Calibration is very easy and covered in the manual. Then it will agree with your best meter at 0V and 20V which is better than no calibration at all. Make sure you use the corresponding manual. AFIK the one in the Paul Rako article is for the version with the large horizontal meter (2005). There is another one for the unit with the thin vertical meter (2005A). From my experience let it warm up at least an hour before calibration. If you have not had it on much since you bought it I would let it burn in a couple of days straight before calibration. FWIW
Regards, robrenz
I think I have the same unit, the meter is the large horizontal one. I did a "burn in" after I received the unit and then I tested the stability by dialing in 10.000 V (9.993 + vernier) measured with a Rigol DM3061 6½ digit DMM. After several days the drift was about 30 µV, amazing! Now I haven't used the PSU in some time, but the dial and vernier is untouched since last time. So, I just powered it on again, connected to the same DMM and I will check on it in a few hours!
I use a cheep 230-115 VAC 50 watt transformer by the way. When I measure the internals of the 2005 and the output voltage, the 2005 is slightly underpowered (95-105 VAC depending on load), but the small wal-wart type 230-115 transformer seems not to get too hot and the 2005 works really well.
For some reason, all the neon bulb was burned so I have them replaced. Always have exotic spare parts!

(No, they where of course not an exact match. I needed lots of shrink tubing to make them fit in the old holders.)
Edit: When I think of it, I did a 0V calibration, but not the 20V one. Some rainy day... But I really wish I could come up with something to make it a really stable current source without external components/electronics. Maybe replace the current shunt resistor?