Jan, I could send you a calibrated 7805 in the mail, I can measure it to about 0,01% uncertainty at the moment.
So you bought 100 pieces and measured them all on a very good meter ?
Nice one.
No, it means I can measure the output of a random 7805 fairly accurately and can give you that number such as 5,1234V and if
you remeasure that same 7805 with your aneng meter and it reads for instance 5,120V we can be fairly certain your aneng is operating in spec.
But I believe your issues are already solved?
...the volt-nuts-disease is real.
Yes very real, I've started with two 2000 count dmm's that were a few counts apart (both well within spec) but my ocd
couldn't handle it so also got a 30.000 count bench dmm and a 10V vref, life was good for a while until a 20.000 count
bench dmm arrived, that was also just a few counts out compared to the other and the vref, but I couldn't be certain If either of them was 100% correct.

So a few years and several external calibration rounds later I have one(!) 2.000.000 count bench dmm and 9 vrefs and various thermal sensors/probes with data logging etc. etc...
Still use those same 2000 count meters tho, lol.