Faulty panel diagnosis from the ground

Hey everybody, bit quiet in this section so thought I would try a post!
I have a 1Kw array on the roof made of four 60 cell 260W panels in series.
Recently the peak power available has dropped from around 1100W to 750W, so I cleaned them as this is usually the reason for a drop in peak power, however that didn't fix the problem, just raised the peak to around 850W (yup the were dirty) , so in desperation I got out the DVM!
ISC was still a very healthy 9.3A (spec 9.1A) but VOC had dropped from the specified 150V to around 136V.
More interestingly the MPPT controller was running them around Vmpp 108V instead of the normal (and specified) 120V.

These things are inaccessible big time being on the roof but I suspect a panel fault as a faulty controller would not affect VOC (IMOP).
I am thinking along the lines of a shadow diode gone short circuit, there being 3 strings of 20 cells in each panel so 12 strings in all and one diode short on the array would reduce the output voltage by 1/12th.

I don't think its a corroded tabbing wire (1st thought) as I couldn't see any substantial corrosion through binoculars and my guess is such a fault would affect ISC.

Does anyone here have any experience of array faults, do my symptoms sound familiar ????

I do think you lost an array based on loosing 1/12 the voltage. A shorted diode would do it. Your weakest panel will dominate Isc/

I had to design and buy all of the hardware to do a "Utility demand-side management" study that was published. Another person installed everything and someone else did the data collection and analysis. At the time, there was heavy competition for my time by management.
We had a matrix-management scheme at the time. One manager was assigned to manage me as a resource and would also manage tie sheets and vacations. My time was then divided among other principal investigators.

The data aquisition system was connected with a technology called "data over Voice" (Internet didn't exist) and done in a different building because the house being studied was occupied. They had to be able to simulate a shower remotely and they investigated using heating elements to supplement hot water heating without an inverter. The Utility measured energy consumption. I know I had to measure insde/outside temp, humidity and solar irridiation. I'm pretty confident that the research resulted in the devices for the utility to be able to turn off the air conditioning and water heater remotely.

The wierd part was was there was a grid tie inverter that never worked or didn't work when the panels were changed, No one could figure it out. It did not take too long for me to figure it out. Although Voc was right, when the inverter was able to run, it shut off because the operating parameters were out of range. The manufacturer/installer reconfigured the arrays and all was well.