Design Concept:

Click to view attachment

Features:

• 720p, 1360 x 768 native resolution, gravity keystone correction (+/-15°)
• Projection diagonal focus 55" to 250" via external lever, throw 1.2 : 1 with screen diagonal 110"
• Inputs: RF for TV tuners (ATSC/NTSC/QAM), HDMI, VGA, component video (YPbPr/YCbCr), composite video (RCA)
• Controller features:
  
  • Monitors LCD power on status, turning lamp and fans on/off when LCD is powered on/off
    
  • Monitors LCD temperature and rotation of fans, shutting down lamp on error
    
  • Runs fans for 15 minutes after lamp is turned off to cool unit down
    
  • Indicator LEDs: Power, TV On, Fans On, Lamp On, Cool Down On, Fan1/2/3 OK, Overheated
• Case features:
  
  • Rounded corner detail for visual appeal
    
  • Rolling base with shelf storage for DVD player, speakers, and cables
    
  • Projector quick release from base for transportation

I was looking for a fun project and WOW did I find it working on a Do-It-Yourself Projector. Learning about and playing with optics/lenses, woodworking, metal working, electronic controller design and build were all enjoyable parts of my projector build experience. And the successful result was a highly featured big screen TV.

When I started the DIY Projector project, I already had a Sanyo Z3 (LCD 720p) projector in a basement home theater setup. So the plan was to build the DIY projector for fun and then find some other use for it. I belong to a swim and tennis club and for some time we have talked about doing movie nights for the kids. Also, watching televised sports is big where I live and I was thinking that a big screen TV at the club would be used a lot for 'sports' parties. So I decided that the DIY projector should end up at the club; this decision influenced many of the design features incorporated in the projector. For example, the base has wheels to make moving and storing the unit in a closet easier. The unit pivots easily to make aiming at screens of different heights convenient. The focus mechanism was designed for screen diagonal sizes from 55" to 250" to allow for many viewing scenarios (e.g., for dark movie night settings the screen can be as big as possible, for sports viewing in semi-lit settings the screen can be smaller and brighter). Keystone correction works when the unit is tipped upward (if on the base on the floor) or tipped downward (if hung from the ceiling). I also built connections for a remote switch box with IR extender to allow the unit to be permanently mounted in a closet for a rear projection setup some time in the future.

I found the Lumenlab forum pages full of useful information and ideas and took much inspiration from them. Using Google search with

site:lumenlab.com subject_term1 subject_term2 -excluded_term

where "subject_termX" is what you are interested in finding and "-excluded_term" lets you exclude terms, works very well to isolate the information sought.

Must reads from the forums:

• DIY Projector Guide from the Lumenlab wiki
• State of the DIY Projector pinned subject from Lumenlab forum > Projector Builder > Beginners
• A TO Z Tweaks

I found the free 3D modeling tool Google SketchIt to be invaluable for visualizing size/shape/look of projector parts. I would design a model for a part, check the fit with the other part models, adjust as necessary, then use the SketchIt dimension tool to make a blueprint for fabricating the part in my garage or shop.

Subject topic links:

• Experiments and Testbed
• Experiment Results
• LCD
• Viewsonic N1630W Strip/Disassembly
• Beseler VuLite II
• Projector Light Source Choice
• Case Design
• Case Construction
• Lens Focus Mechanism
• Lamp Position Adjust Mechanism
• Light Box
• Controller Design
• Fresnel Mounting
• LCD Sled
• Heat Management
• Testing and Tweaking
• Misc. (case rolling base, LCD AntiGlare, DIY Screen, cost roll-up)
• Future Enhancements

Experiments and Testbed:

Click to view attachment

I got started (after reading / exploring the Lumenlab forums, of course) by buying a 19" LCD monitor (size 386mm x 290mm) with broken backlight from a coworker, followed by sending off for a variety of both projection and fresnel lenses.

From the Lumenlab store:

• S15 220mm fresnel, S15 330mm fresnel - size 319mm x 252mm
• S15 triplet projection lens - focal length 320mm, diameter 60mm
• Pro 220mm fresnel, Pro 650mm fresnel - size 431mm x 406mm

From a company in Taiwan called 3dlens.com I obtained :

• large 330mm fresnel - size 395mm x 395mm
• small negative focal length fresnel - focal length -300mm, size 200mm x 250mm

From a company in Pennsylvania specializing in surplus optics called SurplusShed :

• Delta II (L3789) Big screen TV projection lens - focal length app. 250mm, diameter app. 140mm / length 160mm (huge)
• duplet objective lens - focal length 500mm, diameter 70mm
• front surface mirror - size 406mm x 292mm

I built a 'testbed' for experimentation from threaded rod and MDF/plywood. See picture above. With the testbed I could easily try various lenses and quickly adjust spacing between elements.

• Two of the frames were fixed to hold the rods in place. I used 1/4" rods but they were not stiff enough to hold the movable frames without sagging; 1/2" or larger rods would have been better.
• I used a specialty household light bulb (KX-2000) mounted on the rear fixed frame as a light source. See picture below. This bulb from lightbulbemporium.com had the advantage of being app. the size of the arc of the projector bulb I planned to eventually use. Also, it has standard household base so the bulb fit into an old light fixture I had lying around.
• The movable frames were cut in a shape to allow them to hang from the top two rods similarly to how hanging folders work in a filing cabinet. I made a movable frame for each fresnel lens and projection lens I had, cutting an appropriately sized hole for each lens.
• I printed a computer desktop image on overhead projector transparency sheets which I taped to a piece of plexiglass mounted in a movable frame to act as a fake LCD.

After some calculations using the focalcalc143.exe program found on the Lumenlab forums, I was able to project experimental images with the testbed !

Note: Keeping the lenses clean was a challenge especially around the MDF/plywood. From a suggestion in the forums, using cheap rubber gloves when handling items really seemed to help keep things cleaner.

Click to view attachment

Click to view attachment

Click to view attachment

Experiment Results:

focalcalc143.exe produced results that I could verify with the testbed. I highly recommend the program. I spent some time digging into the math/formulas/theory behind the program's calculations. See wikipedia/optical_lens for nice descriptions and explanations.

Click to view attachment

http://dev.physicslab.org/asp/applets/opti...ors/default.asp is a very nice java web application for ray tracing that allows for experimentation with lenses and mirrors and includes real time interactive changing of focal lengths and distances between items.

(continued next post)

Experiment Results (cont'd):

Click to view attachment

Projection Lenses

Lumenlab S15 Triplet : Of the 3 projection lenses I initially had, the S15 Triplet performed superbly and was the only one that gave acceptable results. Good focus edge-to-edge on my fake LCD screens up to app. 17" diagonal. The only concern with this lens was a short throw distance when used for front projection; however, short throw distance is a great benefit for rear projection. Specs: Focal Length 320mm, diameter 60mm.

Lumenlab Pro Triplet : I have read only good things about this projection lens, but it was out of my initial price range and I have no first hand experience with it. I mention it as the ideal against which I measure other lenses that I experimented with. Important parameters: Focal Length 500mm, diameter 100mm.

SurplusShed duplet objective lens with fl~500mm / dia~70mm : I had high hopes for this lens since its focal length matched the Pro Triplet and its diameter exceeded the S15 Triplet yet it was very inexpensive. However, it proved to be not appropriate for my projector. The first issue was the long lens housing was blocking light from the front fresnel making the testbed projection image darker than expected. This problem was resolved by cutting away much of the lens housing. The second and fatal issue had to do with 'field of view' of the lens. I found that only about the middle 50% of the projected image would be in focus when used with the Pro Fresnels (focalcalc143.exe reported 'Triplet Field Angle ~40°'). I believe this is a limitation of the duplet design; the S15 Triplet did not show this type of issue until experiments where focalcalc reported Triplet Field Angles of greater than 90°. One can reduce the field angle by using a smaller LCD panel or longer focal length duplet/longer projector but these options were not for me. For example, an LCD panel of 7" diagonal produces a triplet field angle of 20°, or a duplet of focal length 1000mm with 16" diagonal LCD produces a triplet field angle of 20°, which I believe would then focus properly.

Note: I read some forum threads describing how to make a duplet from eye glass lenses that sounded like it would be fun to try (see link ) but the SurplusShed duplet cost less than buying the eye glass lenses alone and saved me the effort of building a housing for them - so I went with the surplus.

SurplusShed Delta II BIG lens : I also had high hopes for this lens but it was unsuitable. The throw distance was very short causing the projector to be in the way for all front projection viewing angles. The focal length was so short at app. 250mm that I could not adjust the fresnels sufficiently and always had very dark image edges. There seemed to be some edge-edge focus issues also - I recall reading a thread that suggested this lens may have been made for a curved surface CRT making focus on a flat LCD panel a challenge. I did take the lens apart to find 3 large lenses inside.

I performed an interesting experiment with the minus_300mm fresnel from 3dlens.com and the Delta II lens. With the negative focal length fresnel almost touching the Delta II, I could use a 650mm front fresnel to properly light the entire projection image for a fake 15" LCD which seemed promising. By calculation the minus_300mm fresnel with Delta II gives a combined focal length of 500mm when the lenses centers are separated by 100mm. However, the negative focal length fresnel that I have is less than ideal for this application; the fresnel is made of soft/flexible material and would have to be mounted to plexiglass for support and each projector element the light passes through reduces the overall projected image brightness. Additionally, the negative focal length fresnel corners are greatly rounded reducing the usable lens area. Also, the fresnel center of focus is not at its physical center line but instead is app. 1/3 from the bottom edge further reducing the usable lens area. So I abandoned the idea of using the Delta II lens for my projector.

Beseler 18" from VuLite II Opaque Projector fl~450mm / dia~140mm : This is the lens I eventually decided to use for my projector. It is as large as the Delta II; its large size simplifies fresnel adjustment to get a fairly bright projected image. The focal length was such that could effectively use the Pro Fresnels that I already had when I won the Beseler on ebay. ( Note that I am considering obtaining a 500mm fresnel from 3dlens.com to test if I could achieve better (read brighter) results; I wish I would have ordered the 500mm fresnel at the same time I got the large 330mm fresnel to save the relatively costly shipping from Taiwan. ) The Beseler will not be suitable if I decide to change my projector to rear projection as the room constraints at my swim/tennis club require a shorter throw distance lens. For a future rear projection do-over of my design, I will use the S15 Triplet.

LCD:

Click to view attachment

Rosewill R910J : I bought a 19" LCD monitor with broken backlight from a coworker initially to use in my projector. I was fairly sure that only the backlight was an issue because the monitor would show an image for a second when powered up but then go dark. The Rosewill R910J 19" diagonal 4:3 1280x1024 6-bit/262K color stripped easily but unfortunately had an FFC (flat flexible cable) issue making it undesirable for the projector. Additionally, I had been thinking about how to drive the projector and the R910J only input being VGA implied the need for a dedicated personal computer for movie watching and additional hardware to tune in TV for sports. To allieviate the need for a personal computer, I looked for DVD players with VGA output but did not see anything that looked good to me. I also looked for a standalone TV tuner but the units with good reviews seemed pricey at near the cost of a complete LCD TV.

Fortunately, I was able to repair the broken backlight and make use of the R910J elsewhere. The inverter board had 2 identical circuits on it. One of the circuits had a visibly damaged capacitor but the other circuit was intact. I measured the voltage across the intact capacitor to be app. 80v ac rms. I unsoldered the intact cap and measured its capacitance to be 0.22uF. I obtained a couple of 0.1uF 220v caps from RadioShack and installed them in parallel at the defective cap location. I replaced the other intact cap and the monitor worked !

Click to view attachment

(continued next post)

LCD (cont'd) :

Click to view attachment

Viewsonic N1630W : This LCD TV has many features that are desirable for my projector design and is the LCD panel I used:

• standalone unit so no need for dedicated personal computer at same cost as good standalone TV tuner
• multiple inputs (VGA, HDMI, composite video, component video) allowing for convenient connection to DVD, game platform, PC, cable settop box, etc.
• built in TV Tuner with NTSC/ATSC/QAM capability
• High Def (720p) capable / native resolution 1360 x 768
• Smaller screen size of 15.6" diagonal allows use with S15 Triplet, potentially smaller projector size than a 17" or 19" LCD panel

I found a N1630W on sale at a local Office Depot store for under $200. I see it at many online retailers at about that price.

(continued next post)

LCD (cont'd) :

Winner !!!

Click to view attachment

LCD Strip/Disassembly:

The Viewsonic N1630W was straightforward to strip/disassemble. See the list of steps below. I followed some advise I found in the forums that suggested taking the LCD apart, measuring all parts, make cardboard models of the parts, put LCD back together then set aside until needed for final assembly. The idea is to keep the LCD clean and prevent accidental breakage. After the initial strip and reassembly, I completed almost all of my case construction before I touched the LCD again.


Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

Click to view attachment

(continued next post)

LCD Strip/Disassembly (cont'd):

Click to view attachment

Disassembly of Viewsonic N1630w LCD TV

• 15.6" diag, 1366x768 native resolution
• 1280x720 mode for true 720p, 1280x720 image size 14.7" diag
• contrast 1000:1 (dynamic 10,000:1)
• built in tuners ntsc/atsc/qam
• inputs: hdmi, component, vga, s-video, composite (rca), coax rf
  stereo audio (rca), 3.5mm pc audio
• audio output: 3.5mm headphone (can connect to external speakers)

Steps:

1. remove stand per user manual
   a - remove 4 philips head screws from bottom of base
   b - remove base foot
   c - remove 4 philips head screws from base neck
   d - remove base neck
2. remove 6 philips head screws from back
   a - 2 machine thread into hdmi connector
   b - 2 small from component connector
   c - 2 large near bottom rear housing
3. Separate rear housing
   a - find snap tab release slots at bottom on right and left
   b - inserting flat screw driver will pop 2 of 5 bottom snaps
   c - slowly work around edge of housing prying front away from back to clear all snaps, 5 on bottom, 4 on right, top, left
4. set rear housing aside
5. remove right-rear audio connector plastic cover using small flat screw driver in cover snaps to expose 1 philips screw
6. peal metalized tape from video connector metal cover
   remove 2 vga connector post/screws
   remove 4 philips hear screws from rear metal frame
   set metal cover aside
7. remove 3 philips head screws holding button circuit board to front housing
8. pry speakers (right and left) off retaining posts
9. release front housing snaps holding lcd frame to separate front housing
   3 snaps at right, 3 snaps at left, 2 bigger snaps at bottom corners
   set front housing aside
10. remove rear metal housing
    a - peal metal tape between metal housing and lcd panel (3 pieces at top, 3 pieces at bottom)
    b - unplug backlight cables top/bottom left side by pressing down on plug latching mechanism with small screw driver while pulling
    c - remove 4 philips head screws (2 left, 2 right) from metal housing sides
    d - remove 1 philips head screw at component video connector
    e - unplug cable from controller to lcd panel at top left of metal frame by squeezing cable connector small metal clips and gently wiggling free the cable plugs in with gold fingers away from pcb/connector metal plate toward pcb
    f - remove / set aside metal housing containing combination power supply/inverter circuit board, controller circuit board, button/IR receiver circuit board
11. remove metal housing around lcd
    a - remove 4 microscopic philips head screws rear four corners
    b - lift plastic covering flex circuits
    c - remove 3 very small philips head screws
    d - release metal snaps (4 each top/bottom, 2 each right left)
    e - separate front and rear metal and set front aside
12. remove lcd from rear metal frame that holds the backlight

Beseler Vulite II :

Click to view attachment

Many references are made in the forums to successful projector builds using the lens from a Beseler Opaque Projector. I was fortunate to win an ebay auction for a Beseler VuLite II with a very low bid of $17. (The ebay seller misspelled the item as 'bessler' and there were no other bids.) Shipping was a app. $22; so for around $40 I obtained a very nice lens. Spec's are focal length 457mm, diameter of app. 140mm, length 160mm, weight 15 lbs (glass lenses are heavy). Additionally, the beseler was constructed of fairly heavy gauge aluminum (1.5mm thick) which I was able to use to fabricate custom metal parts for my DIY projector. This gauge aluminum is a dream to work with - it cuts easily with a jig saw or hack saw, files quickly with a metal file, and even sands smooth with 100 grit sandpaper. I priced aluminum sheets at local home improvement stores but could not come close to $40 for the amount of metal in the VuLite and the home store gauge was much lighter. It was also quite fun disassembling the Beseler VuLite II to see how it was designed and how it worked - a nice piece of equipment. All in all, this was of great use for the DIY projector build and I would go this route again when building another projector.

The Beseler's 1000w halogen light was not acceptable to me for the DIY projector due to high heat output and poor color reproduction in the projected image as compared to the Metal Halide bulbs more commonly used in Lumenlab style projectors. See this forum post which includes a 'best bulb chart' and explanation of CRI (Color Rendering Index).

However, I did re-use the VuLite raise/lower mechanism (including external lever) as a lens focus device for my DIY projector. The focus mechanism works very well providing lens travel of app 5" (125mm) and allows for quick and accurate image focusing.

I also re-used the lens clamp and power cord (for wiring). Left over parts that I may use in future builds include a nice front surface mirror, a piece of heat resistant/tempered glass, the original VuLite lens focus mechanism, and additional aluminum stock.

Click to view attachment

Click to view attachment

Projector Light Source Choice :

Click to view attachment

The biggest advantage a DIY Projector has over commercial projectors is long lamp life. The Sanyo Z3 projector in my home theater is a great unit, but its lamp life is rated at 3000 hours (4000 hours economy mode) which is typical of its category. So if one were to watch the projector for everyday TV use, say 6 hours a day, then the bulb would need replaced in 500 days or about 1.5 years. Replacement bulb for the Sanyo Z3 costs app. $300 last time I checked. Contrast that with the typical Metal Halide bulb used in DIY projectors with life of 20,000 hours or 10 years at the same 6 hours/day usage. Cost of the bulb (and ballast to drive it) is also much less than the typical commercial projector bulb.

After reading the forums and consulting the 'best bulb chart' I decided to go with a Ushio S400DD bulb which I obtained from atlantalightbulbs.com from their retail store in Tucker, Georgia (driving distance for me). At the time of purchase, I got the bulb, mogul base, and ballast for $120.

I have been happy with my choice thus far as the projected image looks quite good to me. However, there are a couple of negative points to this lighting solution:

• weight - the ballast weighs almost 30 lbs . The kit came with strong brackets that I attached to my case with a dozen screws. Even so, if the projector is
  ever dropped I am sure the screws won't hold in the MDF/plywood case. Additionally my final projector configuration is quite heavy and the ballast is the single heaviest item in the build.
• size - the ballast is quite large and I had to dedicate a large space in the case to locate the ballast. This impacted my case design contributing to the
  relatively long design.
• heat - the ballast generates a significant amount of heat. I added a special duct to direct air flow over the ballast in the final projector configuration
  which keeps the ballast at a reasonable temperature of app 100° F. Before the duct was added, I measure the ballast temperature at 195° F which is too hot.
• noise - the ballast emits an audible buzz while operating. The sound is not objectionable but is noticeable if one listens for it. For the final projector configuration I actually adjusted the fan rotation speeds so that the noise of the fans was just quieter than the ballast buzz. The fan noise helps mask the ballast noise, and the overall sound from the DIY projector is not a problem for my application.

Case Design:

Click to view attachment

I wanted the DIY Projector enclosure to be more elaborate than a plain box but I was not inclined to be too complex for my first build. I settled on the idea of rounding the corners of the case. I used Google SketchIt to model a few concepts (1", 2", 3" radii) and liked the larger rounds. Lowes home improvement store had 6" (150mm) diameter black PVC pipe in 24" (600mm) lengths that I then cut into fourths using a table saw and wooden jig. The jig securely held the pipe square to the table saw's table and fence for straight cuts.

Click to view attachment

At the same time I started calculating the size the case would need to be:

• width=475mm was chosen to be the width of the Pro Fresnels plus 1/4" (6mm) gap on each side for wood to hold the Fresnels. This saved me having to cut the fresnels width-wise.
• height=355mm was chosen to be the height of the LCD panel plus the 2" (50mm) for the LCD controller plus 1" (25mm) air gap plus 1-1/2" (40mm) slop for adjustments.
• length=1060mm was the most complex value to choose. I wanted to be able to focus a screen from 50" to 250" diagonal. focalcalc143.exe provided values for distance between the middle of the projection lens to the LCD, with the difference being app 5" (125mm). I planned to fix the position of the LCD and front and rear fresnels. So the projection lens focus mechanism would need to provide the 5" (125mm) travel. I also wanted the projection lens to completely retract within the case for storage. And I needed roughly 9" (225mm) for the massive ballast to be placed behind the projector lamp. Thus the final length was ballast space plus 220mm for rear fresnel plus 20mm fresnel-LCD gap plus focalcalc143.exe minimum LCD-lens distance plus half the Beseler lens length (85mm).

...about 15x the volume of my commercial Sanyo Z3 projector.

Case Construction:

Click to view attachment

I decided to build most of the case out of 1/2" Medium Density Fiberboard (MDF). Although I have a table saw, I made most of the MDF cuts using a hand held rotary saw held against a metal straight edge clamped to the wood. The rotary saw leaves a somewhat rough edge; so I would leave the piece a little long and then use my router with straight-cut carbide laminate bit (also held against the metal straight edge) to finish the cuts. The routing leaves a very nice/clean edge.

For the large rounded corners of the case, I made a pattern out of a small MDF piece. I traced the curve using one of the PVC pipe quarter pieces that would eventually be the rounded edge. I rough cut the pattern with a jig/sabre saw and then sanded it to a smooth round. The pattern was then clamped to the back side of the piece to be shaped and I used the router with the ball-bearing of the straight-cut bit riding against the pattern. I used the pattern for the 8 rounds of the main case plus 4 rounds on the shelf that I added to the rollable base later.

Click to view attachment

(continued next post)

Case Construction (cont'd):

The rear end piece received holes for three 120mm 12v dc fans. The fans would be directed to blow air out of the case. The front end piece received a centered hole for the Beseler projection lens, plus a hole to either side of the lens where cool air would enter the case. Based on a suggestion read in the forums, simple air filters were built from small square wooden frames with pantyhose stretched over each frame.

Click to view attachment

The holes in the rear and front pieces was cut using the drill press and a hole and wheel cutter accessory. The sides, ends, and bottom of the case were assembled together with biscuit joints and wood glue. Two hardwood cross pieces were attached with screws through the case side walls. These cross pieces have slots to hold the case top. They also have slots that hold the quartered PVC pipe pieces.

Click to view attachment

(continued next post)

Case Construction (cont'd):

As nice as the routed MDF edges were, they still did not present a suitable finished edge. I decided to finish the cut MDF edges with 1/4" x 3/4" pine trim from the Home Depot home improvement store. I read up on steam bending wood on the internet and after several tries was able to form the trim around the case edges and rounded edge corners. My wood steam setup is shown below: 1) tea kettle, 2) on my gas grill, 3) steam feeding a PVC pipe, 4) held on a step ladder with bungee cords. I would steam the trim for 20 minutes, then quickly clamp around the case edge. After letting it cool and dry for a hour, I would glue and nail the trim in place. I also bent wood trim around the edge of the shelf mounted on the rolling base.

Click to view attachment

Click to view attachment

Case Construction (cont'd):

The interior of the case was painted with Rustolium BBQ Grill High Temperature Flat Black spray paint. This paint worked quite nicely on the metal parts. And it adhered well to the hardwood pieces. However, it did not work so well on the MDF and comes off when touched. This is not a huge problem for the inside of the case, but I wish I would have used the BBQ grill paint only in the high temp areas (e.g., "light box") of the case and had used a different black paint elsewhere.

The exterior of the case was painted with exterior semi-gloss white paint on most surfaces. The America's Best brand paint from Home Depot covered well and looks great on the MDF. I did paint the exterior ends using the BBQ Grill paint and had the same problem with the black paint coming off when touched. To resolve this issue, I put three light coats of clear polyurethane on the black exterior parts using a foam brush. This changed the black from flat to glossy and also sealed the surface.

Some MDF edges (e.g. holes cut in the case) could not be finished easily with wood trim. I found some black felt with adhesive backing from Wal*Mart worked very well to clean up the MDF edge and hide any unevenness in the hole cuts.

Click to view attachment

Click to view attachment

Lens Focus Mechanism:

Click to view attachment

The projection lens is held in a wooden frame. The frame is mounted to the sides of the case on drawer slides purchased from Home Depot. The slides I bought were too long, so I disassembled them, cut each section in half, then reassembled them to make two 7" slides from one 14" slide.

Click to view attachment

The raise/lower mechanism from a Beseler Opaque Projector was modified and used as the focus mechanism for the DIY Projector. The mechanism works by twisting the actuating rod which turns the double-wheel against the L-shaped pipe, moving the pipe forward or backward. The actuating rod extends through the bottom of the case where an aluminum lever is attached. Turning the lever moves the Beseler lens in or out. From left to right, the picture above shows the lens fully retracted, half retracted, and fully extended. The Beseler lens clamp (blue metal ring) is also visible in the picture. The clamp is pinned to the wooden focus slide with two finish nails. The clamp securely holds the Beseler lens as it slides through the center hole in the front piece of the case.

Lamp Position Adjust Mechanism:

Click to view attachment

The mechanism for adjusting/fine tuning the DIY Projector lamp position was inspired by a SketchIt design posted by fellow Lumenlabber WindCalmer. (In SketchIt, perform File->3D Warehouse->Get Models then search for 'WindCalmer'. Other interesting SketchIt models exist; try searching for 'lumenlab' or 'projector' or 'beamer'.) I built my bulb adjuster from scrap aluminum from a Beseler Opaque Projector housing and 1/4" bolts and nuts. Using high temperature epoxy from Home Depot (J.B. Weld), four nuts were attached to the smaller top plate. Bolts are inserted through the epoxied nuts, and these then capture the bottom plate with 4 additional nuts per bolt. The bottom plate nuts spin freely with each bolt. This arrangement allows the top plate to be moved closer to or farther from the bottom plate in a controlled manner. The top plate can also be tilted by rotating some of the bolts more than the others. Two hardwood pieces were mounted to the DIY Projector case side wall and two metal plates are mounted to the hardwood pieces. The bulb adjuster slides forward or backward in slots on the hardwood pieces for additional lamp position adjustment. Four small nuts were also epoxied to the bottom plate to hold screws used to lock the bulb adjuster in position.

Click to view attachment

I later replaced the rear bolts with much longer versions so that the new longer bolts would hold the Pro Reflector bracket used on the DIY Projector. The Pro Reflector bracket was copied from work by another Lumenlab forum contributor who used a ThermalTake 80mm fan guard with the center cut free. The picture below is from his post / I apologize for not being able to give proper credit. After cutting out the center, the cuts were filed then sanded smooth. Large flat washers and #8 bolts and nuts then capture the Pro Reflector to the fan guard. (The reflector is not rigidly captured but instead has app. 2mm of movement in all directions to prevent cracking.) Custom brackets then capture the reflector holder to the two long bolts of the bulb adjuster.

Click to view attachment

Click to view attachment

Light Box:

A separate "light box" compartment was built at the rear of the projector to house the projector lamp and ballast. The front wall of the light box provides a convenient mount for the precondensor lens. Also since the air flow is controlled through the opening in the front wall, the air flow is directed across the precondensor cooling it and helping to prevent lens cracking. This front wall also removed the need for a lexan or tempered glass heat shield to protect the rear fresnel; and since some light would be lost passing through a heat shield the benefit is a brighter projected image.

Click to view attachment

The precondensor lens is mounted to an aluminum plate and captured by four 1/4" bolts, nuts, and large washers. The lens sits on the flat edges of 1/4" nuts. The lens is not captured tightly but has app. 3mm of movement in in all directions. The precon plate mounts to slots in the light box metal front wall with 1/4" x 4" bolts. The plate can be moved by adjusting the bolts providing front to back precon adjustment. The slots in the metal front allow adjustment of the precon from side to side; the light box metal front is screwed through more slots to hardwood supports. The support slots in the metal front allow for top to bottom adjustment of the precon.

The metal for the light box front came from a square metal duct purchased at Home Depot. The duct metal is not as nice to work with as the thick scrap aluminum from the Beseler Opaque projector but the duct metal was inexpensive and straight and works well for fabricating metal parts that don't have to carry much weight.

Click to view attachment

A MDF piece forms the lid for the light box. A large MDF piece forms the top of the case. The top sits in grooves cut in hardwood cross piece supports at each end, and is supported by the rear fresnel frame in the middle of the case. The top is held in place by four magnetic cabinet door latches (see latch in light box picture above next to the precon mount picture, brown plastic item mounted on case side wall). To make the case top easier to remove, a handle was made from hardwood and attached with screws.

Excessive light was leaking through the fans. A metal piece was fabricated to resolve the issue. Added to this metal piece was a heat shield and air duct designed to keep the ballast cool. In the picture above, see the ballast mount location in the 3rd of 4th picture, then see the light/heat shield in place (before being painted flat black).

The light box lid was found to get too hot when the projector was operated. A metal duct was fabricated and attached to the lid with screws. This duct directs air along the top of the light box between the metal and MDF which keeps the MDF substantially cooler during projector operation. Holes were added to the hardwood cross pieces that form the light box front wall for the additional ducts.

Click to view attachment

Controller Design:

Click to view attachment

The controller provides these primary functions:

• Turns the projector lamp on when the LCD TV turns on; Turns the projector lamp off when the LCD TV turns off
• Monitors the cooling fans and turns the projector lamp off if the fans stop working properly (error condition)
• After the LCD TV turns off, the controller runs the fans for 15 minutes to fully cool down the unit
• Monitors the temperature of a probe and shuts down the lamp of the temperature exceeds an adjustable value (error condition)
• Runs nine status LEDs (Power On, TV On, Fans On, Fan1 OK, Fan2 OK, Fan3 OK, Cool Down, Lamp On, Overheated) reporting current state of the projector
• Provides electrical isolation ( ESD protection ) between an external switch box and the LCD TV circuitry

The controller is powered by 12v and requires app. 1250mA maximum (350mA fan1, 350mA fan2, 350mA fan3, 100mA relay, 9 LEDs x 5mA each, 5omA misc.). All connections between the LCD TV and the controller are at the LCD TV button board. All connections between the LCD TV and the controller are via optical isolation to both protect the LCD TV from static discharge and prevent unwanted ground loop issues.

(continued next post)

Controller Design (cont'd):

The picture shows the controller placed in the front of the projector near the projection lens. The lamp relay is in front of the controller. The wiring block to the right of the controller board connects the various AC items (listed left to right - Ballast Ground, Ballast Hot, Ballast Neutral, Relay In Hot, Relay In Neutral, Controller supply Hot, Controller supply Neutral, In Hot, In Neutral, In Ground).

Notice the bracket that holds the controller power supply bricks in place.

The controller and LCD electronics are toward the front of the projector where it is cooler.

A 3/16" deep groove is cut in the case wall for routing wires to the rear of the projector.

The ballast and fans are mounted at the projector rear.

Click to view attachment

Controller Schematic (3 pages)

The controller circuits can be broken down into several functional areas:
1) TV On detection (page 1)
2) Fan Control and Fan Rotation Detection (page 1)
3) Temperature Detector (page 1)
4) Lamp Relay Control and Logic (page 2)
5) Cool Down Timer Circuit (page 2)
6) Remote Switch Box Isolation Circuit (page 3)

Click to view attachment

Theory of operation: (page 1)

1) TV On - The LCD TV turns on a blue LED on the TV button board when it powers on. This LED was disconnected, and the LED of the six pin opto-isolator IC is wired in its place. Thus, the LCD TV controls the on/off status of the six pin opto-isolator. The LCD TV was found to not drive the opto-isolator LED sufficiently, so a second external NPN transistor was added to give more gain. The output of this circuit is the signal TV_ON. A Green status LED reflects the state of TV_ON.

2) The FANS_ENABLED signal is the logical OR of several signals including TV_ON, LAMP_ON, OVERHEATED, COOL_DOWN_TIMER. If any of these signals are asserted then the fans should be on. A Green status LED reflects the state of FANS_ENABLED.

FANS_ENABLED drives a NPN transistor in each of the three fan control circuits. When the NPN is on the fan ground is connected to controller ground and the fan is powered and spins.

Each fan has its own rotation detection circuit. Pin3 of each fan will have a square wave from 0v to 12v present when the fan is spinning. If the fan is powered but stops spinning, then Pin3 will go to either 0v or 12v. The rotation detection circuit works by low pass filtering the square wave ( 47k, 1uF ) which produces a voltage between 2v and 10v when the square wave is present. This voltage allows both the NPN and PNP to power on which lights the FAN_OK green LED and a LED in the sixteen pin opto-isolator IC. If the square wave is not present then either the NPN or the PNP will power off and the green LED and opto-isolator LED will be off. Note: To increase the detection voltage of the NPN transistor a LED is placed in the NPN transistor base path. This LED does not visibly light, but it does act as a 2v zener diode in this configuration.

3) The temperature sensor circuit works by monitoring the resistance of a Positive Temperature Coefficient (PTC) thermal-resistor (thermistor) and changes the circuit output status when the thermistor resistance exceeds the resistance of an adjustable resistor (potentiometer). The circuit is designed such that it reports OVERHEATED if the thermistor is unplugged or fails in the open circuit configuration (the more likely failure case). The potentiometer was adjusted during DIY Projector construction for OVERHEATED to occur at 110° F. A red status LED reflects the state of signal OVERHEATED.

(continued next post)

Controller Design (cont'd):

Click to view attachment

Theory of operation: (page 2)

4) Lamp Relay Control and Logic - The sixteen pin opto-isolator forms a logical AND of signals TV_ON, FAN1_OK, FAN2_OK, FAN3_OK, TEMP_OK (opposite of OVERHEATED). Only if all these signals are asserted will the projector lamp be turned on. Note that to prevent glitches on the lamp relay, a delay ( 220k, 100uF ) circuit of app. 20 seconds is employed. An amber status LED reflects the status of the NPN transistor the drives the lamp relay.

5) Cool Down Timer Circuit - The cool down timer circuit keeps the fans spinning for 15 minutes after the projector lamp turns off. When the lamp turns on, the 100uF capacitor is charged almost instantly through a PNP transistor to 12v. When the lamp turns off, the capacitor starts to discharge through the 4.7 Meg Ohm resistor such that the voltage on opamp pin5 begins to drop from 12v to 0v. Two resistors set a reference voltage on opamp pin6 of app. 2v. When the voltage on pin5 drops below the voltage on pin6 the state of the opamp and signal COOL_DOWN_TIMER changes. An amber status LED reflects the state of signal COOL_DOWN_TIMER.

(continued next post)

Controller Design (cont'd):

Click to view attachment

Theory of operation: (page 3)

6) Remote Switch Box Isolation Circuit - The DIY Projector has jacks for a remotely mounted switch box. The switch box provides the same controls as the LCD TV button board: Power on/off, Channel Plus, Channel Minus, Volume Plus, Volume Minus, Input, Menu. The idea was that for a rear projection setup where the projector would be located behind a wall and not be user accessible, someone operating the projector would use the remote switch box to control the rear projection TV functions. It was decided to use CAT5 cable since it is commonly available and to keep cable costs down. To provide electrical isolation and protect the LCD TV circuits from static discharge, opto-isolator ICs are used. The switches of the Remote Switch Box control the LEDs of the opto-isolators; the transistors of the opto-isolators then active the switches on the LCD TV button board.

It occurred to me that being able to control a rear projection setup with the IR Remote control would be a nice feature. I found an interesting web write up on an IR Extender circuit and implemented it inside the Remote Switch Box. To support the IR Extender feature with cheap cabling, I used telephone wire for the additional signals needed.

The internal cable connection between the controller and the LCD TV is implemented with an old personal computer motherboard-to-floppy disk cable. The cable quickly and easily plugs into the controller making projector disassembly straightforward.

(continued next post)

Controller Design (cont'd):

Remote Switch Box Schematic

Click to view attachment

(continued next post)

Controller Design (cont'd):

Click to view attachment

Circuit Construction

Powerpoint was used to make layout/wiring diagrams. All pins were made to align to a 0.1" grid in Powerpoint. The circuits were then built on fr4 perf board which was predrilled with a 0.1" hole grid. The layout/wiring diagrams were printed and taped to the perf board and the electrical components placed right on top of the diagrams. The paper of the diagrams is a good enough insulator so as to not affect the circuit operation. And all soldering was done on the rear / non-diagram side of the circuit boards so as to not burn the paper diagram.

I bought most electrical parts from online retailer www.digikey.com or local RadioShack stores.

In the PCB layouts below, the red, green, black traces are wire runs on the back of the board. The silver traces are top/component side jumpers.

(continued next post)

Controller Design (cont'd):

Controller Layout

Click to view attachment

Remote Layout

Click to view attachment

Bills Of Material

Click to view attachment

Click to view attachment

Fresnel Mounting:

Click to view attachment

My DIY projector uses a 'split design', that is, arranged as rear fresnel, LCD panel, front fresnel.

The rear fresnel sits in slots cut into two pieces of hardwood that are then screwed to the sides of the case. See picture above. Not shown in the picture is a third slotted hardwood piece that acts as a cap. Small blocks at the bottom of the side slots hold the rear fresnel to create an air gap of app. 40mm.

The front fresnel is glued into a slot cut in a round wooden dowel which sits in holes drilled in small hardwood blocks that are screw mounted to the sides of the case. (The strings in the picture were just holding the front fresnel in the slot until the RTV glue dried.). The front fresnel remains vertical even when the projector is tipped upward or tipped downward, thus providing gravity keystone correction.

The width of the case was chosen so that only two cuts were needed to size the rear fresnel. After measuring several times to assure the fresnel optical center remained centered with the projection lens and lamp, I cut a section from the top and a section from the bottom of the fresnel. From a Lumenlab forum suggestion, I used a home laser level to align centers (the level can make both vertical and horizontal level lines). The laser level was aimed through the projection lens at the front of the case to the center of the middle fan at the rear of the case. I also used the laser level to adjust the left to right and top to bottom placement of the lamp.

It worked out that I only had to make two cuts on the front fresnel also (although this was more luck than planning). I rotated the 17" x 16" fresnel by 90° to be 16" x 17" to achieve a width that worked with the dowel and blocks. After measuring several times to assure the fresnel optical center remained centered with the rear fresnel, I cut a section from the top and bottom of the fresnel. The front fresnel is app. 30mm wider and 30mm taller than the LCD viewable area, and measures 406mm x 225mm.

To cut the fresnel, I first wrapped the lenses in two layers of plastic food wrap (Lumenlab forum idea). The fresnel to be cut was then lightly clamped to a work bench using a metal straight edge. A sharp utility knife was used to cut the plastic fresnels. Using the straight edge as a guide, I made 60 progressively deeper cuts, 30 cuts right to left and 30 cuts left to right. I then was able to snap the fresnel into two pieces along the cut. On one cut the snap was not a clean break ( I tried to cheat by doing less than the 60 cuts ). So I completed the cut using a razor saw.

Because the fresnels are held by wood that is screwed into the case walls, there is a fair amount of adjustability provided. Additional adjustability was provided in the front fresnel by drilling the holes for the dowel off-center in the wood blocks. By rotating the blocks 90&#176 at a time, the front fresnel can be 20mm, 40mm, or 60mm away from the LCD. I ended up using 60mm to provide the greatest mount of keystone correction.

Lumenlab forum lists maximum practical keystone correction to be +/-15° when tilting the front fresnel. Since I wanted the maximum keystone correction, and my fresnel was 225mm tall, I calculated the LCD-to-fresnel spacing needed to be 60mm.

Click to view attachment

LCD Sled:

Click to view attachment

The LCD panel and LCD electronics are held in a "sled" built from MDF and hardwood pieces. The sled is screwed to the bottom of the DIY Projector case. The LCD glass sits in the metal frame from the LCD TV and is captured by the two vertical hardwood pieces in the picture above. There are 1/8" grooves in the wood pieces and finish nails are inserted through the metal frame into the wood to hold the frame in place. The LCD TV electronics were composed of two PCB's (printed circuit boards; controller and power) mounted in a metal frame. I decided to use the LCD TV metal frame unmodified for convenience and since it probably provides shielding for the electronics. The controller/power board metal frame lays horizontally in the sled and is attached with screws. A short cable runs from the controller PCB to the LCD panel driver PCB.

There is a third LCD TV PCB that contains 7 push button switches, the IR remote control receiver, and power on/off LEDs that I refer to as the "button board". A short cable runs from the controller PCB to the button board. The button board was mounted to the front plastic housing of the original LCD TV. I was able to re-use the section of plastic from the LCD TV housing that held the bottom board as seen in the picture below. An access panel was fabricated out of scrap aluminum from a Beseler Opaque Projector. The access panel contains the AC Power Entry Module, a 7 Amp circuit breaker, mounts for 6 status LEDs from the custom controller, jacks for the remote switch box (RJ45 and RJ11), and the LCD TV button assembly. A hole cut in the DIY Projector case bottom gives access to the LCD TV inputs for VGA, HDMI, component video, and composite video in addition to the access panel I fabricated.

(continued next post)

LCD Sled (cont'd):

Click to view attachment

The LCD TV connections for RF and audio (headphones out) were on its side panel. I made custom cables with panel mount connectors to route the RF and audio signals to holes drilled next to the vga/HDMI connections.

Click to view attachment

A ground loop problem caused the audio out to work intermittently with external powered speakers. To resolve the problem, I built a small PCB with two audio isolation transformers (right / left channels).

Click to view attachment

Heat Management:

Click to view attachment

My DIY Projector uses three 120mm personal computer case fans to exhaust hot air from the projector. From experiments with cooling personal computers, I knew that large fans spinning relatively slowly provide good air flow without producing excessive noise. I adjusted the fan speed using series resistors plugged between the fan connector and fan power source. I tried a few resistor values before settling on 33 ohm / 1100 r.p.m.

Air flows in through two 120mm diameter round vent holes at the front of the case. The air passes through shear air filters made of pantyhose stretched over small wooden frames. The air filters can be removed for cleaning; but it is easier to just blow them off occasionally with a can of compressed air from inside the case. The LCD panel and LCD sled form a wall that directs the air flow to a 2" (50mm) air gap between the top of the LCD panel and the projector case top. The rear fresnel and LCD panel form a channel through which the air flows cooling the panel and fresnel. Air then passes through the 40mm gap between the rear fresnel and projector case bottom. The air flow then splits into three parts. Most of the air continues to the "light box" passing over the pre-condensor lens, lamp, and reflector. Some air is directed by metal pieces that form an air duct guiding air flow along the "light box" bottom and then across the lamp ballast. Finally, some air is directed by a metal air duct that guides air flow across the top "light box" wooden panel. All hot air then flows to the fans where it is exhausted from the case.

Using a touchless temperature meter, I have measured various internal case temperatures and am satisfied with the current arrangement:

• Before installing the lower duct, the ballast temperature was quite high at 195° F but with the duct it has not exceeded 100° F.
• The LCD panel temperature after several hours of viewing was app. 85° F or 15° F above room temperature.
• The hottest location in the current design is the top of the "light box" which is made of MDF. Maximum temperature with the duct is 135&#176 F - a reasonable value (an online MSDS (material safety data sheet) lists MDF ignition temperature at greater than 425° F). Without the duct, the wood heated to app. 160° F and one could smell the MDF "baking".

The temperature probe on the custom controller has been set to shut off the projector lamp if the temperature exceeds 110° F. The controller will continue to spin the fans until the over-temperature condition is cleared. Hysteresis in the temperature sense circuit requires the temperature to drop below 95° F to clear the overheat condition. The controller also runs the fans for 15 minutes after the projector lamp is switched off to assure the unit has cooled down.

Click to view attachment

Click to view attachment

Click to view attachment

Testing and Tweaking:

Click to view attachment

My DIY Projector had the potential for outdoor use (at the club) and brightness is often discussed in the forums as a weakness of this type of projector design. So I wanted to optimize for brightness wherever possible. The most common adjustment was just moving the projector lamp holder assembly forward or backward and then also adjusting the position of the reflector and/or precondensor lens. I found the gas welder glasses pictured to be useful when working on the projector when the lamp was on.


I also have my Sanyo Z3 home theater projector to compare against (see it hanging from ceiling at top right of middle picture).

For accurate measurements, a light/lux meter was used.

I used PGEN.EXE, a test pattern generation program from the Lumenlab forums, with pattern "solid color" selected and the lux meter to make brightness measurements.

The Sanyo native resolution is the same as 720p high-definition (1280 x 720), so all measurements were made with this screen setting.

I projected against a flat white painted wall. I adjusted the Sanyo to produce the same picture size (84" x 47.25", diag. 96.4") as the DIY Projector.

Click to view attachment

I measured the Sanyo Z3 as a benchmark. I then measured the DIY projector first without reflector or precondensor lens, then with reflector but no precondensor, then with reflector and with precondensor.

Click to view attachment

First some comments on the Sanyo Z3. The Z3 has a very nice lens shift feature allowing up to 50% left/right and 50% up/down shifting. I take advantage of the Z3 lens shift in my home theater; the Z3 effectively is located perpendicular to the upper left corner of the projected image. Using lens shift does degrade the evenness of the projected image and I believe the Z3 would have shown virtually no vignetting in the values in the table above if I were not using lens shift.

I was initially somewhat disappointed in the DIY brightness values as compared to the Z3 (essentially, 1/5th center and 1/10th edge). I could achieve brighter screen edges and corners but only at the expense of a less bright image overall. It was encouraging to see all measured locations increase roughly 33% when the reflector was added. It was even more encouraging when the precondensor was added as it increased all values but the gains were more at the screen edges.

My disappointment faded after watching a few hours of TV and movies, and I find the DIY Projected very watchable with the brightness much less of a concern to me. The final comparison to Z3 about shows app. 1/3th center and 1/5th edge brightness.

In fact, I find the DIY Projector to be a better viewing experience when watching standard definition TV than what my old home theater setup provides. This is not a fault of the Sanyo Z3; the Sanyo receives its signals from a Comcast/Motorola settop box. This particular settop box produces great High Definition content but is routinely criticized for its handling of standard definition sources. Inside my DIY projector, the Viewsonic N1630W handling of standard definition signals is superior to the settop box.

Click to view attachment

Click to view attachment

(continued next post)

Testing and Tweaking (cont'd):

Click to view attachment

Click to view attachment

I did try the DIY Projector at the club under overcast sky but the image was just too dim to be watched during the daylight hours. Even with the DIY Projector moved as close to the screen as could be focused (diag. app . 55") it was still just too dim. However, later that evening and just after dusk the DIY projector image looked really really good. I was able to tune in a High Definition TV channel of some documentary on Public Television. The image looked spectacular even when the projector was pulled back to the full 110" diagonal size of the screen. So club movie night and Monday Night Football will be perfect; Saturday afternoon sports watching will not be practical.

Misc. (case rolling base, LCD AntiGlare, DIY screen, cost roll-up):

Case Rolling Base

Click to view attachment

The projector was mounted on two 1" pipe legs. The threads of the pipe into the 90° pipe elbows are stiff enough to hold the projector angle when tilted. Two hardwood blocks (that mimic the case shape and rounded corners) are attached via several screws from the inside of the case walls. The base is made from 2" x 4" lumber and MDF/plywood. A set of four furniture casters are mounted on the base allowing the projector to roll on smooth surfaces. To allow for easier transport, the projector can be disconnected easily from the base. The legs separate in two after removing a locking pin. A 2" x 4" block was screwed to the bottom of the case to protect the focus handle when the projector is set on a flat surface.

Click to view attachment

To accommodate a DVD Player and the Personal Computer speakers used with the projector, a small shelf was built on top of the base. The shelf also has two compartments for extension cords. The speakers are secured to the shelf with velcro; the speakers can be removed and placed up to 25 feet in front of the projector to achieve proper stereo sound. The shelf is placed so as to prevent the projector from being over-rotated on its mounts. (See picture above.)

(continued next post)

Misc. (case rolling base, LCD AntiGlare, DIY screen, cost roll-up) (cont'd):

LCD AntiGlare (AG)

Click to view attachment

The LCD panel has a layer of antiglare (AG) film that negatively affects the brightness of the projector. The AG film scatters light hitting the front of the panel which is exactly the way the DIY Projector is configured. Much discussion of the AG film can be found in the Lumenlab forums.

Based on comments in the projector log of RCDavid, I used the "polish method" with Mothers Mag & Aluminum Polish. I measured the light from a fluorescent lamp at a fixed distance with a lux meter. Then I measured the light passing through the unpowered panel before applying any polish with the light and meter at the same spacing. I measured again after 6 coats and again after 12 coats of polish were applied. The amount of light passing through the panel initially was quite small, at app. 6.5%. I was very happy to see a 50% increase in the panel transmisivity after 6 coats, to app. 10% light passed. However, I saw no improvement after 6 more coats for 12 total coats.

It is my belief that the polish is not actually polishing the AG film. Instead the polish fills in the uneven AG film surface making it smooth so that more light passes through the panel and less light is scattered. I could see a slight oily sheen on the panel after polishing. I am happy with the results but do have the concern that over time the polish may dry out and the improved transmisivity will be lost. I also believe that the improvement could have been achieved after 1 or 2 coats instead of 6, and that additional coats beyond that do not provide benefit.


DIY Screen

Click to view attachment

A DIY projection screen was made from blackout cloth, a 3" PVC pipe, and wood trim. (The "wheels" pictured at the end of the PVC pipe turned out to be unnecessary; the wheels were made from MDF and were cut using a hole/wheel cutter and drill press.) Two pre-made wooden shelf brackets are used to mount the screen to a wall. Holes were drilled in the top of the brackets to receive 1/2" dowels. Holes 9/16" diameter were drilled through the PVC pipe app. 1" from each end. The pipe sits on the brackets with the dowels through the PVC pipe holes when being used or rolled up for storage. To raise or lower the screen, a person at each end lifts the pipe off the dowels and twists the pipe to raise or lower the screen; the pipe is replaced on the dowels to complete the operation.

Screen construction was straightforward. The PVC pipe was first hung from two coat hangers. The blackout cloth was attached using masking tape to the pipe and adjusted until level and wrinkle free. A piece of 1/4" x 3/4" x 8' wood trim was held against the pipe and cloth and screws driven through the trim and cloth into the PVC pipe to secure the cloth in place. Two additional pieces of wood trim were glued with adhesive caulk to either side of the screen bottom edge to add weight and hold the cloth straight when in use.

The pipe, 3 trim pieces, dowels, and shelf brackets were obtained from Home Depot. The 54" wide blackout cloth was obtained from JoAnn Fabrics, a local craft store. The screen size is 96" x 54" for diagonal 110". The pipe extends past the screen app. 4" at both ends.

Cost roll-up:

• ViewSonic N1630W LCD High Definition TV - $200
• Light engine (bulb, ballast) - $120
• Controller (perf PCB, relay, circuit breaker, ac entry module, misc. electrical parts) - $80
• Wood (two 1/2" x 4' x 8' MDF sheets, 1" x 4" x 6' and 1" x 6" x 6' red oak hardwood) - $75
• Precondensor lens (diameter fl~ plus shipping) - $70
• Fresnels (17" x 16" fl~220mm, 17" x 16" fl~650mm, plus shipping) - $60
• Pipe (two 1" x 2' lengths, two 1" 90° elbows, 5 1" flanges) - $60
• Projection lens and scrap aluminum for custom metal parts from ebay-won Beseler Opaque projector - $40
• Paint (3 cans high temp bbq grill spray paint, 1 gallon exterior white, 1 multipack of foam brushes) - $40
• Fans and fan grills (three 120mm 12v computer case fans) - $24
• Lamp Reflector, shipped - $20
• misc. (felt, nuts, bolts, washers, drawer slides, ...) - $20
• 6" diameter x 2' PVC pipe for case rounds - $15
• Remote Switch box (perf PCB, electrical parts, housing) - $15
• DIY Screen (pvc pipe, blackout cloth, wood trim, wood dowel, two shelf brackets) - $35
• DVD Player (HDMI output, upconverting) - $40
• Computer Speakers - $25
• Testbed (sheet of MDF, threaded rods, light fixture, krypton light bulb, inkjet transparency) - $90

Grand total: $1029

Being able to brag, "I built a big screen TV in my garage", priceless...

Future Enhancements (fl~500mm front fresnel, AntiGlare Removal, Rear Projection, etc.)

fl~500mm front fresnel

The present projector build uses a front fresnel with focal length of 650mm and a projection lens with focal length of 457mm. For the brightest image to be achieved, the lamp is moved backward and is not at the rear fresnel focal point of 220mm. And the light passing through the LCD panel between the fresnels is not completely parallel/collimated.

I suspect that some percentage of light transmission is lost as it passes through the LCD due to the light not being collimated. There is a fresnel with focal length of 500mm available from 3dlens.com that I may obtain and test in this projector. I have no estimate on any improvement that might be gained.

AntiGlare Removal

The "polish method" was used on the present projector build to improve light passing through the LCD panel antiglare (AG) film. However, I am concerned that over time the polish will dry out and become ineffective. The best case would be to remove the AG film altogether, and much discussion can be found in the Lumenlab forums on this subject. At some point I may consider removing the AG film; but at present the risk of damaging the LCD panel out weighs the benefits I perceive would be achieved with the AG film removed.


Rear Projection

When I started the DIY Projector my goals were to make a unit that could be used for movie night and for sports viewing in the covered pavilion of my swim/tennis club. The projector performs fine for movie night but is not bright enough for sports viewing except at night. So only part of my goal has been achieved.

A potential solution for the sports viewing would be to redesign the projector as a rear projection unit. I have read that rear projectors can perform acceptably in situations with moderate levels of ambient light. My club's pavilion has a closet at one end that could accommodate the projector. (A major side benefit of the rear projection setup is that the projector would be protected from weather and vandals since it could be secured in the closet.) I already have a large fl~330mm fresnel and the Lumenlab S15 fl~320mm projection lens that could be retrofitted to the current projector. The retrofit should be straightforward. It turns out that going from front to rear projection without making any projector changes would cause the image to be reversed from right-to-left. However, if one then uses a mirror the image is reversed again and would appear in the correct orientation on the rear projection screen. So I believe the retrofit would only involve swapping the front fresnel, swapping the projection lens and moving the focus mechanism somewhat, and adding a mirror at the projector front external to the case. (It probably would also require a modification to the front case wall.)

There are some concerns with rear projection however. The closet depth is app. 6' and I calculate (using focalcalc143.exe) that my desired screen size of 110" diagonal would be impractical and a smaller screen would have to be used. Additionally, I am still unsure whether the DIY Projector would be bright enough for day time sports viewing without actually trying it. The grey rear projection material I sampled appears to suitable at a cost of $70 for 55" x 108" (3 yards) piece. The wall of the club pavilion would also need to be rebuilt, a big project.


etc.

etc. etc. etc. ...

this is an awesome build with great detailed information and comparisons. Thanks!

The polishing method smoothens the etch like texture of the AG and not fills it in. I've done several polishes in the past and cleaned off the panel with alchohol of any residue left by the compound. Your polished AG should be ok and remain intact over time without any affect to it. However, if you do decide to remove then you invite a higher risk of ruining your polar (graining) and thereafter replacing. So, yes there is much to consider if removing but I've come to realize from experience that protecting the polar is vital. Unless, of course you find a panel without AG.

Thanks for alleviating my concern that the AG might revert to its previous (eg. unpolished) state over time.

Thanks for alleviating my concern that the AG might revert to its previous (eg. unpolished) state over time. I am glad I did the polish since I did measure a substantial improvement in the light passing through the LCD panel after the polishing. I only wish I would have measured the improvement after each application/buff of polish. Since most or all of the measureable improvement was in "coats" 1 to 6 and I saw no improvement between coat 6 and coat 12, I think I could have saved myself some time and effort. Thanks again.

Thank you. You probably noticed that I wrote up the plog after completing the projector.

Thank you. You probably noticed that I wrote up the plog after completing the projector (although I did take lots of pictures and made notes to myself throughout the build). I felt it was important to try to give something back since I gained so much information and inspiration from reading forum contributions. Hopefully there is some useful tidbit in my writeup. I also tried to have at least one picture with each entry since I think it keeps the log visually interesting and saves so much typing (since 1 picture = 1000 words ).

Wow that was an impressive writeup (and an impressive build). I appreciate all the detail that you added to your plog. I know that many set out with that goal but hardly ever follow through with it, especially at the end. I'm glad that it turned out well for you. It sure looks nice!

As others have said, very impressive build.
It's quite a testament to your designing skills that the finished project turned out just as imagined with the plans created with the Google Sketchup program.

If I were you, though, I'd cover those windows at the club with some blackout cloth and throw those Saturday/Sunday afternoon sports parties, anyway.