Pulse proportional thermostat

[malc-c]

Guys,

I was thinking about designing and building my own thermostat for controlling the heaters in a bank of vivariums. Currently there are two manufactures who make a pulse proportional theremostat, but none who make dual / quad stats in one unit.

They work by reading the temperature from a remote probe and comparing it with the value set on a control knob. Here is a description from one of the manufactures site

The sequence of events are as follows: The temperature is set on the dial. The cage and heater are cold so the heater is on full power.As the temperature in the cage nears the set temperature the unit cuts the power down to pulses.The degree of pulse (i.e. the ammount of time the heater is on and the ammount of time it is off) is dependent on the cage temperature. In the early stages of approaching the set temperature the 'on' pulses will be long and the 'off' time will be short. In a correctly set up cage, as the desired temperature is neared, the 'on' pulses become shorter until the cycle levels at approximately 50% on and 50% off. In a cage where the heater is of the correct capacity, the cycle should remain at this point. The heater is only powered enough to maintain the temperature set. It does not have to continually start the heater from cold. This vastly improves heater element life and is far safer for the animals as the heater rarely, if ever, needs to run at full power.

I would welcome suggestions on what chips to use, and what temperature sensors (DS180? or LM35) would be ideal. I have some 240 solid state relays I was intending to use to drive the heaters, so all I really need is to work out the reading of the temperature and the pulsing method, and then replicating this for 4 outputs. I would still like to use a pot to set the temperature, but it would be nice to have an 16 x 2 LCD or sets of 7 segment LEDs to display the current temperatures the probes are reading.

Would welcome your suggestions, or if it's already been done, any links to similar projects

[Acetronics2]

Hi, Malc

see here :

http://www.picbasic.co.uk/forum/show...3&postcount=11

...

Alain

[malc-c]

Alain, thanks as always... 266 pages of excellent bed time reading

[malc-c]

Alain,

The pdf is based on exercises using the basic stamp and discrete components for a to d conversion etc. Before I get my EasyPIC board out, what would be an ideal Microchip PIC to use ?

[Acetronics2]

Hi, Malc

relies on the sensors you use ...

LM 35, 335 .... ( voltage output ) will need 4 ADC Inputs ... Obvious, eh ?

DS 18x only need 4 Bidirectionnal I/Os ( Onewire "net" will lead to deal with DS IDs ... eh ! )

Add a nice 16 or 20 X 4 LCD ... say 1 full port

4 Pushbuttons or better a Keyboard ( looks more " pro " LOL ) 1 full port

4 R/G status Leds ... or 8 leds ... 1 Full port


let's "proto" aboard a 16F877A or 16F887 ... no soucy if a 28 pins is sufficient ( 876A or cheaper 886), and use of a faster or Bigger ROM ( if necessary ) 18F4520, 4620 or 2520/2620 will be easy ...

GLCD aren't so good, generally, because of the PID loops - BUT here there's no hurry for calculations ... or driven with it's own serial Pic Driver.

What about ???

Alain

[Charles Linquis]

Always use an 18F part if you aren't concerned about cents.

[jetpr]

I use LM358 read from 0c to 1000c and thermocouple type k
NumSamples CON 48
'ADCIN EGTChannel, EGTSample '0 to 255 EGT sample
EGT=0
EGTSample=0
For i = 0 TO NumSamples - 1
ADCIN EGTChannel, EGTSample
EGT = EGT + ((EGTSample * 255) / 60) 'The equation gives the actual Celcius reading
Pause 5
Next i 'Add it all into EGT.
EGT = EGT / NumSamples 'Divide EGT by the Number of samples