Returning hobbyist seeking circuit design assistance

[HerpHandler]

Let me give a little of my background info. Years ago I got my Batchelors in EET, and promptly went into a field that used aproximately 1% of what I learned. I then went back to school to get as much programming training as I could. And then continued in the same field that used less than 1% of that. And as the old saying goes "Use it or lose it" Well Id say its lost :cry:

I used to be an active electronics hobbyist but fell out off the hobby for reasons beyond my control.

Well Im back and raring to go.

Im looking to design and build a control circuit for a resistive heat load (thermostat essentially) But not just an on/off type. Im looking for proportional control. By proportional control I mean I would like to take a resistive load using 120vac and vary the voltage to the load based on a temperature reading at a probe. For example set a setpoint of 75deg F. probe sees 72, Vout 120vac. Probe now sees 73 Vout drops to 80Vac etcetra etcetra.

I know there are commercial versions of this available, but with todays availabilty of PIC chips I would like to tailor my own circuit to my needs (ie. add high/low temp alarms etc)

Its been so long since I fired up my breadboards, soldering irons, and meters Id like to start off small by just getting a basic schematic started that would allow the load control, and Ill work on other bells and whistles in the future.

I assume that using a Micro Controller will be my best bet, and I have started looking into them. PIC being the ones i most likely will use due to FREE samples.

So, being I have rambled on for way too long already. Could someone find it in their heart to put me on the right track as to how to begin an endevour like this.

 

[computer]

If you're looking for an easy method to get into chips, I'd suggest the **broken link removed** range of chips. Easy to program and can be converted to normal PIC chips by overwriting the bootstrap.

 

[panic mode]

varying voltage is very uncommon way of controlling large loads such as yours. common method is to change duty cycle the load is on.
The smoother response the faster rswitching time. For DC loads it is common to use PWM for example. For AC load (probably your case) it is
common to control delay from the zero cross. The triac will stay on until
next zero crossing (end of current half-period).
this method is cheap, efficient, smooth etc but it does generate noise
so it is good idea to put filter to reduce such effects.

 

[Nigel Goodwin]

For AC load (probably your case) it is
common to control delay from the zero cross. The triac will stay on until
next zero crossing (end of current half-period).
this method is cheap, efficient, smooth etc but it does generate noise
so it is good idea to put filter to reduce such effects.

That's why you don't use it for heating elements, it's usual to use 'burst fire' control instead. This only switches at the zero-crossing point, so it doesn't generate interference like phase shift control.

Basically you turn the element ON and OFF over a period of time.

For example, assuming 50Hz mains - you have 100 zero-crossing points per second (which is a nice convenient figure). By turning it ON for 50 of them, and OFF for 50, you get 50% power - likewise ON for 10 and OFF for 90 is 90% power. You don't have to base the power over a 1 second period, it's probably common to use a longer period (say 10 seconds?), the thermal inertia of the element smooths it all out.

You obviously can't use this for a light bulb :lol: but it's perfect for heating elements.

 

[stevez]

Herp - if I were going to tackle this project (electronics is my hobby, not my profession) I'd try to break it down into smaller parts - less to go wrong and more opportunities to celebrate success (even though they are little successes).

A significant part is the PIC - that might be a good place to start. As Nigel's input suggests you'll likely end up with some variation of on-off control so your PIC is essentially going to be an on-off timer with a duty cycle that is determined by the difference between the sensor input and the setpoint. You could start the project in the middle and work your way out. Get the PIC up and running with a LED or two to demonstrate the output - and a pot to act initially as a temperature sensor. Learning how to do this may not be a huge undertaking for you but for some it might be enough to do all at once. The pot and LED will allow you to understand what you need to do in order to adjust the sensitivity (how much or little the output changes in response to a change in input).

From there you might replace the pot with a sensor and get that working. You'll still only be lighting a light but you've made progress. From there you might work on the output. You could take a shortcut and simply drive a relay - while not elegant it could work. Keep in mind that the on-off cycles that you'd need for common space heating can be measured in minutes or at most, seconds. The relay could eventually be replaced with a triac. You might also build the triac portion separately - simulating the output of a PIC with a switch. Debug that then marry it to the PIC.

The essence of all this is to bite off small chunks to get your skills honed and so that you work the bugs out of small sub-systems.

As I've surfed the various sites for information on PICs I've seen plenty on proportional control of temperature as well as Proportional/Integral/Derivative. If you surf enough you may find that someone has described what you want and has provided the code for the PIC. That is certainly a way to do it though you may miss out on some opportunities to build skills.

Good luck, have fun.