Very interesting.  What series resistor did you have these results with?  Very interested to compare spec'd temp co. Regards, Chris  Sent from my SMRTphone -------- Original message --------From: Jerry Hancock <jerry@hanler.com> Date: 2/20/17 2:51 PM (GMT-06:00) To: kc9ieq <kc9ieq@yahoo.com> Cc: Discussion of precise voltage measurement <volt-nuts@febo.com> Subject: Re: [volt-nuts] Fluke 332B I took some of the 100K 1/4 watt resistors from my DAS-46 and heated them with a soldering iron.  They ran high pretty quickly.  I then used a cheap, 1% Chinese brand metal film and they ran high just about as quickly.  The 5% Chinese brand ran low a lot faster.  Just bringing the soldering iron near them  Interesting in that using one of each, the resistance stayed about right on the parallel value.  I then used a high quality Vishay and I couldn’t get it to move with the soldering iron without touching it.  These resistors cost about .40 per at Mouser.   I was just using my Agilent DMM so I’m sure they were moving, just not within the resolution of the meter.  The bottom line is that the carbon comp I replaced with the 50 cent per Vishay was a good move.  Had I used the 5% I have, it would have been about the same.  > On Feb 20, 2017, at 12:36 PM, kc9ieq <kc9ieq@yahoo.com> wrote: > > Very interesting, very curious to hear your conclusion!  > > My thought would be to replace these with standard value 5% resistors having good temp co, as calibration should surely make up for any subpar values--  my thinking is that temp drift would be a more major consideration for overall stability.  If this is a false assumption of would certainly like to learn why.  > Perhaps the old Allen Bradley carbon comps were special in this regard, but the data sheet I've seen for currently available comp resistors had a horrible temperature coefficient--  much worse than the "better" film resistors available today.  I stock the Vishay PR02 metal films for rebuilding old tube stuff, which have a temp comp of +/- 250ppm/K.  There are much more stable options out there, but I chose this line because of the 500V rating and dark red/brown color which blends into an old chassis more so than tan or bright blue.  > > Regards, > Chris >