Electronic Temperature Monitoring and Control.
Background.
The system for monitoring temperatures has evolved over a number of years starting with a single inside temperature sensor feeding a data logger. I soon tired of the need to wake up the logger just to have a quick glance at current temperatures. In 2003 a remote temperature display was added together with a sensor preamplifier unit. An independant electronic temperature controller and a PAR sensor were added in 2010. The long term plan is to replace all of this analog system with a micro controller and logger system - if I ever get the time!.
Sensor Preamplifier.
 |
Sensor Preamplifier. |
The purpose of the sensor preamplifier is to amplify the low level signals from the temperature sensors and to provide auxillary channels and power for other sensors. The output from the preamplifier feeds to the house via an underground cable.
The schematic for the temperature sensor preamplifier can be found in the Temperature Sensor Preamplifier Schematic Page.
Temperature sensors.
The temperature sensors are based on the change in voltage drop of a diode
with temperature. The temperature sensors were constructed by
machining a small thimble from stainless steel and then encapsulating the diode and cable into the
thimble with epoxy resin.
 |
Temperature Sensor. |
To guard against incorrect readings because of heating by direct sunlight, the outside temperature sensor is shielded from direct sunlight by a curved piece of plastic painted white. The shield was constructed by cutting away the top and one side from a plastic bottle. The inside temperature sensor is of similar construction with the addition of some reflective thermal insulation.
| 
|
Inside temp sensor. | Outside temp sensor |
Despite the use of a silicon diode as the temperature sensing element, the calibration of the preamplifier unit is surprisingly linear over the range that is of interest. Calibration was carried out using well mixed water of varying temperatures in an insulated container. The temperature reference was an accurate temperature meter having an uncertainty better than 0.01 Deg. C.
|
Inside Temp Sensor Calibration. |
|
Outside Temp Sensor Calibration. |
PAR sensor.
The Photosynthetically Active Radiation(PAR) sensor is based on a photo diode operated in the photovoltaic mode. A BiMOS Operational amplifier amplifies the current from the diode to give a 5v output at maximum irradiation. The photo diode is mounted inside a diffuser machined from P.T.F.E. The spectral response of the PAR sensor has not been adjusted in the 400-700 nm range as is the case with commercial sensors so the PAR readings should only be regarded as indicative values.
The schematic for the PAR sensor preamplifier can be found on the PAR Sensor Schematic Page.
|
Home made PAR sensor |
Remote Temperature Display.
The temperature display below alarms below 2C and above 45C and in the event of power failure. It was built from salvaged components and what could be found in the 'junk box'. There is also a current loop to monitor heater activity. The current loop load sensor is an important feature and offers positive indication that the heater is actually drawing current rather than just being offered power by the controller.
The schematic for heater load sensor can be found on the Heater Load Sensor Schematic Page.
At the bottom there are outputs for connection to a data logger where the signals from remote sensors and alarm events can be monitored. The pairs of LED's give an immediate indication of any temperature violations, the second LED of each pair remains illuminated until the reset button is pressed. The unit also has an internal battery so that the alarm can sound in the event of mains power failure.
 |
Remote Temperature Display. |
Data Acquisition.
The data acquisition set up captures the inside and outside temperature sensor output voltages, PAR sensor output voltage and the load sensor event output that monitors the heater power feed. There is also an event input from the fan controller.
The data logger was a skip find many years ago. It has 4 DC inputs and can log 8k of data. Data is uploaded as a tab delimited file using the original manufacturers software running under wine. The limitation of this set up is having 5 channels and only a 4 channel logger. At the moment event input connections have to be swapped over depending on whether the cooling fan or heater is being controlled.
|
Grant Instruments Data Logger - a relic from the 1980's rescued from a skip. |
Temperature Controller construction.
The controller was constructed as an experiment in response to the poor performance of bimetallic thermostats found in commercial greenhouse heaters at the lower end of the market. It was made from salvaged electronic components from a UPS and housed in a waterproof plastic enclosure. The sensor circuit is based on a ntc thermistor and two op amp comparators.
The first comparator is set up to control a heater and the second comparator is set up to control a fan for forced air cooling in the summer months. Each comparator drives a relay capable of switching mains voltages at currents up to 10A. Getting the hysteresis right involved a bit of trial and error because in addition the the theoretical value from the circuit design there was the thermal mass of the sensor to consider.
There are event outputs for interfacing to a data logger and for future expansion, there are also control voltage outputs for driving solid state relays. The temperature sensor for the controller is independant of the temperature display and monitoring circuits. The added advantage of using this set up is that I can use just about any type of heater and because there is no watering of plants done between November and March when the heater is in place, a domestic fan heater can be used with the thermostat disabled or set on its maximum setting.
The schematic for the controller can be found in the Temperature Controller Schematic Page.
 |
Electronic Temperature Controller. |
Data Processing.
The uploaded data from the data logger is first imported into an Open Office spreadsheet to allow formatting of the date and time as a single Date &Time column and is then saved as a tab delimited file, one file per month of data.
Each file is then plotted using gnuplot. Prescaling factors and calibration coefficients are applied on the fly during plotting.
The corresponding gnuplot scripts used to produce each chart are linked from each chart page.
2010 Data.
The data plots below are quite large images and you may need save the image locally and then open it with a graphics package.
The 20th and 21st of October were the first cold nights for 2010. The 26th of October had a dense cloud covering and heavy rain. In the evening there is also one point where the outside temperature warmed up rapidly and overtook the inside temperature.
Change of logger unit. Backup Heater testing 12:00 to 1300 on 6th November following merging of load and
controller events into combined signal. Backup heater functional test during night of 6-7th November.
Backup heater thermostat setting reduced at 13:00 7th Nov. 22nd November, Adjustment to set temperature and hysterisis.
28th November, the coldest overnight temperature for November for a few years.
Continuation of unusually cold weather.
A few cold spells but milder in the middle of the month.
Some periods of sunshine giving elevated inside temperatures. Main door left open between 12:00 and 16:00 on warm days. There some gaps in the data where I was unable to download the logger before it's memory was full.
Extended periods of sunshine giving elevated inside temperatures and the odd night time frost. Main door left open between 12:00 and 16:00 on warm days. There some gaps in the data where I was unable to download the logger before it's memory was full. Shading material placed on south facing side on 24th.
Above average temperatures for the time of year and extended periods of sunshine giving elevated inside temperatures. Bubble wrap was removed on 16th April. Main heater removed and replaced with the circulation fan, temperature controller is now controlling the circulation fan.
A few overnight cold snaps early in the month.
The last few days of June with clear skies as can be seen from the PAR data has given rise to elevated day time temperatures. Unfortunately the logger was not downloaded in time and the data for 30th June to 9th July was not recorded.
