Heat treat oven temp variations

Joined
Jun 3, 2017
Messages
811
A few weeks ago i started a thread where i built a 120vac oven. It is working great and i have heat treated about 2 dozen springs and about 8 blades in it so far (i like to make springs in bulk, and blades on a per knife basis).

I have it set for 1475 and have not changed it at all. When i first start a batch of 6 springs, it starts with temp that cycles from hovering in the high 1450s to low 1460s, and will occasionally bump up to 1468 to 1472 or so and start to fall back to low 1460s and might drop to high 1450s, then hover a bit and bump up to 1468 to 1472 again, and so on.

Around spring 3 or 4 (i heat one spring at a time to prevent heat loss while quenching one spring), it will start to cycle differently. It will hover around 1480 to 1485 and jump up to around 1495, then drop to right at 1475 and then climb again to 1480s and hover, then jump up to 1495 again, and so on.

Around spring 5 or 6, it will stay right in the sweet spot 1470 to 1480.

Is there anything i can do to stabilize the temp cycles for the first few springs????

I have tried leaving it on for about 20 minutes before doing any heat treating, and it seems to fall right into the 3 to 4 spring area of running hot, as expected. I dont really want to run it 40 minutes before doing any heat treating. I assume the issues are small chamber size of 4" by 4.5" by 9", which is about 0.1 cubic feet. Also i suspect the bricks themselves retaining heat is messing with the pid controller. But i really feel i could actually just sit there with a switch and eyes glued to the temp display and do a better job controlling the temp.
 
Last edited:
The longer you leave the oven on the smoother the cycling will be. What controller do you have and what is the element trigger frequency. Some pids have long delays between cycles. This is to prevent burning out relays. But if your using SSR's you can cycle them very quickly. What happens is the pid knows it can't turn on for x time and it knows the temp is dropping. Once it can turn on it hits the gas and shoots right past the set point. It then turns off and the temp drops and the cycle starts all over. By letting the oven warm up fully you reducing the amount of temp the oven looses when the pid is off which means it does not feal like it has to blast right past the set point.
 
Ah, i see in the manual how to change cycle time. It looks like it needs to be as low as possible, which is 2 seconds. I also see how to start the autotuning. I may do that.

What exactly is an alarm in this context? I assume it isnt an audible alarm. It doesnt really say what an alarm is.
 
Actually it is an audible alarm. The alarm is used to signal when a process is out outside the temp parameters.

As JT mentioned the furnace will fluctuate temps until the bricks have fully soaked up heat which will stabilize the swing once fully heated. The alarms will signal a high or low condition based on parameter settings for this type controller.
 
Last edited:
Since my oven is very responsive (more than 1 degree a second) it says fuzzy logic isnt quick enough and to set P to 1. I bet autotuning and looking at the parameter values and making sure they are in the right range will make a difference. Any fluctuation less than +/- 20 degrees will make me happy.

I suspect i start the autotuning once the oven reaches 1475?

If i have to, i will let it run 30-40 minutes and then start HTing 12 or so springs at a time to make it worth it.
 
An alarm (in control terms) is simply a change-of-state that occurs at some specified condition. It's not something about which you should necessarily be alarmed.

For most of us, the alarms in the controller fall into the same category as the thermocouple types we are not using: we don't need them and can safely ignore them, but they are available if somebody should need them.

An example of an alarm in use might be as follows:

An industrial cooker heats a product to a temperature of 220 degrees and holds that temperature for 2 hours. An alarm is programmed to trigger an output relay at 218 degrees and this feeds a 2-hour timer. When the timer times out, either a flashing light/buzzer can tell the operator to do whatever he needs to do, or it can send a start signal to initiate the next step in an automatic process. If the product to be heated will have a different starting temperature on a cold winters day to a hot summers day, this could be a better way of doing things than simply using a timer on the heating step.

For a HT oven, an alarm setting likely to be useful might be a low deviation alarm. Setting this to, say, -2 deg would cause it to trigger when the temperature reaches 2 degrees below the setpoint, when it could switch an output relay to sound a buzzer telling the operator to start timing the soak. Using a deviation alarm, rather than a fixed-temperature alarm, would ensure that the buzzer always sounds 2 degrees below the setpoint, whatever it may be, eliminating the need to change the alarm setting for different steels/setpoints.

One thing to be aware of with the Autotune is that it will tune for optimum control at the temperature at which you tune the controller. You need to pick your tuning temperature for the temperature range you intend to run the oven. With a high-power heating element, tuning at Austenitizing temperature will usually give the best response for Austenitizing, but will tend to overshoot significantly if you run at Tempering temperatures.

Tuning at your lowest tempering temperature will give optimized control at tempering temperature, but slightly slower response at Austenitizing temperature (usually a slower approach over the last few degrees to the setpoint). If you will use the same oven for tempering and Austenitizing, tuning at tempering temperature is advisable: a slightly slower approach to Austenitizing temperature will not cause any problems. Overshooting the tempering temperature can easily cost a Rockwell point or two.

If you have a separate tempering oven, by all means tune at Austenitizing temperature.
 
Yeah i made the mistake one time trying to temper in the oven after heat treating. I was foolish enough to think that if it cooled down to 550 degree on the display that it really was 550 degree. Turns out the bricks were much hotter and soon as i tried to use a spring and loaded it onto a blade it bent like a wet noodle haha. It must have lost several rockwell points.

I will mess around with the settings next time i HT and see what happens.
 
I just started heat treating 2 blades. The cycle time was already 2. I did the autotune and it stays within 1475-1476 consistently. I am very happy!
 
Back
Top