During yesterday's livestream on my Youtube channel, I got SL935 cable connections fixed up and all assembled in 732B-7001 chassis.
YoutubeSchematics for resistor standard was recovered, and oven settings captured. Oven inside SL935 is set to run at +35 C, instead of warmer +45C like 732B.
Lower oven temperature make sense, as main purpose why this prototype was build is metrology, and majority of metrology labs running 20-24C lab temperature. So +35C setpoint allow good 10C headroom against ambient and does not introduce too much thermal stress into the resistor assembly.
This also meant that I had to add 25.5Kohm resistor in series of R547 on charger board, to make IN CAL led working as expected, otherwise it detected thermistor temperature signal is too low and IN CAL was off.
Initial data after reassembly with oven working revealed <0.5ppm difference to standalone measurement I made week ago.
1 Ohm resistor is just simple parallel connection (maintaining 4-wire Kelvin layout) to 10 x 10ohm from networks.
10 KOhm output is taken using 9 x 19 ohm series from network + 10K+10K/10K+10K calibration network. 2.111111 Ohm from 19R network is likely just a trim to get <1ppm accurate 10K.
According to my 3458A (DIY calibrated in January 2017 vs double 10K foil resistor transfer vs 0.55ppm uncertainty calibrated SR104 (August 2016)) median resistance settle around 9999.9944 Ohm, just half a bee's dick short of ideal 10K
Will swap to another 3458A (same date calibration) to log for longer time.
I'm happy with the investment so far, given that I spent half of the price of Fluke 742A/IET SRL-1 and get both resistances.
For tempco measurements probably best way would be to collect some 24-48h data with oven ON, and then repeat same 24-48 measurement with oven OFF, keeping room temperature same (to reduce chance of error due to own HP 3458A tempco). 40K VHP103 inside 3458A expected to be worse than this SL935 box, so ACAL can't help much here either.
Eventually I plan to send this box for low uncertainty calibration lab over few years, to establish drift and long-term stability rate of both 1 Ohm and 10Kohm, but for now I'll just log it for few months to make sure it's stable.