Hi all; as per the title really! I’ve just bought a pair of the MK138 thermostat kits and would like to know if they can be commoned up to a single NTC sensor without lots of additional buffering circuitry? Will be putting them together tonight but any thoughts would be appreciated in advance. basically I’m after a three state control of mains fans, specifically: OFF - SPEED CONTROLLED (via variac) - FULL SPEED
much obliged…
The Guvnah
This has never been tested.
We’d recommend to use two separate sensors and put them very close to eachother, so that they measure an identical temperature.
Crikey; I’m at ‘the cutting edge’ already! 
Thanks for your thoughts Vel. I’ve been trying to apply my limited electronic knowledge to this and thinking around the various factors that are going to have an effect. my main concern are the combined tolerance variations between the same components on each board all combining to produce different response curves from the two boards. Not least the slight differences that are bound to exist between any two randomly selected NTC resistors. That was the original inspiration for using a common sensor between the two boards; so that they are both singing from the same hymn sheet as it were and both will have identical input signals to work with.
Thinking further I can see that I’d probably have to “bookmatch” every input side component (in particular R4, R5, R6 and C3) so that differential voltages don’t arise in the shared circuitry and get ‘exported’ to the other board. 
Then there’s the question of how the ‘paralleled’ NTC affects the individual input impedance of each board?
I’m planning to double stack the two boards as closely as possible to enable them to be mounted in the limited space available in the finished unit. It really is tight for space in there and there is a lot of 230vac cabling in it as well so the unit has also to be shielded in a casing/barrier to ensure segregation of ELV and the mains which of course takes up more valuable casing real estate. The intention is to take the potentiometers off the board and panel mount them, the connector blocks will be replaced by hard-wiring. The LEDs will be taken ‘off-board’ to the front panel and I’m using my own relays in DIN rail sockets. By laying the electrolytic capacitors on their sides I can stack the boards just 11mm apart on pcb spacers making a final finished unit thickness of about 26mm! Not bad! 
The only other question was the maximum permissible distance between the NTC sensor and the circuit board. I’ve seen a figure of around 25mtrs used in connection with other controllers that use NTCs? I can’t see that a few milliohms of added resistance that would be introduced by 5 mtrs or so of signal cable would be much of an issue and not something I’d need to worry about compensating for or trimming out. The capacitance might be a different matter though, not sure on that
The salvation of it is that the two switching temperatures will be at least 10 degrees apart so it’s not like I have to tune the two boards to within half degree c. or anything. The objective was to streamline it as much as possible and avoid encapsulating the two NTCs together. It seems an ‘inelegant’ solution if it could be done with just one and would also require a 4pole plug & socket for the sensor rather than two pole, not that it would be much of a pain to accommodate this if req’d. I suppose there’s only one way to find out isn’t there.
Incidentally I’m also modifying the upper range of both units to give a maximum switching point of around 40deg. C.
Oh yeah; is it possible to upload a pic or two of the finished working unit/smoking remains for the benefit of readers? Imageshack or such-like?
The Guv.
Small differences between parts are compensated by adjusting the potmeter…
5 meters of cable for the sensor should not be a problem. Keep it away from mains carrying cables.
Sorry, this forum does not allow pictures. Use a free web-based picture base instead.