"The Black Knight Rises" - Week 4 Progress

With less than a week remaining until the 4th annual Southern Fried Gameroom Expo, I was now in overdrive working 16+ hour days.  The lack of sleep was beginning to take a toll on my ability to work.  I began to notice that even small obstacles would completely derail progress as I struggled to think through solutions.  I'm beginning to regret the challenging 4 week timeline.

Still, I pushed forward.  This post is longer than most, as a lot happened this week.


Day 22 - Saturday:
The cabinet felt very close to completion, and I was determined to spend half a day and knock it out.  I had applied the Danish Oil on Monday, and after giving it over 4 days to cure, it was time to apply a polyurethane protective coat.

But first I wanted to finish the trim out.  During parts sourcing, I had stumbled onto reproduction Black Knight cabinet side rails, with the dual flipper/Magna-Save button cut-outs.  Of course, being sized for an 80's style cabinet in which the side rails extend to the very back of the cabinet, they were too long for a modern standard sized cabinet with a pivoting backbox.  To make them work, they would have to be trimmed to length.  I also needed to shorten the backbox hinges.

I put a ferrous metal blade on my compound miter saw and got to work.  The hinges were stamped from thick gauge stainless steel, and cut like butter.  The siderails, while much thicker than modern siderails, were still thin and soft, and more than a couple times I slipped with the saw and bent the rails.  I thought for sure I had damaged them beyond repair, but luckily with a bit of straightening and a more meticulous cutting process, I was able to get them finished.

Cutting the 1980's era Black Knight side rails to make them fit on modern cabinet.

Working again on the cabinet, I aligned and drilled the holes for the backbox hinges.  I also had to notch the side rails for the hinge bolts, and the rails were too tall.  I then assembled and tested the side rails and backbox hinges.  This cabinet build is similar to building a hot rod - merging old with new with unique, plus lots of custom fabrication.

Metal cutting finally done, the backbox pivots!

Read on to see how the final build week turned out...


I then cut the cabinet button holes.  The Black Knight side rails made it easy to perfectly locate the side buttons.  Hopefully this will appease my critics of my Modern Firepower cabinet, who all like to point out that I put the buttons way too low.  They're not wrong.  This time I also only installed a single start/credit button, as the dual buttons on Modern Firepower continue to give pinheads fits.

I also drilled out a few holes for the ball shooter.  This time, I used a different locating method versus what I did for Modern Firepower.  You may recall, for MF, I first installed the playfield, then used a pinball sized dowel rod with a pointed tip, resting in the ball shooter lane groove, to locate where the ball shooter rod would bisect the cabinet.  It was only after final assembly that I noticed that the ball shooter mounting bracket nestled almost perfectly into a cutout in the lock-bar assembly.

This time, after installing the lock-bar assembly, I installed the ball shooter mounting plate flush against the cutout.  I then used the mounting plate to indicate where to drill the through-holes for the ball shooter assembly.

Finally, I located the mounting positions of the legs and drilled the holes for mounting them, and installed the leg retaining brackets.  Since the leg positions seemed spot on for the Modern Firepower cabinet, I simply used those measurements to determine where to mount the legs on the new cabinet.  I had to measure top-down, from the top edge of the cabinet to the top bolt, to make sure I got the positioning correct, as the new cabinet design is not the standard height on the lower front half.  Also, it is interesting to note that while the front retaining brackets mounted inside the cabinet above the bottom panel, the back retaining brackets mounted underneath the bottom panel.

Drilled and cleaned, ready for a protective polyurethane finish.

Unsurprisingly, this half-day of labor had taken more than a full day to complete.  I'm certainly not the fastest woodworker.  Exhausted and with the sun setting, I wasn't even able to brush on the polyurethane.


Day 23 - Sunday:

Before breakfast and even coffee, I gave a very light sand to the cabinet, wiped it down with Naptha followed up by a tack cloth, and brushed on a couple coats of satin polyurethane.  It was at the last minute that I had changed my mind from using the flat "Soft Touch" poly which I originally planned to use.  I kept noticing during cleaning with Naptha that the wood looked richer with a glossier finish, and the matte look of just the Danish Oil appeared a bit cheap.  I debated between semi-gloss and satin, and decided to go with satin as I really don't like glossy wood.  The satin also provided additional scratch protection versus the flat.

Allowing the polyurethane time to cure before applying a third coat the next day, I turned my attention back to the playfield.  Time to begin wiring.

First, I determined where the Chameleon IO Controller would be located on the playfield and mounted it in place.  The Black Knight playfield is very busy, and with the main playfield having a large hole cut into it for access to the upper playfield, a good location was hard to find.  The small size of the controller gave me a few options, and I finally settled on a large, centrally located and mostly un-populated area just above the right magnet, which only obscured access to a single GI lamp.

With pre-made wiring harnesses in hand, I began wiring up the switches.  I decided to use blue and yellow wiring harnesses for switches, red and white for LEDs.

Pre-made wiring harnesses for switches and lighting.  The cables are 3 feet long, making it easy to reach all corners of the playfield from the centrally located IO controller.

I was especially pleased with my 9-wire harnesses.  For the Modern Firepower build, I had used pre-made wiring harnesses that I purchased along with the U-HID controllers.  The harnesses were both very expensive ($10 each) and fairly short.  My new Chameleon IO Controller was designed to use up to 21 harnesses (14 for lighting and another 7 for switches); a $200+ harness cost would almost triple the cost of my IO controller. Reducing the electronics cost was a primary consideration for this second generation build, and that extended beyond just the electronics to include the wiring harnesses as well.   I was able to have these new harnesses manufactured to my specs for wire gauge, length and colors for a fraction of what I paid for the U-HID harnesses.

The wiring harnesses also greatly accelerated the process of wiring the playfield.  I would plug a wiring harness in the controller, route the wires to an area of the playfield, securing them as I routed, then cut, strip and solder the wire to each component.  For both the switches and the lighting, I used shared power and individual grounds, so for the power wire I would daisy chain the blue wire from switch to switch.

The hardest part of wiring the switches was actually removing the old wires and solder, since these were mostly the original switches from 1980.  Within a few hours, I had all 32 switches wired up.

Switch wiring complete.  Yellow are the ground signal lines, and blue are the shared power lines.

Though each individual wiring harness includes 1 power wire and 8 ground signal wires, in several cases I uses the power wire from one harness to supply power to switches using the signal wire on another harness.  I wasn't concerned about overloading any circuits or wiring, as switch circuits have high-resistance and consume almost no power.  I probably could have used a single power wire from one harness to provide power to all switches.

Also note that I used yellow wires for every switch, not individually colored wires.  This made it easy to wire up, as I didn't have to stop and make sure I was wiring the right colored wire to the appropriate switch.  It's also cheaper, as having individually colored wires is much more expensive to manufacture.  Later, I will map each physical switch to the corresponding port on the Chameleon IO Controller using a mapping routine built into my software.

With the switch wiring complete, I began work on the LEDs.  The process was nearly identical to the switches, a red wire for shared power, and white wires for signal wires.  Though I hoped that this process would go quickly, I soon found that wiring up my ChameLEDs was a slow and tedious process.

This all stemmed from two design choices.  One was making the ChameLED PCB capable of controlling RGB LEDs, requiring 3 signal wires plus power.  The other design choice was making a connected strip of 8 ChameLEDs sub-boards that could be wired from one end of the connected larger assembly.  This meant I needed to be able to run up to 24 signal wires, plus power, from a single sub-board with a working space barely larger than a postage stamp.

I also wanted to engineer a wiring solution that would be both fast and cheap.  Using wiring headers with nearly 30 connections would be very expensive and large, but PCB holes are practically free and space efficient.  So my solution was to have 27 20-gauge thru-hole type pads located on the PCB.  During PCB assembly, solder paste would be applied to the holes to fill them with solder.  Now, during playfield wiring, all I had to do was cut a wire to length, strip about a millimeter of insulation off the tip of the wire, position the wire in place over the appropriate pad hole, and re-melt the solder in the hole with the tip of a fine pointed soldering iron.

Done properly, the solder melts in a few seconds, and the wire effortlessly slips into the hole.  Remove the heat, and the wire is locked into place.

In reality, this only worked effortlessly about 20% of the time.  Typically, I would have to continually re-position the soldering iron tip trying to get the solder to melt, and in a few cases this took several minutes to accomplish just for a single wire.  The wire would quickly get too hot to hold, so I had to use a pair of pliers to hold the wire in place on the hole.  Often I would apply too much down pressure trying to force the wire into the hole, and the wires would bend.  I would then have to straighten the wire tip (sometimes by cutting and re-stripping the wire) and try all over again.

Wiring the ChameLEDs was a slow and challenging process.

The reality of my design is that I had created the equivalent of trying to thread a needle, while holding the thread with a pair of pliers, and simultaneously applying 650F of heat to the needle to melt metal filling the needle's eyelet.  In retrospect, not my finest design moment.  The ChameLEDs had failed to solve any of my design goals, and created new problems in the process.

Worse, because of the vertical nature of the wire insertion, my wiring job was not nearly as attractive as what I was able to accomplish for Modern Firepower.

I was only able to get about about 16 ChameLEDs wired up before it was time to get some hard-earned sleep.


Day 24 - Monday:
At the break of dawn and partially rested, I returned to the arduous task of wiring the LEDs.  With somewhere in the neighborhood of 80 lights on the Black Knight playfield, I had a lot of work ahead of me.

With practice I was able to wire the ChameLEDs faster and faster.  In a handful of cases, I was able to cut, strip and solder a wire to a ChameLED in 15-20 seconds, though in most cases it still well over a minute.

Unlike the switch wiring, I made sure not to share the power wire from one harness with the LEDs connected to another harness.  This is because each bank of 8 LED outputs is power limited to 1A.  With up to 8 LEDs per bank, that's a generous 125mA per LED, enough for some seriously bright LEDs.  But if the power wire is shared to other banks, it could easily carry more than the design limitation of the Chameleon IO Controller.  By wiring up just 8 LEDs at a time (one harness), it was easy to make sure that I kept the power wire with the same 8 signal wires.

Before the day was done, I had completed the wiring for all the lights.

LED wiring complete.  Even had a few spare wires left over for repairs if needed.

I also gave the cabinet some attention.  Since it had been over 24-hours since the last coat of poly, I lightly sanded the entire cabinet, cleaned again with Naptha followed by a tack cloth, and applied a third coat of poly.  Though I briefly thought about applying a fourth coat, I felt I had pretty much achieved the look I wanted, and called it quits.

3 coats of satin finish polyurethane gave me just the sheen I was looking for.

Day 25 - Tuesday:
The last bit of wiring I needed to accomplish was the solenoid wiring.  With only 20 solenoids, I expected this to go very quickly.  But first I needed to make a new Chameleon Power Driver to replace the one I fried testing the magnets.  Instead of just making one, I decided to make six new power drivers.  This would give me spares in case I needed them, plus extras for show and tell at the expo.

As I was assembling the surface mount components on the PCB, I was alarmed when I couldn't find some of the wire headers in my box of parts.  I spent hours searching to no avail.  I looked back at my purchase logs, and sure enough I should still have 10-15 wire headers available... somewhere.  After wasting entirely too much time looking for the lost bits, I placed an order to overnight the parts from Mouser.  My wallet felt the pain of the overnight shipping charges, and in general I've dropped a sizable amount of money on expedited shipping charges to make sure I got parts in time to be ready for the show.

I was also worried that the solder paste would go bad overnight - it is only rated for about 8 hours once applied to the PCB before you have to bake it in the reflow oven.  While each power driver is not terribly expensive, throwing away 6 partially assembled power drivers would certainly add to the financial pain.

While I couldn't get all six finished, I was able to get one assembled and baked.  Within minutes of coming out of the oven, it was installed on the upper playfield, and 7 solenoids were wired up.  Before proceeding further, I decided to test it to make sure it was operating correctly.  I hooked up the 38v power supply I got for The Black Knight Rises build, plugged it in, and switched on the power.

Immediately all seven solenoids fired, BAM!  It's quite a startling sound when so many solenoids fire in unison.  Chances are, you've never heard how loud a solenoid is when it is not muted by the thick cabinet walls, they are easily three or four times louder.  Seven together is an altogether special experience.

As fast as I could, I switch the power back off, and the solenoids relaxed to their normal resting state.  I then toggled the power several times, and each time all seven solenoids fired.

The good news is that the power driver was handling the power demands without missing a beat, so yeah, at least I had that going for me.

Off of the playfield and onto the test bench the power driver went.  An hour later, I was still befuddled.  I felt I was missing something obvious, but my sleep deprivation and general state of exhaustion prevented me from unearthing the mystery.

Comparing the defective unit to a working unit, I could find nothing visibly wrong.  The most likely candidate was that I had installed the wrong biasing resistors on one of the transistors, as the wrong resistors would cause the on/off switch to occur at the wrong time.  The 0402 sized resistors, so named for their size of 4mm x 2mm, about the size of a flea, proved very difficult to measure even with fine tipped probes on my multi-meter, but eventually I determined all values were correct.

I then considered that I had possibly installed some of the smaller transistors backwards.  Using a microscope (yep, these are truly tiny parts), I reviewed the orientation markings on all the components, again comparing bad vs. good, and again determined that everything was assembled correctly.

The list of remaining possibilities haunted me.  Bad PCBs?  Bad solder?  Bridged solder connections?  Components out of spec?

At this point, Tuesday was coming to an end.  Solenoid wiring could not be completed, and new wiring headers for the remaining Chameleon Power Drivers wouldn't arrive until the following day.  The detour had also taken time away from cabinet assembly.  With only Wednesday and Thursday left for assembly and game configuration, it was now obvious that The Black Knight Rises would not awaken for the show.

A change in strategy was required.  Instead of taking a working pinball machine to the expo, I would instead take a "work in progress" for a bit of show and tell.  This decision immediately lowered my stress level.  With two, maybe two and half days remaining, I felt certain I could complete physical assembly in time for Friday.


Day 26 - Wednesday:
The shipment from Mouser arrived surprisingly early - I hadn't paid for the Early AM delivery.  Within an hour I had positioned the new headers on the 5 power driver PCBs and baked them in the reflow oven.  A cursory inspection appeared that they had all survived the ordeal, though electrical testing would have to wait until after the show.

With the last coat of polyurethane having been applied early Monday afternoon, the cabinet was now ready for final assembly.

The first step was to get the wheels mounted.

"Wheels?", I'm sure you're asking.  Yep, this baby is going mobile.  The typical pinball machine is large and heavy, only designed to be stationary and to resist the aggressive English of the most hard-core players.  For this new cabinet, one of my design goals was to be able to transport it easily to trade shows.  Larger companies have the resources and manpower to throw at such tasks, but for me moving even a single pinball machine requires renting trailers, dollies, and finding manpower (Troy to the rescue).

I had already designed the cabinet to be smaller and lighter, and for mobility my plan was to fabricate a wheel assembly that could be installed and removed and in matter of minutes for transport.  I wanted the wheel assembly to be both lighter than a dolly, and off-road capable.  I also wanted the wheel assembly to work on all my machines, so I sized it to fit both the Modern Firepower cabinet, as well as the new cabinet.  It took a few hours of careful measuring, plus a trip to The Home Depot for some mounting bolts, to get my new wheel assembly done.  In the end, I think I have the worlds first mobile pinball cabinet.

If for any reason The Black Knight doesn't rise, I can re-purpose this cabinet into a soap box derby car.

Next I turned my attention to mounting the playfield support rails.  I had used wood rails for my first cabinet build, and never liked the solution.  Recently I found and bought some Stern metal playfield slide rails and pivots on Pinball Life at a really attractive price.  I really wanted to use them, but didn't know how to install them.

Luckily I found a nice video on YouTube illustrating how to lift a Stern playfield, and the video included an in-cabinet angle that addressed all my questions.

I then searched for my old test playfield in which I had cut a hole in the center for locating the rails on the Modern Firepower cabinet.  I couldn't find it anywhere.  I think I must have foolishly tossed it away, thinking I wouldn't need it again.  These new slides and pivots were very different than using wood rails, so I couldn't simply take measurements of the Modern Firepower cabinet.

Luckily, I found an old test playfield that a CNC company had mocked up for me (one with completely unrealistic holes, but they were just showing off what they could do), and it had the right external dimensions.  I cut the center out of it to make a new mount locating jig.

Then when I went to install this test playfield I found that one of the playfield hangers couldn't be attached to the playfield because the right side cutout was several inches wider than a real playfield.  I quickly found a spare piece of walnut plywood of the right size, glued and clamped it in place, and waited for it to dry.

I glued a scrap piece of walnut plywood to the test playfield so that the mounting bracket could be attached.

I used this down-time to get the cabinet legs bolted on and the cabinet leveled in preparation for mounting the support slides.

Finally I was able to continue.  I got smart this time, and used a long bar clamp to slowly raise the far end of the playfield while watching the digital level's readout, getting to exactly 6.5 degrees.

I used a long bar clamp to raise the playfield until I achieved a 6.5 degree rake.    At the sides of the cabinet you will notice the Stern playfield slides waiting to be installed, and one of the pivots mounted on the playfield.  You can also see the four extra large t-nuts in the bottom of the cabinet, used to mount the wheels assembly for transport.

From there it was easy to hold the slide rails in place and mark the screw locations.  With the support slides installed, it was time to trim out the cabinet.

I now had a decision to make.  I had purchased the old style Black Knight side rails, and cut them to fit the modern style cabinet.  But in case that had gone horribly awry, I also had purchased a spare set of modern cabinet side rails.  Since both would fit, I now could use whichever looked better.

The new style rails are slim, and help show off more of the wood.  Ignore the protective vinyl coating, they're stainless.

The original Black Knight side rails obscure more of the beautiful wood, but gives a retro look and is "slimming".  Bonus points if you can name that car...

Troy thought both looked good, but it didn't take long for me to settle on the original Black Knight rails.  After all, the project is The Black Knight Rises.  I found the thicker rail to provide a slimming effect on the cabinet, making it appear to be even thinner in the middle.  Also, if I didn't use them now, I wasn't sure when I would use the rails.

Finally I got the rails, buttons, coin door and ball shooter installed. I slid in a filthy piece of glass (no time to clean it now) and levered the lock bar into place.  Removing the protective vinyl from the stainless parts, I could now, for the first time, appreciate the fruits of my labor.

Almost complete.

Around this time I got a shipment from Ponoko, the new acrylic panels for the backbox.  The box was smaller than I expected.  I eagerly opened the package, and with newfound horror I realized that there was no way that these panels were the right size, especially the monitor trim panel.

With growing relief I placed the trim on the monitor and found it to be the right size after all, and that it was just a weird optical illusion.  The panels also appeared to be the right size to slide into the backbox.

Again the sun had set on me.  I had spent nearly the entire day on just the cabinet.  Only a day and a half remained until I had to load the trailer and head out to the show.


Day 27 - Thursday:
Eager to get the backbox done, I started off by test fitting the new acrylic panels.  I had routed a 1/8th inch slot into the sides of the backbox, slightly larger than the acrylic panel thickness, but as I slid the panels in, I found they were grabbing in various spots.  Of course, I had applied a few coats of polyurethane, and this seemed to be gumming up the works.  With an Exacto knife I spent an good hour carefully opening up the slots, trimming away the poly, until the panels slid into place.

To finish the backbox, I would also need a panel upon which to mount the speakers and DMD.  I already had the CAD design, but I needed to tweak it slightly in width to fit the actual finished dimensions of the backbox, which was slightly out of spec due to the dovetail joinery.  I then created a gcode file for the CNC machine.

Troy arrived, and immediately lent two huge helping hands.  My CNC router, which had given me so much trouble over recent months, needed a quick tune up before use.  I had purchased new GT2 belts, and we spent an hour or two getting those installed and tensioned. With the router finally ready for use, we taped down a 3/4" slab of MDF and cranked the router up.

Now, I'm not exactly sure what I did wrong, most likely multiple things, but the routing did not go as planned.

Since one of the main problems I had been struggling with was backlash and belts loosing tension, I had decided to set the feed rate really low.  Sure, it would take longer to route, but I had other things to work on in the mean time.  A low feed rate would keep stress on the axis belts low and minimize deflecting, ensuring I would get a good result.

Since I was running the feed rate so low, I also set the spindle to 50% speed, 12k RPM.

I watched it for a while, manually vacuuming dust as it slowly progressed, as I'd yet to make a new dust shoe for the new spindle.  It seemed to be doing fine on its own, and I took a short break to let the dogs out.

When I returned I immediately noticed a burning smell.  The MDF board appeared fine, so it seemed the smell was coming from the spindle, and I discovered that the spindle was too hot to touch.  Not sure what to do, my first reaction was to slow the spindle down to 9k RPM, still plenty fast for the slow feed rate.

Within a minute the spindle came to a halt, and an error message appeared on the VFD:  E.O.C.A.  The spindle seemed even hotter than before, so too the large steel spindle mounting bracket.  I quickly performed a feed halt on the CNC before the bit snagged the wood.

Not sure what to do, I left the spindle to cool down.

In the mean time, I got to working on the acrylic panels.  The plan was to paint the back-side black, and leave the front side unpainted.  The clear acrylic would then have a glossy black appearance, like a super thick gloss coat.  Instead of simply using black acrylic, I chose this route so I could mask off the DMD window and leave it unpainted.

Originally, I had thought it quickest to first glue the finished MDF speaker/DMD support panel in place, and then trace the DMD cutout onto the acrylic.  With the ongoing CNC disaster, that plan was out the window.  Instead, I spent an hour carefully measuring the exact size of the the DMD panel, and the planned mounting location, and got the window masked the hard way.

Troy took care of the painting.  I had two different types of paint, and unsure of which would work better, Troy tried both on some test scraps of acrylic.  Surprisingly, the paint that promised adhesion to plastic was plagued by bubbles and spots.  A few hours and several coats later, the acrylic panels looked even better than I had hoped.  We left them to cure overnight before installation.

Checking up on the CNC, the spindle was still hot to the touch after over an hour of cooling down.  I positioned a fan blowing on it, and a few minutes later it was finally cool.  Unfortunately, every time I tried to start up the spindle, it would momentarily start to spin up, then stop with the same E.O.C.A error message on the VFD.

Half an hour of research later, I concluded the spindle was toast - the error message indicated an electrical over current condition and that the wires were shorted out inside.  Running the feed rate so low through MDF created a very fine dust, which likely found its way inside the air cooled spindle, killing it.  One or more of the spindle settings on the VFD may have been wrong, causing a premature death.  Or running an air cooled spindle at slower speeds may have reduced the cooling effect too much.  Most likely it was a combination of all three.

Regardless of the root cause, there was no way I would be able to finish the speaker/DMD panel in time.  The design was too complex to cut by hand with the tools I had available.  My show and tell would be a little less showy.

The next task was to get the monitor mounted into the backbox.  I had left the metal mounting tabs on the monitor, which was originally used to affix it to the plastic monitor housing.  The plan was to 3D print mounting strips that would be screwed into the backbox, and to which the monitor would then be screwed.

It took an hour of careful measurements to come up with the design, and then 6 hours to print the first prototype.  The tolerances were so tight, I needed to make some small tweaks and print them again. I didn't get the final designs printing until around midnight.  Hopefully the prints wouldn't fail and I would wake up to working parts.

Since progress on the cabinet was stalled, Troy and I turned our attention to Modern Firepower.  Since The Black Knight Rises would be a non-working demonstrator, my first priority had become making sure Modern Firepower made it to the show in good working condition.

One of the mistakes I made in the Modern Firepower cabinet design was using a large, 32" LCD panel as the backglass.  The backbox is so wide, it will not fit through the door to my office where the machine lives.  The only way to get through the door is to remove the backbox from the cabinet, which we accomplished fairly quickly.  On the plus side, it removed over 50 lbs of weight, making it easier to get the cabinet down the stairs and to the garage.

We also removed the playfield for the trip down the stairs.  This only took a minute or two, since the playfield is connected to the cabinet by just a USB cable, a 50V power cable, and a 5V/12V power cable.

Next, we drilled four holes and mounted the wheel assembly to the cabinet using t-nuts and long bolts.  This made it so much easier to move the cabinet out of the office, though it was a challenge making the 180 degree turn halfway down the narrow stairway.  With a bit of muscle, we managed. I hate to think how much harder this would have been with the heavy backbox and playfield attached.

With everything in the garage, we reassembled Modern Firepower.  First the backbox went on, then the playfield went back in as easy as it had come out.  All the wires were hooked up, and we finished off the day by giving the cabinet a good cleaning.


Day 28 - Friday (A.K.A. Delivery Day):
I only live about 15 minutes away from the Galleria/Renaissance Waverly Hotel in Atlanta where the 2017 Southern Fried Gameroom Expo is being held.  With a game delivery deadline of 3PM, I had figured that we had until 2:30 to get on the road.  My day started at 6:30 am (T-8 Hours), with a quick run to U-Haul to pick up a 5x8 enclosed trailer.

Within an hour, I was home and assembling the backbox.  The Acrylic slid into place easily, and the new 3D printed monitor mounts worked well enough, with a fit so tight I almost didn't need screws. It took a long time to get the screws installed, solid walnut is very hard.  T-6 Hours.

The red 3D printed monitor mounts are screwed into the backbox, sized perfectly for positioning the monitor.

Next I drilled seven ventilation holes in the back panel of the backbox.  I sized these to work with some grommets that I would order later.  For now, the ventilation holes were to serve two purposes: cooling, and providing a handhold to actually open the backbox! T-5 hours.

Thankfully, about that time Troy arrived.  Before proceeding with further assembly, Troy hooked up the PC to the monitor in the backbox and made sure everything was working.  During the monitor de-casing (the process of removing the LCD panel from the plastic monitor cabinets the are installed inside), I had to separate the control boards from the LCD panel.  Weeks later I had remounted the control boards directly onto the LCD panel (a 3D printer and epoxy are great problem solvers), but I was concerned I might have wired something wrong during reassembly.  Luckily, Troy found everything to be working perfectly.

In the meantime, I quickly modified my PC mount design to be VESA compatible and mount directly to the monitor instead of the cabinet, and got that printing.  The print would take about 3 hours, but would finish around 1:30 pm, still plenty of time.  T-4 Hours.

Next, I spent an hour or so working on a slide show of pictures from the build.  The plan was to let the Window's screen saver simply cycle through the pictures, so anyone examining the build could see the behind the scenes process.  I thought this would be better than a lifeless cabinet.

T-3 Hours.  Realizing time was running out, I gave up on the slideshow.  Way too many pictures to wade through.

With the backbox acrylic installed, we were finally able to mount the backbox onto the cabinet (the backbox hinges hold the acrylic in place, so the acrylic has to be installed first).  Then we mounted the back panel on the backbox, drilling holes for the hinges and screwing it into place.  Troy worked on installing the power strip, and PC and Monitor power supplies in the cabinet.  Troy also worked on some final details like covering the multitude of nail holes in the side rails.  Chrome thumb tacks proved surprisingly simple and effective.  T-2 Hours.

Installing the back panel on the backbox.  Yes, even the back is walnut.  Only corners, not costs, cut here.

Next, we pulled the playfield off of the rotisserie, and installed the Stern playfield edge slide brackets and pivot brackets to the playfield.  It was at this time I discovered that a playfield side rail screw was blocking the correct positioning of the pivot bracket - obviously this playfield had never been designed to use these playfield slides.  With no options, I had to adjust the pivot location by a few millimeters, making sure to go closer to the center of the playfield, else the playfield would not properly rest on the cabinet slides.

The wiring was nearly complete.  Still need to wire the bottom seven solenoids... after the show.

 
All of these little unexpected challenges seem to be the standard when making a one-off custom machine.  And while quick and easy to resolve, an extra 5 minutes here and an hour there add up to significant lost productivity.  Yes, this is also called "foreshadowing"...

With the playfields slides, pivots and hangers installed, we dropped the playfield into the cabinet for the very first time. It slid in smoothly and found its new home, even with the adjusted pivot bracket mounting location.

A view to a kill...er interior.  Look how clean it is!  The bottom panel is two pieces, with the first foot or so steeply raked.

Looking inside, I was expecting to only see a couple inches of clearance between the cabinet floor and the tallest playfield components. It seems that there is closer to 4-6 inches of clearance.  Looks like I can go even thinner for my next cabinet design.

There's quite a bit more than the 2" of clearance I had planned for.

The new VESA compatible PC mount was now finished, and fortunately it fit perfectly.  I would throw the PC in the travel bag with the cabinet legs and tools, and install it at the show.

Inside the backbox, the red 3D printed PC mount attached to the VESA mount on the monitor.

Running out of time, we re-installed the filthy playfield glass (no time to clean it now).  This was it, The Black Knight Rises was now ready, as good as it could be, for showtime.  T-1 Hour.

The Black Knight Rises, ready to go to the expo.

Not a good photograph.  The acrylic panels in the backbox are really shiny and blend in with the LCD screen.

With time, minds, and bodies exhausted, we then got the Modern Firepower cabinet covered and loaded onto the trailer.  It was at this moment that harsh reality began to sink in.  My new cabinet mobility solution involved leaving the front legs installed on the cabinet (they work great as steering handles, and allow you to rest the cabinet on the front legs instead of the back side).  But this now posed multiple problems on the trailer.

First, the 5' wide trailer wasn't wide enough for two machines side by side against the front wall.  No biggie, we thought, simply turn the cabinet sideways and secure them to the side of the trailer - but with the legs attached the cabinet couldn't be rotated either.

I looked at the clock:  2:15PM.  We had spent 45 minutes getting Modern Firepower packed and loaded.  We needed to be on the road in mere minutes.  I still had to let the dogs out again and lock up the house.  The extra time needed to remove the front legs on both cabinets was just too much, so I made the call:  The Black Knight Rises doesn't ride today.

We secured Modern Firepower and a few minutes later we were on the road.  We got to the expo just in time.  I asked about being allowed to make a second run to bring the knight, but was firmly told no, and for good reasons too.  The show was more than full with too many machines, Friday traffic in Atlanta is brutal, and beyond everything else, my new machine wasn't yet operational.


Week 4 Complete - Game Over?
Having missed the 2017 Southern Fried Gameroom Expo, what now for The Black Knight Rises?

Well, first I need to relax and recover.  I pushed myself really hard these past four weeks.  I did the work of two men across at least eight different professions, and I'm flat out exhausted.

But the project is far from over.  In the coming weeks I will continue, and finish, The Black Knight Rises build.  On the to-do list is: resolving the electrical gremlin that was affecting the Chameleon Power Driver, assembling a new Chameleon IO Controller, fixing the CNC router and getting the MDF cut for the speaker/DMD panel, installing the speakers and DMD, wiring up the cabinet buttons, updating my Chameleon Pinball Engine software with a few new routines for Black Knight, and configuring the game rules, sounds and animations.   Hopefully it is really only a week or two of solid effort.  But without the added pressure of a pending deadline, I'll take my time from here on out, and it might even take a few months.

I may also go ahead and redesign my ChameLEDs.  I have some great ideas for the next revision, and it might be worth the effort to strip off and redo all the lighting to test it out.  But not before completing the rest and ensuring everything else is in proper working order.

I'll continue to post build updates as The Black Knight Rises.

If the build turns out as good as I hope, I may take The Black Knight Rises on a road trip to Orlando for the 2017 Free Play Florida expo in November.

Thanks for reading!!!