Incubation chamber

[Ales97]

Hello,
I'm working on a project involving 160 w, 230 V, 8,5 m long heating cable wound on a copper chassis. The heating element heats approx. 2 cubic meters of air inside an insulated rack. I'd like to put the chassis on the floor of this rack and measure temp somewhere near the middle of it.
The question is if there is any simple way to control the amount of the current ("heat"?) flowing through the heating element. Maybe I am overthinking this, but I saw a solution with SSR and using extra low frequency PWM pulses controlled by PID and it just doesn't seem right to me.
I want to achieve direct control of the current flowing through this cable with an ambient temperature sensor - If the settemp is almost reached, it lowers the current, and vice versa. I suppose that mosfet could do the job instead? Or should I switch the current manually to get the fine tuning done (So it doesn't switch too often) and let the PID handle the SSR like a bang-bang.
I still struggle to understand few things but, if the cable emits 18W/M and I lower the current, I can get lower wattage per meter, right?

I hope that my question is clear and I'd really appreciate any suggestions and responses.

 

[pc1966]

Most heat control systems are of the "bang bang" sort where the thermostat simply switches the heating element on or off depending on the temperature, and with a few degrees of hysteresis so the amount of cycling (and thus wear on any mechanical switch) is kept to a reasonably low level.

To vary the power using a PWM system is unusual and only needed if you require accurate control and/or thermal cycling presents other issues with your application. For most AC system the PWM is achieved using a triac (or back to back thyristors) where they switch at a varying phase angle of the AC supply. This is how most light dimmers operate (but without temperature feedback, of course!)

The Wikipedia page on that:

Phase-fired controller - Wikipedia

en.wikipedia.org

 

[Lucien Nunes]

a solution with SSR and using extra low frequency PWM pulses controlled by PID

Sometimes known as 'burst firing', this is quite a common method. It has the advantage that if the power device uses zero-crossing detection to synchronise the switching points to the power waveform, it does not generate fast rise-time switching edges hence does not need a suppression choke as it does not generate RFI. A common hardware implementation that is better than bang-bang but requires less setting-up than a PID, has a proportional band possibly of variable range, between permanently-on and permanently-off, within which the burst duty-cycle varies from 100% to 0. Obviously with whole-cycle burst firing the transfer function from duty-cycle to heating power is linear, unlike that with phase-angle control.

 

[marconi]

CH102 4-Digit Dual Display PID Temperature Controller (48mm x 48mm) (Thermosense Direct) - https://www.thermosensedirect.com/acatalog/ch102_4-digit_dual_display_pid_temperature_controller_48mm_48mm.html#SID=260

MC810B Digital Heating Controller, Temperature Controller - http://www.electrothermal.com/product.asp?dsl=765

R38S-HARRB-A-E-01 - 100-230 V supply 75mm x 33mm PT100 input PID controller, 2 x relay outputs | Tecnologic UK - https://www.t-uk.co.uk/product/r38s-harrb-e-01-75x33mm-pid-controller/?gclid=Cj0KCQjwsYb0BRCOARIsAHbLPhFyadwW5t4CXMze7TirsioU-3ll220xAASoKqxnOndCsFdXxs-nmrsaAg6hEALw_wcB

 

[Ales97]

Thank you all for the response.A circuit with MOC3041 with integrated zero-crossing circuit and BTA16 could be the solution? Trying to make it simple as possible, but also very accurate. That is why I came with PWM.
Also trying to figure out even heat distribution, because the rack is 2 meters tall. The temp is going to be higher near the ceiling. I was thinking about installing a 100 mm flexitube and a DC ventilator leading the air from top to bottom to circulate the air and even up the temp that way. And to control it, maybe putting NTC near center of the measured volume and another near the ceiling. The setpoint would depend on deviation between those two NTCs. It is just an idea, I am a noob. Just asking you to push me in the right direction. Thank you again for those and any further replies.

 

[pc1966]

Simplest and most certain is you just buy a commercial temperature controller, such as those Marconi has pointed you to. For a zero-crossing relay you can get them ready built as well, such as:

RS PRO Solid State Relay, 10 A rms Load, Panel Mount, 280 V ac Load, 32 V dc Control | RS

uk.rs-online.com

When you say "very accurate" what do you mean here?

• Absolute accuracy?
• Stability of temperature w.r.t time?
• Uniformity of temperature over volume?

You will find it very hard to get a temperature sensor with better then 0.5C absolute accuracy unless paying calibration laboratory grade prices! Getting good resolution, on the other hand, is easy.

 

[DPG]

Agree with the above - why re-invent the wheel. Unless you really want to build your own.