[volt-nuts] MAX6350 ?

[Andreas_-_Jahn at t-online.de]

   I made some measurements over about 1 week (156 hours)
   with a similar device: the MAX6250ACPA
   It is specified with typical 20ppm/1000h.
   Instead of having a constant 25 degree temperature
   (which would be the correct way to get the datasheet spec)
   I did some thermal cycling every day between about
   20 and up to 40 degrees (centigrade) and recording the
   voltage at the 25 degree points.
   The ageing was about 100uV during this time
   including 15uV thermal hysteresis.
   This would be 20ppm/156 hours or when projecting with
   the square root to 1000 hours about 50ppm/1000h.
   Since I investigated only one single device without pre-ageing
   I cannot state wether this value is really a typical one.
   If you really need a maximum drift over age in my opinion
   the only way is to pre-age the devices and select those who
   meet the desired target value.
   On the other side I have 2 other 5V buried zener devices (LT1027CCN8-5)
   which I have compared to 2 different Keithley 2000 after pre-ageing
   and one year later again. After having corrected the temperature error
   of the 2 zeners I got about 5uV difference (1ppm) over one year.
   (1 digit difference on one of the both Keithleys).
   So either all 4 devices did the same ageing or they are dammned good.
   By the way: within the year I did some comparisons to a 7V LM399 zener.
   Either the LM399 or the two LT1027 devices drifted about 8 ppm in the
   one
   direction and then back to about the original value.
   Since temperature effects are calculated out I think that relative
   humidity
   between winter and summer together with the plastic case of the LT1027
   plays a role.
   with best regards
   Andreas