If anyone is interested in watching my eventual success or failure at this on the forum, I have an approach worked out now.
At least two heated containers for salts (NaCl/CaCL2 mix for treating, NaNO2/KNO3 for quench, I don’t really want to have cyanides anywhere near me if I can help it) are required, in principle mild steel vessels should be sufficient but I need to do some calculation to verify the load they could support when at 2/3 of their melting point (there will be an upper limit to the size of bath feasible with a given wall thickness, which may not be a linear relationship due to the amount of salt needed to provide sufficient thermal mass)… if not Inconel is just about feasible, (again, this imposes a size restrictions, this time through cost, as it rapidly becomes prohibitive as the scale increases).
Heating would be delivered using water cooled induction coils driven by the slightly dubious 1.1kW solid-state ocilator kits available on eBay and elsewhere.
Temp control I can cover using one of the spare PLC’s kicking about, 12 I/O’s ought be sufficient for several thermocouples per vessel and on-off switching of both driver circuits.
The two major challenges are both related to heat-flow.
In an ideal world each vessel would be in a refractory jacket to minimise energy use, but this increases the risk of accidentally melting the vessels on startup due to the crude on-off method of temp control, I need to better understand the thermal conductivity and convective flow characteristics of molten salts before making a judgement on this.
Secondly, the parts themselves are liable to be inductively heated if the frequency of the ocilator circuits is not high enough… I can either put a thermocouple directly on them and allow that to take over control once the part goes into the bath (requires making suitable probes), increase the frequency of the drivers (requires more complex electronics), or try to model the heat flow from the part to the salt bath and programme the PLC to account for this and add energy to the system in pulses that can be equalised with the salt bath.
Without doing all the maths, and at least one trip to the library, I’m not sure this approach is workable… But it still might be at this stage.