Sorry folks, I couldn't post before due to
this cookery, but now I'm with you all
First hats off to
mimmus78, this was unexpected joy to see someone actually going to order my volt-nut design
The issues with soldermask opening is from my gerber file, I forgot to update the file B01. My own boards were corrected B02-version.
However those will not affect anything in measurable way, so it's still alright to use. If it bothers you, soldermask can be carefully scrapped off, but I'd leave it as is.
Other thing is manufacturing tolerance specs. I obviously did not expect someone to order the PCBs on usual hobby-level run fabs, so Gerbers are at 4mil/4mil spec. It does not have to be that precise, so I'll update it to be more friendly and replace GERBER on the article.
As of resistors, I would try to have best stability/tempco resistors for 1K/13K and 120 ohm ones, as these are key for LTZ1000 operation temperature and current. 70K can be less fancy and cheaper, and even vary in value 65-80K. You can see on HP 3458A A9 ref board there are just usual metal-films, and that board is totally fine to get <0.05ppm/K tempco, I tested it.
Opamps I use on my modules are controversial LTC2057 and now ADA4522-1 on this latest module covered in this thread before. No issues discovered.
Module can work with any input voltage from 10V to ~20V without problems and should not show dependence from input voltage. Regular LDO like 7815 with few decoupling caps will do job fine. Avoid switching PSUs though. There is no polarity protection on board, so double check connections and wiring before turning power on!
Normal current consumption with 1K/13K heater setting is ~28-32mADC steady state. It warms up to operation spec within 5-10 minutes.
Be careful with output too, it's not protected and directly connects to zener in LTZ1000A. If you short it or inject current into zener you are likely to cause stress/permanent damage the chip.