| General > General Technical Chat |
| PWM'ing a 250W heater. |
| (1/6) > >> |
| paulca:
I am trying to heat a plant incubator for getting ahead with some seedlings before summer. It's a pretty small space and a 250W desktop heater is able to heat it as long as the extractor fan is very low. As the temperature rises however the thermostat on the heater cuts it out around 26*C, it sits idle for many minutes while the temp falls right down to 14*C when it clicks back on and rises pretty rapidly back up to 20+C and then slowly to 26*C when it cuts out again. So the mechanical bi-metalic strip in the heater is way too slow. I got the bizarre idea to automate the heater using a "smart" switch and instead of choosing a basic target temp, current temp feedback loop, I considered instead using a PWM cycle and a tracker algorithm to adjust the duty cycle to acheive and maintain the target temp. Maybe a 1/60th Hz frequency, meaning the heater will switch ON, OFF, ON every minute with the ON time varied by the algorithm. Is this even slightly sane? EDIT: 1/60th Hz might be a bit rapid and wear the relay out. Maybe a 3 or 4 minute cycle. |
| paulca:
Just to add the competing technique suggested to me is to place a barrel of 20 litres of water in there, preheated to around 22*C. The idea being the thermal mass will soften the saw toothing of the temperature. Additionally that barrel of water could be heated with a 250W fish tank heater and the only heat in the incubator would be passive heat from the water barrel. |
| mikeselectricstuff:
plenty of triac-based temperature controllers out there - why reinvent the wheel |
| Gyro:
The idea of skipping complete (or multiple mains cycles) is common for high thermal mass items like heaters. It is commonly known as burst fire control. It's usually done with a Triac or SSR rather than a relay though, and at mains zero crossings. |
| Ian.M:
I assume the heater is fan assisted. First hack the heater so the fan stays on when the thermostat contacts open. As-is, the residual heat from the element holds the thermostat off until the rest of the enclosure cools significantly. If you maintain airflow so its thermostat is sampling the general air temperature, not the locally elevated temperature inside the heater, you'll get far smaller temperature swings. If that's not good enough, you'll need to add a temperature controller, that uses cycle skipping to control a mains load, preferably running a PID algorithm, and hack the heater so the controller feeds the element not the fan, which should stay on continuously, as it wont like cycle skipping or frequent stop/starts. |
| Navigation |
| Message Index |
| Next page |