If it reads 0 with distilled water, it still may not be calibrated. For it to work properly it needs to be calibrated to the SG of the water you will eventually be using it to test, i.e. 1.025ish.
I calibrated mine using distilled, as per the instructions. I then found out when it read 1.025 it was actually 1.028
IIRC, about 1/3 to half of the people who tested theirs found them to be off, usually reading about .003 lower than reality. So, by saying you've never had a problem with yours, you may actually be saying you've never had a problem running a higher SG.
Anyway, I guess stability is most important, and the refreactometer still works in that sense.
It was just very disappointing to hear about this inaccuracy, especially after reading hundreds and hundreds of people tout the accuracy of refractometers over swing-arms and float style. Turns out that it just may not be so.
Few of the refractometers out there are designed for NSW. They are designed for Sodium Chloride. A high quality refract will be wrong by .0015 merely because NSW has a lot of ions that aren't in pure NaCl.
I've been adjusting to zero with distilled and mentally adjusting for the .0015 for years. This is not a new issue. The new issue is the swarm of el-cheapo Chinese refracts that are off by more than .0015.
You can make your own standard. Randy Holmes-Farley wrote an article about how to a while back. It's easiest to use 53 micro Siemens fluid for pinpoint salinity monitors.