Fluid heater 3kW on PWM control?

[bill nahd]

Hello tech's,

I think i get a 3kWatt heater challenge to deal with.
Someone asked me if i could make a simple fluid heater control (for a distiller). It should work like a gas stove; when the fluid is boiling too much, just turn the knob a bit down...

I was thinking of a PWM control, but with 3kW?!?
Maybe a very 'slow' PWM control with a solid state relay?
An ssr switches at zero crossing, is that good for a fluid heater? (or: is it a resitive load?)
Or is it anyhow better to use a real pid? -heater control? (example schematics plzz )

I don't have details yet about heater type and distiller size and so, but i'm sure you can point me to the right direction.

Many thanks in advance!

 

[Les Jones]

You could use the very simple method of phase angle control using a triac. Just a high power version of a light dimmer. You could think of this a PWM with a repetition frequency of 100 hz. (Or 120 hz if your mains frequency is 60 hz.) A solid state relay with zero crossover switching would be a better solution as some power companies do not like phase angle control as it distorts the mains waveform. It also produces RF interference. You could drive the solid state relay with a low frequency PWM signal (Say 1 to 10hz) using an NE555 or a couple of comparator chips.

 

[Nigel Goodwin]

You don't use PWM for high powered heaters - you use 'burst fire control' - nice and simple, easy to do, and efficient.

Essentially you simply send a number of mains cycles, and then none - so 20 cycles ON, 20 cycles OFF would give 50% power.

 

[bill nahd]

Thanks for your quick answers Les and Nigel.

Hmmm i for some reason a 'power dimmer' did not jump to my thoughts. I think that can stay very simple then.
But, as i understand from your considerations, a 'slow' PWM (or busrst fire control, thank you Nigel) is a good idea too.

I know enough for now, thanks again.
*Thread closed.*

 

[Nigel Goodwin]

Thanks for your quick answers Les and Nigel.

Hmmm i for some reason a 'power dimmer' did not jump to my thoughts. I think that can stay very simple then.

Unfortunately it's not as 'simple' as you might like to think, as you need to heavily filter it to help reduce the massive amounts of interference generated, which is why phase control is only used for small loads.

But, as i understand from your considerations, a 'slow' PWM (or busrst fire control, thank you Nigel) is a good idea too.

It's not 'my' idea, it's how it's always been done, from when thyristors (back before triacs) first came out.

Burst fire, using zero voltage switching, produces no interference, and thus makes the design a lot simpler and cheaper.

 

[chemelec]

You could use a Thermistor to sense the Temperature of the liquid in the distiller and have it control the gate of a SCR or Triac.

 

[KeepItSimpleStupid]

Nigel:

I don't buy your phase angle fired used for small loads. 100 A isn't small as described here: https://www.eurotherm.com/products/power-control/power-controllers/7100a/

PID control, just basically makes the setpoint agree with the measured value and compensates for undershoot/overshoot. PID requires step changes in the setpoint to set the tuning constants. I have not done PID control with auto-tuning, but have implemented PID in software. I also did some wierd PID control by controlling the proportional band (setpoint) of another controller. i.e The controller was designed to control the oven temp, but I wanted surface temp, so i read the surface temp and used my own PID to modify the oventemp.

Auto-tuning controllers are great. In general, the PID controller might output a 4-20 mA or voltage signal and the load controller implements the phase angle firing, or slow cycle triac or Zero cross pulse. Phase angle firing gives you the finest control. Slow cycle triac gives you the lowest interference and slow cycle triac is essentially relay control.

The power controller can be just that. It can control power or it's input is usually proportional to V^2. Furthermore there may be "tungsten" controllers where the load is non-linear. In order for the power controller to operate into an inductive load, one has to incorporate current limit.

Finally, in one system I designed we used 1500 W DC power supplies for low voltage custom tantalum heaters. This allowed EASY computation of power.
Previously, we used phase angle fired controllers operating into variacs. The new design made the panels smaller and we didn't need to build custom hardware (Variac/meter panel) and the values were actually correct for actual power measurement. The panel space was 1 RU (Rrack Unit) for the power supply/metering and 1 DIN for the controller.

 

[gary350]

This is what you need. You can set the temperature to what ever you need. It has a ramp control with memory so it will never over shoot the set tempeature. It is rated 3 amps it can be used to turn on/off a large relay or motor starter for high amp loads. You will need thermocouple wire 12" is all you need. I use these at work and the thermocouple can be 15 to 100 feet from the controler but you and use any type copper wire between the control and thermocouple. It is not true the entire length of the wire must be thermocouple wire. I run 25 feet of copper wire from the control to the work area then I attach the thermocouple to the copper wire with wire nuts. I make my own thermocouples by twisting thermocouple wire ends together then clamp on a wire stake on terminal. These are accurate to .5 degrees F.


**broken link removed**

 

[Les Jones]

As the requirement is to control the rate at which the liquid is boiling controlling the temperature is not the solution as a liquid boils at a fixed temperature. 1 degree below the boiling point it will not boil. Increasing the power input will not raise the temperature. It just increases the rate of boiling. As the distillation progresses the boiling point will probably increase as the lower boiling point components boil off.

 

[KeepItSimpleStupid]

Interesting. I wonder if you could control vibration of the liquid? No clue though. **broken link removed**

==

As for gary, thermocouple extension wire is used to make long connections. Each connector can be thought of as two 1/2 junctions and it's important to make them isothermal. I've seen a LOT of crappy ways to measure the temperature at the instrument. In fact, a device my Measurement Computing was affected greatly by a fan nearby. Cold Junction compensation is usually valid for a subset of temperatures.

K thermocouples are probably the most used. T is used for cryogenics and room temperature measurements. Then there are isolated junctions, exposed junctions and grounded junction thermocouples. I've made most junctions: torch for platinum, spot welder for some and a carbon block and spot welder for others.

 

[dr pepper]

I maintain poly machines where I work, they have custom thyristor modules driven by a plc driving multiple heaters up to 75a, the thyristor modules are just fancy solid state switches.
You dont have to use a pid if your not that bothered about accuracy, on/off control will work but you will get overshoot, the overshoot depends on the system but as a rule the more powerfull the heater the more the overshoot.
A good way of doing it is to have a temp sensor on the heater itself which can be used for overtemp protection, and another sensor on/in the product to provide the pv.

 

[alec_t]

I'm with Les Jones (post #9). It's power, not temperature, the OP wants to control. If it were me looking for an off-the-shelf solution I'd try to source a good old-fashioned 'Simmerstat' control from a domestic stove, and if necessary use it to drive a slave contactor/SSR.

 

[gary350]

I built a still about 15 years ago and I used one of these controllers and it worked perfect. My still was 5 gallons with a hot water heater element inside. There is a copper pipe that comes straight up 6 feet makes a 90 degrees turn then goes down. The thermocouple was attached to the 6 ft tall copper pipe about 6" from the top. Ethanol boils at 173.1 degree F so that is the starting point for the controller. Ramp is set at 10% and the control turns on a 30 amp motor starter that is connected to the 240 volt heating element. Liquid starts out at about 70 deg F and heats up 90% to about 162.7. The control does a 10% shut down then turns on/off at a certain test rate while it takes a test of the real temperature increase. The mini computer decides what to do and it will either speed up or slow down to slowly approach the target temperature 173.1. It takes about 30 minutes for the mini computer to caliberate itself and the alcohol boils off with minimum water.

If you have 90% alcohol you can distill it again and get about 95%. Then you mix it with SALT. Salt soaks up the water but not the alcohol, the liquid you have is 100% ethanol.

5% water is ok with me I was using it in a rocket engine. Water turns to steam its volume increases 1800 times so the water produced just as much power, maybe more than the alcohol. The water vapor makes the exhaust gas white color you dont get that with 100% alcohol.

 

[bill nahd]

Well thank you all for the information.
I learned a lot.

gary350 : Thank you very much for the info. He bought a stainless steel beer barrel of 50 litres that will be modified with heater element and copper pipes and so. I will pass the tips to him, he'll need them.

I like **broken link removed** solution. Will try this, to begin with a cycle time of 10 secs. replace the transistor with a solid state relay of 25A.

Thanks.