First up, we need to mount our ballast. We mounted ours on the rear of the projector. In order for the mounting bolts to clear, we had to offset it from center. This position places one mount bolt between the reflector and the lamp mount, and the other is easily in the clear.
A pair of 3/8-inch holes allow the wiring to pass into the enclosure. Note that we've got plenty of clearance for the LCD's inputs.
To keep things cool in the lamp compartment, we ordered a few bits from Mouser electronics. We scored a thermal switch(Mouser part 802-STC-120), an 80mm 120VAC fan(670-OA125AP111WB) and a 15amp solid state relay(653-G3A-210B-DC5). The switch will turn on when the temperature gets high enough. (Just a note, we'll be replacing this fan with a pair of 80mm Panaflo fans. This sucker certainly works, but a pair of DC Panaflo fans will be far, far quieter. (For reference, the five Panaflo's in our old Sony CRT projector are quieter than this thing.)
To keep things easy to understand, we've split the circuits into two diagrams. The solid state relay (SSR) will power up the lamp ballast when it sees between 5 and 30 volts DC, just like a regular relay.
The cooling circuit is similar in simplicity. The thermal switch will power up the fan when the temperature rises above between 113º to 127º F, the switch closes and fan is switched on. (The switch isn't supremely accurate, but it'll do the job. This is the simplest fan circuit we could think of.) With this configuration, the fan will run after the lamp is turned off until the enclosure is cooled below the trigger temperature.
There's quite a bit of room under the cover panel we modified. We mounted the SSR with a pair of machine screws where it'll clear the case.
To run power into the ballast, we drilled a 3/8-inch hole at a 45 degree angle leading away from the SSR. We ran the wire from the ballast along the inside edge of the enclosure and through the hole to the SSR. We cut and stripped about a foot of the outer insulation off the power leads. We cut the white lead short, stripped the ends and connected it and the piece we cut off to the load terminals of the SSR.
To provide the 120VAC to the SSR, we flipped the power supply board over and soldered the long leads from the SSR to some handy exposed pads on the the power connector of the LCD power supply. The two load wires are directly attached to the power connector terminals, but the ground was safer and easier to connect over near the chassis ground screw hole.
To mount our thermal switch, we made a small L bracket from some of the material we removed from the LCD enclosure. Then we pop-riveted the switch to the metal, using the mounting brackes that came on the switch.
We mounted the switch near the lamp inside the enclosure where it won't interfere with the light path.
To supply power to the fan, we drilled another angled hole. Then we spliced the black lead into the lamp power lead. We used the screw terminal of the SSR to connect the other side of the circuit. (And we insulated the exposed connection with some quality electrical tape.)
Once everything was wired up, we zip tied the excess wiring into a fairly neat bundle.
To make our cooling panel removable, we glued in a pair of crossbars. Then we cut a 1/4 inch ply panel to fit the space.
To mount our cooling fan, we traced the inner diameter and cut out the circle with our jig saw. We decided to mount the fan on the outside of the enclosure (because it's the ugliest hole we've ever cut). We then mounted the fan with some spare hardware.
Before wiring up the trigger circuit to the SSR, we need to test our our new circuits. It's better to do it now when the system is easy to troubleshoot. To get our trigger voltage, we unplugged the power to the LCD board and used some thin wire to tap 5VDC. (The leads are labeled on the LCD board, so it's easy.) You can see the MH bulb lighting up through the hole for the LCD data lead.
The fan spun up exactly as planned once the chamber reached temperature. We'll add a inner metal shield to keep the light from flooding from the fan. (And maybe something along the edges to stop the light leakage.)
To trigger the SSR, we need a 5 to 30 volt DC power supply. The LCD board has a nicely labeled backlight enable. Unfortunately the LCD turns off the backlight anytime the input source is changed. Metal Halide lamps need several minutes to restart, making this solution impractical. Go ahead and disconnect the backlight enable wire to keep that high voltage supply from powering up.
Since we needed a 5-30v source that's powered up when the LCD is on, we tapped the 12v supply at the audio amp chip. We're leaving the speakers out of our projector, so it's a good choice.
It took a couple of tries, but we managed ot solder a small lead onto the Vcc pin of the audio amp. (We used a fine tip 15 Watt Weller.)
We anchored the lead with a zip tie and ran it to the input of the SSR. (Note the electrical tape holding down the AC wiring to prevent rubbing on the pc boards. To ground the DC supply, we spliced into one of the ground leads from the power supply.
To mount the carriers we made, we drilled some slots half way up the carriers. Then we lined them up, drilled starter holes and anchored the carriers with a pair of sheet metal screws. The fresnel should be 220mm from the center of the lamp.
We donned our handy latex gloves and slid the LCD home. The cable just reaches
to make the connection.
We need to cut a slot out of our lexan to allow the PC board to pass through, not to mention airflow.
Lexan is strong enough that we were able to use our table top router to cut out the slot. If you don't have one, you can just cut the bottom edge short, and put in riser blocks similar to the LCD to hold the panel in place.
With the lexan in place, we had to give the lamp and LCD a try. At first we wondered if the palm trees were actually signs of a broken display! Lucky for us, everything is working perfectly.
To trim our lenses to fit, we used a T-square and an utility knife. With the lens aligned, we made multiple passes with the knife until the lens was cut. We had to trim the top and bottom of the lenses to center each panel.
We tried the same thing with the lamp side Fresnel lens. The legs work, but in the end we broke one off because of the brittle material. In fact, our biggest error was forgetting to measure the width of the lenses when we built our carrier. The Fresnel lenses are only 17 inches wide; narrower than the channels we built into the carrier.
To fix our width problem, we needed to fill 1/8-inch on either side. We slid some 1/8-inch dowel rod into the channels.
We shored up the problems by using some CA (superglue) and some 1/4-inch bass wood. We attached a small block to hold the panel at the correct height and added some longer pieces along the lens channel to keep the fresnel firmly located.
To allow for keystone correction, we decided to make an infinitely adjustable mechanism. Two pieces of slotted dowel rod keep the lens straight. A spring holds the upper end against an adjustment screw. Adjusting the screw fine tunes the keystone adjustment. The screws will only be accessible inside the carrier, but once it's set, we'll forget about it.
A trip to the local ACE hardware store (they have the very best selection of hardware that we've found) yielded a 3/4-inch dowel rod, a pair of springs, sheet metal screws, the dowel rod we needed earlier, and those handy 1/4 inch pieces of bass wood.
We slotted the dowel on our table saw. We cut along 2/3 of rod, because that's about all we could safely cut. You could make a safer version using a piece of hardwood to cut the slot into, then cut the hardwood down to a good width.
We cut a piece just short of the width of the carrier. Then we added some blocks to the carrier to hold the dowel in place. The bottom dowel will pivot, but otherwise, it's going nowhere.
We cut the upper down down to be a bit narrower, and mounted a spring on each end.
We made a pair of these adjusters. We tapped them with a 1/4-20 tap and added a pair of screw holes.
With the adjustor installed, we mounted the opposite end of the spring to retain tension through the full adjustment range.
We needed to add a lower barrier just past the LCD to prevent light leakage. We cut a 1 and 3/16-inch strip of bass wood and made two slots using some 1/4-inch pieces. Everything is mounted to the carriers, so they can be moved at any time. Up top, we cut down the long frame piece and placed it over the LCD.
To focus the projection lens, we dug up some cutting board scraps and made a bearing carrier like the one we made for our CNC project. We drilled and tapped the second piece to act as the movement nut. It's smooth and has enough resistance to keep the carrier from moving once it's set.
We mounted each piece with a pair of machine screws. By necessity, the nuts are on the inside in each direction. The board will probably strip out if over-torqued, so be careful. We rounded it out with a knob from we scored at ACE.
To get things going, we mounted the control panel with some zip ties and a piece of lexan as an insulator between it and the metal case. We used a screw to mount the IR receiver with another small insulator. (We're thinking that a nice cover will be a good CNC milling project. Yo Ben
?) With everything wired up, we installed the final panels with a pair of screws at the edges and took the projector down to our theater room.
We fired it up on a temporary mount and started testing it out. We've got lots of room for tweaks, and we need to mount it up for reals.
We put our thermometer into the lamp chamber through one of the fan mounting holes. When the fan was on, it stayed below 120 degrees. Cooling air flows past the projection lens assembly, through the LCD/fresnel assembly and into the lamp compartment.
That's it! We'll be back with a review of the projector once we've spent some time with it and had a proper focus session and get it mounted up. Until then, we did a quick focus and watched Yoda kick some Dooku. So, you ready to take this project on yourself?