Thermal store measurement/control

Thread Starter

geoffers

Joined Oct 25, 2010
475
Hi all,

Wasn't sure of the best place to post this,
We are off grid and have two thermal stores, one in the house heated by wood burner, the other in a shed around 100ft away (cable length) which will be heated by the water jacket of the backup generator.

The stores will pump from one to another depending on where the most heat is

I fitted 3 thermocouple sensors to each store and want to use a pic to decide when to pump and which way etc

Question is which is the best way to achieve this? I have a few cat5 cables I put in so I could,
Use 2 pic on for each store and send the 3 temperature readings down the cable.

Use one pic,
Connect the 3 thermocouple from one store to the cat 5 cable and send the very small signal down the cable.

Have the amplifiers at the remote store and send the amplified signal down the cable.

I know pic don't like long cables and am not really sure which way is best?

There is a 240v mains cable in the same duct as the cat5 but it's armour cable with the outer earthed.

Cheers Geoff
 

Ian0

Joined Aug 7, 2020
9,671
I would recommend thermistors over thermocouples for that temperature range. Thermocouples are dependent on a cold-junction compensation (which is a thermistor) for their accuracy.
A thermistor can be put at the end of a long cable with virtually no error due to cable resistance, perhaps a bit of interference pick-up which can easily be filtered out, and there are no problem with remote power supplies and differing ground voltages.
A remote thermocouple would need cables of the same metals all the way to its amplifier, otherwise another metal-to-metal junction with its own Seebeck effect voltage would be introduced.
 

Thread Starter

geoffers

Joined Oct 25, 2010
475
Thanks, so something like this,
Screenshot_20230118-140813.png
I only picked the thermocouples as the had I think a BSP fitting, it will be a slow system,
Would digital filtering be enough
 

Ian0

Joined Aug 7, 2020
9,671
I would consider sticking the thermistor to the outside of the tank under the insulation - that’s how central heating thermostats seems to work.
You can get matched-curve thermistors for about £1 from Rapid Electronics. I’d use some filtering to remove the noise and interference, maybe just a capacitor, and do the rest digitally, as you will probably be implementing a lookup table to determine the temperature.
I can send you my handy look-up-table-generating spreadsheet if you want.
 

Jerry-Hat-Trick

Joined Aug 31, 2022
545
I favour the good old-fashioned 4 to 20mA communication of temperature between the shed and house. Also, whilst being contrary, how about using silicon diodes to measure temperature - nice and linear. More than one diode in series to get a bigger voltage swing per degree, near enough constant current of 100uA through them with the right resistor and a reliable 5V supply, rail to rail op amp to convert mV to 4 to 20mA. Comparator with a bit of hysteresis to determine direction of flow.
 

Ian0

Joined Aug 7, 2020
9,671
If you want linear, but you don’t want to have to calibrate, then there are some very reasonably priced platinum film sensors available these days, both 100Ω and 1k.
 

Thread Starter

geoffers

Joined Oct 25, 2010
475
Thanks ,

I hadn't considered sticking them onto the cylinder, there's plenty of places I could stick them, it doesn't need to be a hugely accurate system so I think I'll go with some cheap thermistors .

Cheers Geoff
 

Thread Starter

geoffers

Joined Oct 25, 2010
475
Just got back to looking at this, as well as the thermal store I have a cooling system to try and make.

The platinum film sensors look like a good cheap way of doing it.
Found this on the TE website

Screenshot_20230224-132837.png

This looks like a nice effective way of cancelling out the lead length.
 

Thread Starter

geoffers

Joined Oct 25, 2010
475
Just going through the numbers to work this out,
DIN class A is +-0.15 % (one application is very fussy about temp the other a bit looser so will have to go with the fussy one!)
With a temp coefficient of 0.00385ohm/ohm/degC

Comes out as 3.9ohm change of 10*C between 0 and 10*C

If I'd like to have +- 0.1*C around 0*C I'm looking at a change in voltage of 0.00039 v/v over the Wheatstone bridge above? Have I got that right?

So I'm going to need a instrumentation amp for each sensor.

Anyone suggest a good/cheap one, I do have a couple INA122 but they are getting a bit expensive now.

Cheers Geoff
 
Top