home-brew PID controller complete

Started by Arcs_n_Sparks, February 26, 2006, 03:37:28 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Arcs_n_Sparks

Well, I finally got my home-brew PID controller finished. The control loops are working okay, but still need a little tuning. I also need to do a little code work on the rotary encoder used to set temperature, as it is not as smooth as I would like it.

Front view; includes three digit dual color LED display, rotary encoder for setpoint, and programming port on top.


Left side view; power input jack for electronics (lower left), modular AC power input plug, and input for ET-73 thermistor probe (top right)


Right side view; modular AC power output plug


On the BS, with red digits indicating setpoint


On the BS, with green digits indicating measured temperature. Unit is cold, with 100 degrees being low end of measurement band I programmed.

SMOKEHOUSE ROB

looks great[:)] good job on posting the photos[;)]

GrillinFool

Arcs, looks great, nice size. Have you tested the tempature swing range yet?

The Grillin Fool

iceman


gary_CO

Very nice!

I just have to ask... What is the db-looking jack on the top of your box for, or is it something that was already present in the box?

Inquiring minds want to know...

Arcs_n_Sparks

GrillinFool: swings a couple of degrees around the setpoint. I need to adjust the Integral loop constant to eliminate that. (If that was your question). Since I use integer arithmetic, the best I can do is +/- 1 degree around the setpoint.

gary_CO: that is a programming port that allows me to reprogram the microcontroller that is inside. I can change the code without opening it up.

IKnowWood

Lifespan?  
Replication costs?  
Patent?
Engraved and signed?
IKnowWood
Coming to you from the DelMarVa (US East Coast that is)

Look up Our Time Tested And Proven recipes

Arcs_n_Sparks

IKW,

All good questions. Rotary encoder is only moving part. The one I picked is a 24 position detent unit good for 30,000 cycles (a cycle is one full rotation CW, and one full rotation CCW). The only thing that gets warm is the triac (I am phase controlling the A.C.). I have not given it the thumb test yet (rule of thumb in E.E. is, if you cannot keep your thumb on a semiconductor part, it is too hot). For the engineer I am, it really needs to be beaten on for awhile to find weak points.

The Bill of Materials is currently at about $58. However, that does not include any printed circuit board, since the prototype was hand-wired. Definitely need a PCB to reduce assembly time. The guts looks like this (this was to remain classified, since it is UGLY):


The dual color LEDs are expensive ($16 total), so I am rethinking the entire controller/display design. I think I was working for about 10 cents/hour putting it together. Some tooling would be required to speed production (using a Dremel to hand-cut holes is okay for a one-off (and gives more time for whiskey sipping), but not okay for production).

This is not patentable, since there is nothing novel about it.

jaeger

<font size="3"><font color="blue"><font face="Andale Mono"><u>Very Cool</u> Arcs n Sparks!!!</font id="Andale Mono"></font id="blue"></font id="size3">



<font size="2"><b>Doug</b></font id="size2">