How to control temperature accurately...


Thread Starter


My application requires to maintain the temperature of process fluid coming out of the heater at 450 +/- 1 deg c. switching on/off of the heater doesn't allow me to maintain at this temperature accurately. There is an error of +/- 3 deg is found. Is there any other method such as controlling the power of heater or so to accurately maintain the temperature.


Terence B. Creevan

I have worked with heaters that accept an analog input to control its power. You can use a simple PID control loop with a temperature
sensor as your process variable (PV) and the heaters analog input as your control variable (CV).

[email protected]

Matt Warshawsky

With an electric heater you can easily control the output of the heater using an SCR and an analog out channel from a PID controller or PID software. If you can't vary the power of the heater, many PID controllers will cycle the power
to the heater at a rapid rate to simulate partial power (which as far as I know is exactly what an SCR does).
check on some of the equiptment used in the plastics injection industry. there are mold heaters which pump hot oil or water through the mold's bases. this equipment is sometimes available used and may be able to be adapted to your process. i'm not sure about the specifications but there will be many
manufacturers to choose from.

also try a different 'cycle' time on your controler. the cycle or reset rate controls how often the heater is on. for example; if the controler calls for 50% heat, is the on/off time 50% of 60 seconds, or 50% of 2 seconds? this is
part of pid tuning and can get pretty tricky.

hope this helps
Tim Davis

Zan Von Flue

I would also use a PI(D) controller and if that's already the case then try a line phase controller. It's better to control then a relay.
One of the problems with accurate temperature control is not the control response of the control circuit but the ability of the heater to respond effectively. The larger the heater (or heat rate) the more difficult it can be to respond quickly and accurately to a change in the control requirement due, for example, to a variation in the inlet temperature of the fluid. In the oil industry the problem of accurate temperature control in sample streams has been
a major handicap to temperature critical analytical measurements such as of viscosity. Usually +/-1degF is as close as a temperature bath can come. heat exchangers or heaters designed for high heat transfer rates are good at the rate of heat transfer but less good at responding to changes in the heat rate
requirement due, for example, to changes in the inlet temperature or the product stream flow rate.. In a joint paper given at the Texas A&M 2000 instrument symposium Mark Shelley introduced the concept of dynamic mixing in a paper entitled "Direct Measurement of Viscosity" where for the first time heat exchangers could be
used to to maintain close temperature control using dynamic mixing. Quite simply this uses hot and cold stream mixing the way you control the
temperature of water in a shower using the mixing tap. Part of the flow stream is diverted around the heater and is the cold stream. The rest of the flow passes through the heater and is heated above the target temperature. The two streams are then recombined through a static mixer. temperature measurement at this point is used to control the relative proportions of the hot and cold streams. A second temperature sensor in the hot stream only is used to compensate the heat rate for a trend (offset)in the inlet temperature.
This solution can also be seen graphically in the SOlatron Viscosity Systems brochure which can be downloaded from their web site ( "":
go to viscosity then to download ) This provides very fast and accurate response to rapid temperature transients while the slower response of the heater control accomodates trends.
Note that for fast control response the temperature sensors should be very accurate and very fast responding themselves... this may mean direct insertion rather than using thermowells.
None the less, this system produces temperature control in a fast flowing product stream at least as stable as achieved with a temperature bath.

Vladimir E. Zyubin

Hello List,

The algorithm:


Good luck.

(if it will not work you need increase frequency the loop, etc.)

Omer Haluk Yilmaz


It seems you don`t want to redesign the heater, then perhaps it`s a good idea to look for `thyristor controlled drive`

Vladimir E. Zyubin

>It seems you don`t want to redesign the heater, then perhaps it`s a good idea to look for `thyristor controlled drive` <

It was a spinal reaction on the advice "to use an SCR"... :)

But... the small words "etc." means,
there could be a lot of reefs... for example,
there is the supposition (mother of all errors :0) --
the delay (ON -> I -> T -> Sensor) is small enough...
lack of specification.

Best regards. Vladimir