Perhaps someone with expert experience in meter bench-testing can explain a problem I had last year.
On the printed recommendation of a LF leading photographer, who had had two spot meters perfectly matched by a very reputable meter calibration service, I sent my Zone VI-modified Pentax analog spot and my much-less-recently calibrated Gossen Luna Pro-F to the same place, since the readings of the two varied widely and inconsistently. The meters returned, about 1 1/2 stops different.
I had another photographer confirm the reading discrepancy, just in case I had lost my mind.
I set up a carefully-made test target, metered it with both meters (both incident and reflected for the Gossen), recorded the readings in a table, and photographed the three set-ups (direct sun, open shade, interior indirect lighting). After notifying the service of the issue, I returned the meters with results, set-up photos, and further explanation.
When they returned again, the same problem persisted, suggesting that something in the meters' design differences made bench testing ineffective -- and that no off-bench test has been made. I resorted to using the Gossen's compensation feature to get them very close.
My question is why this evident failure in bench testing occurred.
I'd appreciate responses that offer some insight on this, if any is to be had, rather than questions on whether or not I had paid attention to possible sources of flare or infrared-emitting materials in my set-ups, or had used saturated colors that might be read differently by the meters. (None of the above. I'm not perfect, but I have run a lot of tests over the past five decades.)
Thanks.