I made some modifications to the schematic: adding an outlet to plug the element into, meters to monitor the power, a switch to go to full power and changed the dropping resistor to a light bulb. The light bulb serves as a current limiting resistor, just like the power resistor in the original design. I was hoping I could choose a bulb wattage so it would glow dully, but more on that later.
The case was salvage from a computer power supply. The triac came from eBay, the meters from MPJA
and the rest was in my junk collection.
First step is figuring out how and where the bulky stuff fits. The dimensions of the parts provide enough constraints that things converge pretty quickly. At this point I haven't found a space to cram the outlet. The paper templates for mounting the meters are CAD printed at 100%.
The dimmer was too big, so I cut off the mounting tabs. The heat sink interfered with the leads from the dimmer, so it got major violence inflicted on it with a hacksaw and file.
That left room for mounting the outlet.
The leads for the dimmer were to short, so I extended them with a black wire to the lamp socket and a black wire to the triac gate with a yellow lead to the switch.
Another black lead goes from the ammeter to the triac M2 and from there a red goes to the lamp socket.
Two orange wires (twisted together) will go from the voltmeter across the outlet terminals. The two yellow wires to the switch go out of the picture at the lower right.
Then the leads to the lamp socket are terminated. The wires go around the socket to prevent having a sharp bend right at the terminals. The switch wires go off to the left and the voltmeter leads go to the right.
The supply line is connected: green to outlet ground, black with spade terminal is the hot to ammeter, the neutral to triac M1. Because things were so tight in the case, I had to dismount the heat sink for the triac. The heat sink is a CPU cooler less the fan. I ground the heat sink and the triac against a flat surface to confirm I would have good thermal contact. The white ties the ammeter to the outlet. The orange voltmeter leads are terminated across the outlet. The yellow switch leads won't be connected to the switch until the case is assembled.
The heat sink is reinstalled and all the connections snugged down. The case is assembled (omitted to spare the squeamish - there was a lot of hacking, filing, drilling, grinding to get the case halves to go together. The switch mounts on the other half of the case, so it's leads run through the switch mounting hole and are soldered to the switch. Another brief but violent episode as the switch gets its little feet and locking tabs crammed back into the case. Finally, some new holes are drilled in the case to secure it -- the meters covered two of the original holes.
The moment of truth: will it work or will all the lights in the neighborhood go out?
IT'S ALIVE!!! (*deranged laughter that goes on for an unpleasantly long time*)
The dimmer won't bring it up to full power - there must be a tiny delay before the diac fires, possibly because the voltage has to reach some minimum value before it conducts. Sadly, the bulb doesn't glow - the gate current for switching the triac must be very low. I settled on 25W by experimenting with wattages from 7.5 to 200. With wattages over 40W, the controller won't go below about 35%. This requires more experimentation.
Secret Switch Saves the Day. By shorting directly across the dimmer, the switch clamps the triac into full conduction.
make an esthetically pleasing knob for that massive feel of raw power and exotic excitement. A friend pointed out that since my entire house has only two 15A circuits for lights and outlets, I should just wire the thing straight into the mains and dim the whole house at once.