Using a Prema 5017

[Resonant]

I have a question about the use of the Prema 5017. I purchased it about 3 years ago and it came from the estate of a deceased voltnut and it was in good cosmetic condition. It has been calibrated by Prema twice recently last time April 19.
When turned on and allowed to warm up for a couple hours it seems accurate when checked against a calibrated source. However, if it is left on in 7.5 digit mode on 30V range using a 10V standard for more than a day it starts to drift a lot, around 0.000171V. Using my own 34401a and borrowed Keithley 2002 they shows almost no drift over the same period against the same Standard.
I have checked the internal reference LM399 over the same period and that also seems stable. The big difference seems to be the Prema shows a large offset, if you short the leads.
My previous experience has been using US made meters like 34401a and Keithley which don’t develop much of an offset and can be connected up and left for long periods.
So do I have a faulty meter or is a characteristic that the Prema needs regular zeroing to be accurate?

Resonant

[Kosmic]

Mine is pretty stable over multiple days. You should check the electrolytic capacitors. Had to change all of them on the analog board and some on the digital board. With the old cap my unit was drifting badly.

[maginnovision]

Yea, I had the same issue. It was very noticeable because the bad cap was getting much warmer than the others.

I added a screenshot of the graphed data from my issue(reading a short). As you see it got really bad but now it's stable again.

[Resonant]

Thanks for the advice, I have new Caps on order and I'll report back on progress next week. I was slightly surprised to find CapXon and Jamicon Caps in a piece of expensive test equipment whilst none tested faulty the ESR was poor compared with premium brands like Panasonic or Nichicon.
Thanks

[Resonant]

Well I finally got the new Caps installed, I faced a bit of a delay with a problem on the connections on C15 which relies on the though hole plating. This stopped the unit booting correctly but was fixed by adding a link C15 to L2 and no further problems and it seems much more stable its no been on for 24hrs. For reference I used Panasonic FC  range 105⁰C from RS.
Resonant.