directly to what the controller actually logs.

Specifically curious whether the Victron's adaptive algorithm recovers any of that lost headroom, or whether it's genuinely as brutal as the theory suggests. The datasheet numbers assume a dumb PWM setup, don't they?

Also — does mismatch loss compound badly on overcast days when you're already scraping for every watt? That's the scenario that matters most for EV charging from solar at my place.