Resizing a 400x400 honeycomb bed to 400x300

Resizing a 400x400 honeycomb bed to 400x300

I got my Full Spectrum Laser hobby laser used, and unfortunately, it did not include its honeycomb laser bed.

I ordered a small one initially, because it was on sale, even though it was small.  Overall, it measured 295 mm x 295mm, though it had markings in the middle that range 0 to 250.

The enclosure of my laser allowed for a honeycomb about 400mm wide (X axis).  Depth (Y axis) could fit quite a bit, but the gantry takes up a lot of it.  And using the too-small bed as an example, it seemed reasonable to target about 300mm depth for the new bed.

It's surprisingly difficult to find a 400mm x 300mm honeycomb bed these days.  Most, it seems, are geared towards diode lasers, which are 400mm x 400mm.

Prices are coming down quite a bit, too.  I saw one person wanting to sell her used, cruddy honeycomb for $80, but I could buy one new on Amazon for $18.99 (with "free" shipping).

I bought one of those 400x400 beds, and got it in a few days.

 

The bed I got did not have measurement markings on it, like the 295x295 did.  But overall, it was constructed in a similar manner.  There were four edge pieces, which appeared to be extruded aluminum.  Those were connected somehow using rivets, two on each corner.  The honeycomb itself appeared to be strips of corrugated steel, and there were rods going through them.

Naming, labeling, analyzing

 

I named the frame pieces Top (1), Left (2), Right(3), and Bottom(4).  The corrugated strips run parallel to Top and Bottom.  The internal rods run up and down, parallel to the Left and Right frame pieces.

Remove six of the rivets

My first investigative task was to drill out a few of the rivets.   I just used a hand-held power drill and a strong-looking bit.  After some amount of chipping out, the last thin ring of metal breaks off with a satisfying click.  I ended up removing the two rivets for each of the Top, Left and Right frame pieces.

A look under the covers

Once I had the rivets off, I could wiggle the top frame edge off.  This then revealed what was going on under the covers.

 

As it turns out, the honeycomb is quite simply and somewhat accurately built.

The four frame edge pieces are indeed extruded aluminum.  In my case, it appears they're anodized, so they are black on the outside.  They're like square tube, missing one side.

The frame edges are cut at a 45 degree angle at each corner.

The edges are joined using steel L corner brackets.  They're strenghtened by a little crimp to keep them from bending under force.  The rivets held the L brackets and frame pieces tightly.  This is one of the corner brackets taken from a top corner, and before I'd removed the rivet backing.

The width of the corner brackets, and the width of each corrugated steel strip, are quite close in size to the width of each frame piece channel.

Remove the frame side pieces

This picture shows the side and top removed.  Note the rod at the top.

Under the top frame piece, welded across all four vertical 4mm rods, there is a 3mm rod.  While this was butted up against the topmost corrugated strip, it wasn't attached to the strips.  It was just holding the vertical rods in place, and kept a little pressure on the corrugated strips.

Having the top and bottom bars welded like this would keep the whole assembly from twisting, too.  And I suppose it would make the top more rigid.  But really, I think that having tight tolerances in the strip widths and the corner bracket width, with respect to the frame channels, keeeps things from moving too much anyway.  So, I didn't see there to be a great need to re-weld the rods later.

Check for square

At some point, I held the Top, Left, and Right pieces up to each other, and was really surprised to find that they were not all the same length!  I think that was a manufacturing flaw originally, and I just didn't notice it.  (For the picture below, the bottom edges of all three frame pieces are resting on flat ground.)

 

Detach the top rod

I used a Dremel with a cutting wheel and cut down through the existing weld.  Remember to use proper eye and breathing protection when cutting!  Once in far enough, I could use a flathead screwdriver to break the rods apart without too much effort. 

Clean up the weld points on the top rod

I had taken off the top rod in order to remove excess corrugated strips.

I used the Dremel again, this time to grinding away the remaining weld on each vertical rod.  Remember to use proper eye and breathing protection when cutting!  

This left some material on the rod.  I tried to slide strips up and off, but they would get hung up on remaining weld material.

To get things round and smooth again, I stuck a little piece of sandpaper inside my handheld drill, and spun the sandpaper around the weld point until smooth.  I'd then tighten the collet and repeat until satisfied. 

Remove the strips

 

I lifted many of the corrugated strips up along the four rods, and off.  (If the rod weld points aren't clean enough, you'll know when you try to do this step.)

I stacked the strips as I was going.  It's best if you flip every other one so that all the lengths match and the holes line up. 

I didn't know how many I would be removing, so I just kept going until I had about 11.75" remaining (measured from outer bottom frame edge to top strip).

The following picture shows the honeycomb with many strips removed, and long rods accessible.

 

Measure, mark, and cut the four vertical rods

(Prior to this step, if you want to practice using a die to thread the rod, you can.)

This part was tricky.  If you don't take the whole thing apart -- that is, if you leave the bottom rod welded -- then it's hard to measure and make all four rods the same length.  I took a "close enough" approach here.  

The 4mm rods were spaced about 95mm apart.  I had a 1/2" plywood board with holes in it already, so I could press that down each rod to the point where the honeycomb strips were tensioned, and then draw marks.  To get this right, I took off more strips (to account for the 1/2" of plywood) and then put those back later.

 

When done, I eyeball-checked the alignment of the four marks (one on each rod).  In the picture above, I had marks in Sharpie that were incorrect.  The right target is the scratched point closer to the honeycomb.

For cutting the rods, it was easiest to start by using bolt cutters to cut close to the real cut point.  That left room for the Dremel to get closer.

I used the Dremel cutting wheel again to get through most of the way, and then cut with flush cutters at the very end.  Remember to use proper eye and breathing protection when cutting! 

After cutting, I used the Dremel again.  This time, I was trying to chamfer the top, cut edge of each rod.  This would make it safer and also it would make it easier to die-cut threads onto the rod ends.

Use a die to make screw threads at the top of the vertical rods

This was a handheld operation, but it worked pretty well.  (I practiced on the longer rods before step 5).  I wasn't sure what thread to use, but just tried different dies until one fit.  As it turned out to be a #10-24, I felt comfortable using it.  I cut the thread about 1/2" down on each rod.

For this step, I removed more strips temporarily.

One problem I faced was that my tap/die set has a die with arms that are pretty long.  They provide leverage, but in this project they would hit the neighboring rod on each half turn.  I ended up starting each thread with the leverage arms in place, and then removed the arms and continued cutting as I got further down.  

This picture shows the rods now with threads.  Extra strips have been removed so that I could have room to work.

Then I put the temporarily removed strips back.

Use a nut to hold the strips in place, rather than welding a top rod.

I didn't have #10-24 nuts immediately available, so I drew some up in Autodesk Fusion, complete with modeled threads.  I printed those in PETG and they worked straight away.  Each rod got one.

Chop the left and right side pieces to the new length

I put the top frame piece back on, kind of as a dry fit.  This allowed me to get a rough measurement for cutting the Left and Right side pieces down.

I also tested the springiness of the assembly, and I figured I had a 1/2"  of travel, so I could afford to be a little imprecise in my measurement.  The main thing was to ensure that the Left and Right sides (A) had near-perfect 45 degree cuts, and (B) were exactly the same length with respect to each other.

To do that, I lined up both pieces, and stuck them together with blue tape.

I used my chop saw to do the actual cutting.  I used its bevel mechanism, tipping the whole blade over to 45 degrees, and used a separate measuring tool to ensure it was really close to 45 degrees.  Then, I did a first test cut, partly to make sure the blade was up to the task, and partly to get a feel for the cut.  Finally, I marked the cut points, and cut the pieces.

Nobody was injured.  The pieces looked really good.  At this point, I also used some sandpaper to smooth out any sharp edges resulting from the cuts.

Tap the L brackets

The L bracket holes were slightly under 4mm, and so it seemed doable to use #10-24 threads for them now.  I realized around this point that I really had to drill out the Bottom piece rivets, and at least loosen the Bottom piece in order to remove the bottom two L brackets.

I then cleaned off all the remaining rivet material from all L brackets.  This was best done by using a flush cutter to chop off the small, outward protrusion.  Then, I used pliers to pull out the large, inner piece at each hole.

Once that was all done, I used a #10-24 tap on a non-powered drill press to thread the holes.  By this point, I had spent $1.48 for a packet containing ten #10-24 x 1/2" pan-head screws.  Each came with a nut.  I tested each threaded hole as I went, using a #10-24 screw, to make sure I wouldn't be surprised during assembly.

Widen the holes in the frame pieces

Because I was using #10-24 screws, the existing holes in frame pieces had to be widened.  I used a #13 drill but could have gone a little wider.  The idea here is that you don't want the screw to be threading into the frame hole.  Ideally, the screw should slide freely through the frame hole, and catch on the L bracket thread.  In practice, the #13 hole size still was a bit too small, but it ended up being something I could force.

To do this, I finally gave in and removed the Bottom frame piece.

Partially reassemble Top and Bottom

I reattached the L brackets to the holes of the Top and Bottom frame pieces.  That's easiest to do when they're removed from the honeycomb.  Then, I slid each Top/Bottom frame assembly back in place.

The Top and Bottom pieces slide pretty easily back over the corrugated strips at top and bottom.  The side reassembly is harder, just because of how the corrugated strip ends can get offset, and even the slightest change in them can cause the edges to catch on the frame piece.

(In this picture, you also can see the pile of strips that were removed prior to this point.)

Make new holes in the chopped-down side pieces

At this point, the Left and Right pieces were chopped down, leaving the bottom section with its original hole, and the top section of each missing its hole.

I had two chopped-off ends, though, and they served as guides/jigs for determining placement of the new holes.  I lined up the chopped-off piece with each Left or Right piece, stuck them together with blue tape.

 

I went to the drill press and drilled a #13 hole exactly where a hole would have been.  

Assemble and align

Now I had the Right and Left frame pieces chopped and drilled, so all that was left to do was reassembly.  Or so I thought.

Getting the Left and Right frame pieces back over the full set of corrugate strip ends is difficult, but there is a technique that can be used to make it easier.  It involves hooking the Frame edge piece over one side of all the strips, and then  having an extra straightedge that aligns the strip edges. 

After each of Right/Left frame piece was back in place, I simply connected the #10-24 screws through the frame to the L bracket at each end.

At this step, it helps sometimes to loosen the L bracket's other screw, and then tighten them judiciously.  Overtightening one side can cause the ends to become misaligned, resulting in a bad corner and an exposed, sharp edge.

This is one corner with screws in place.

At this point, the whole honeycomb was back together again!

 At this point, I also zip-tied together all the extra strips.  Only one was kind of scuffed up -- the topmost one, due to how I broke the top rod welds -- but the rest are in good shape.  It looks like I ended up removing 25 of them to bring the honeycomb size down by about 100mm.

 

Measuring them, I'm getting about .33mm thickness, both measuring individually and averaged over a compressed stack.  They're nearly exactly 20mm wide.  Similar in flat, not yet crinkled, appears to be available online, e.g., this link on aliexpress (1000mm 301 stainless for $4.46?  That would be prohibitively expensive for this kind of build!).

Installation - fail

As a final step, I tried putting the new honeycomb into the FSL hobby laser, only to find that it wouldn't fit!  That's because the pan heads of my #10-24 screws were standing proud of the frame piece, much moreso than the original rivet heads did.

Flattened screw heads

I got out my bench grinder and ground much of each screw head away.  I didn't do this with any particular precision in mind.  I just wanted some of the material to go away.

This is what a ground-down screw head ended up looking like.  This is one of the better looking ones.  Other ones were quite unevenly ground.  They might be harder to undo at this point, but I suppose there shouldn't really be a need to undo these very often, if at all.

With that, the new honeycomb could fit in the laser bed.

 Conclusion

It's totally doable to take a cheap laser honeycomb bed and cut it down to size.  For me the material cost was about

$20 for the 400mm x 400mm bed

$1.48 for the #10-24 screws and nuts

Some nominal filament cost for the 3d-printed #10-24 "nuts" that I made. 

One crappy #10-24 tap that was made from some metal that wasn't as strong as the L bracket metal. :(

It took me quite a bit of time to do this project, but in part that was due to nervousness, never having done it before.  Having done it once, I could probably do one of these in a few hours.

Things to do better next time

I think I would have a better design for my 3d-printed nuts.

I might consider unwelding the bottom rod.  Since there's so much excess rod material, I could chop off the ends of the rods rather than clean the welding material off.  And, having all four of the 4mm rods free would allow me to (A) make sure they're of uniform length, and (B) die-cut threads on their ends more easily.

In retrospect, I could have laser-etched measurement markings on the frame pieces.  That would have happened prior to any reassembly.  I suppose I could still do that.  One of the beautiful things about this project is that I can disassemble and reassemble the honeycomb now.

Full Spectrum hobby laser

 Full Spectrum Laser Hobby laser - alignment

I'm posting this after getting very far into reviving an old Full Spectrum Laser hobby laser.  It's 4th gen, I think.

In short, the machine was in rough shape when I got it, which also explains why I got it for a relatively low price.  Among other things, it had problems with its power port, input air line, and front panel switches.  It had good water cooling tubes, but no chiller.  The inside light wasn't working.  The locks were missing keys, but luckily they were all in an unlocked position.  There was no honeycomb.  It had no air exhaust pump and its rear panel for the air exhaust was missing, so there was just an odd, rectangular hole in back.

Most everything else turned out to be working well.  The interlock switches on the front and rear panels were good.  The power supply was good.  The driver board was working.  And, of course, the CO2 tube itself was still working.  I couldn't test any of this when I bought it, so I consider myself lucky.  I also found inside the "Ruler" for doing Z height focusing.

Texting, texting

My first test of the laser -- with safety glasses on! -- was to fire the laser using the "Text" button on the power supply.  Yes, it says "Text", not "Test".  That gave me confidence that the tube itself was operational.  I only did that after having thoroughly cleaned a fountain pump that I'd gotten at a thrift store, and set up a Homer bucket for cooling.

From a power standpoint, I run the laser, water pump, exhaust air pump, and nozzle air pump all off of the same power strip.  The water pump and exhaust pumps are not switched, so when the power strip is turned on, they come on immediately.  This is done for safe operating reasons.  The way it's set up, I can't fire the laser without the water turned on.  (And the gurgling of the pump reminds me to check to make sure it's ok before turning on laser power.)

Pushin' my buttons

While the initial test dot fired using the Text button, the front panel buttons for testing the laser did not work.  Eventually, I found that the old front panel switches were broken and weren't trustworthy.  The Laser enable latching switch, in fact, popped its back off when I depressed it, breaking the switch entirely.  (BTW, one thing I really don't like about this machine's design is that if something breaks and falls into the power area, there's a decent chance that parts will fall into the power supply fan cage.)

I bought a two set of 12mm / 0.5 inch push button switches -- some momentary, some latching -- de-soldered and removed all five old switches, and put in the new ones.

Along the way, I found that the Laser enable switch is set up in series with the door interlocks.  Also, the two laser test buttons are wired in series, which explains why the front panel says you have to press both to test-fire the laser.

The front panel's current control (precision pot) was wired to a three-pin header on the power supply.  There also was a dangling, single-wire connector coming from the control board, and it sure looked like it should go to the current control, but I left that alone.

Software, test cuts, and FSL to the rescue

I got great support help from Full Spectrum Laser, even though the machine is out of warranty.  They were most excellent to work with.  The machine is kind of a predecessor to the popular K40 machines of today, but its driver board is specifically for use with FSL's Retina Engrave software.  After some searching, I found a version I could use.  But it didn't really work.

It turns out that you have to get a specific version of the software to run a machine as old as mine.  Yes, you can get the old Retina Engrave if you look around a bit on their site, but you have to get the right version for this generation of the hobby laser.

Once they steered me to the right version, things were much happier!  I could start Retina Engrave on my laptop, connect to USB, and it would connect.  (If that didn't work, I was probably going to have to head down the path of getting a K40 control board and LightBurn.)  I could get it to test fire from the software.

I did some test cuts.  While I could get some things to burn, I noticed charring on the wiring of the red-dot aiming LED while I was doing some etching tests.  Yikes!

I powered things off and left the machine alone for a while, realizing I'd need to go through the alignment process.

It took some time -- really mostly to get up the nerve to do it -- but I eventually watched the YouTubes and went through the alignment steps.  The first mirror was very poorly set up.  Once it was in place, amazingly, I didn't need to do anything to the second mirror or third mirror.  But until the first mirror was fixed, I definitely could see that the laser was aiming at the back side of the cylindrical third mirror mount, and so it was probably bouncing off and into the wiring.

Prototyping the cow

I'm adding a piece to the FSL here as a test, seeing if this approach might later work on a friend's K40.  His laser has been aligned from mirrors 1 to 2, and 2 to 3.  But nothing is coming out the output tube.

Hence, the cow.  Here is the cow, designed and rendered Fusion.

The cow is basically a target holder.  It sits on the gantry in a somewhat predictable and precise way (yes, "somewhat precise"), and is meant to hold a piece of paper.  I don't have the original schematics/dimensions, and don't want to disassemble what's working.  So, I made the cow to help determine where the laser is actually hitting, relative to the gantry bar.

I built the cow with the intent of 3d-printing it.  Since 3d printing has its quirks with respect to dimensional accuracy and shrinkage, I made its legs wider than the bar depth (distance in Y) intentionally, thinking that I would shim it or drill a screw to get it to hold onto the gantry snugly.

The design splits the cow in two.  I used a few printed pins and holes to line up the parts.  I used 1mm chamfers to strengthen the joints where the pins met the left part of the cow.  

I subtracted the left side from the right side to create holes, and then widened those by an extra 0.3mm to allow for shrinkage.  Doing it as two halves made it easier to print flat, and helped avoid problems with supports.  Note that the "left" side of the cow has a rectangular area cut out.  That's there so a piece of target paper can slide in from the top after the two halves are joined.

 

In real life, I printed the cow using white PETG with 0.2mm layers.  The pegs had some extra gunk when done printing, but I could clean them up with flush cutters.  The two pieces could be joined after that with a satisfying squeeze.  I didn't bother to glue them.

This is what the cow ended up looking like, printed and joined.

Printed cow, PETG, 2 sides joined

As mentioned before, I made the legs of the cow intentionally wider than the gantry bar.  I chose to do that rather than spend time on a test print and making it totally accurate.  What would have been a big waste would have been making the legs too close together, and unable to fit over the gantry.

I found a random magnet that worked well as a shim, and that allowed the cow to fit snugly.

Random sheet magnet as a shim

 These are back and front views of the cow, mounted with shim on the gantry.

Rear view with shim in place (friction fit, not magnetically attached)

 

Front view, mounted on gantry

 

I cut out, very crudely, some 40mm-wide pieces of paper to act as a first target.

This is what the paper looked like in the head of the cow.

Time for an initial test fire!  Power dial 3 turns, quick hit on the test buttons, and, poof, we have a hole.

The next step was to build some actual targets.  I used Inkscape and threw together a page with 40 mm x 40 mm grids, with lines spaced 4mm apart.

The following are mountings and test firings.

Grid target 2, and grid hole 2:

Grid target 3, and grid hole 3:

Here are holes 2 (lower) and 3 (upper) for comparison.

In both cases, the target grid was flush against the left side of the mount.  With a little magic help from Adobe Photoshop Elements, I clipped out the Target 2 result above, positioned it, de-tombstoned it, etc.

End result from target measurements

The bottom grid line is at pix 834.  Next grid lines up are 756 and 678, meaning the photo is showing 78 pixels per 4 mm.  The hole center is at pix 688, so it's 10 pixels below the grid line.  That means it's about 4mm + (68/78) mm, computing out as (4.872mm) above the bottom.  Also, the hole is pretty much dead on 20 mm from the left edge of the page.

What's Next

Now that I have a pretty good understanding of where the hole is hitting, and because I know the laser is well aligned, I need to make a cat laser pointer that will point through the hole.  Ideally, I'll have a second cow-like mount, placed on the gantry closer to the x=0 point.

If I can make the cat laser's starting point close to same as the hole, shine the laser down along the gantry and through the hole of the target, it should go through to the output mirror and down.

That all should enable us to do 3rd mirror adjustments without actually using the CO2 laser.

Cube Pro Duo disassembly, part 1 - Escape from The Enclosure

I had the good fortune of finding a Cube Pro Duo for a low price, used.  The prior owner said the motherboard was blown, so it had sat unused for a while.  Even so, it wasn't all that dusty.  We lifted all 75 lbs of it into the back of my car, and off I went.

There are lots of videos out there about the Cube Pro Duo.  It's a big, heavy machine.  It actually turns out that The Claw was built using the top of this same kind of printer, so I'm familiar with its innards.  

I was able to slide it down a makeshift ramp to get it out of the car, and then put it onto furniture moving "glides" to move it into the garage.

The disassembly started with the removal of the build plate (just lift off), top panel (just lift up), and back panel (pull away -- it's held by magnets).  One of the magnets lost its adhesion, so be careful to make sure you have all of them.

My goal at this point was to just get the metal frame out of the enclosure.  I couldn't really address the electronics issues until I got the cube frame out.  (The power supply, motherboard, and other things are at the bottom of the cube frame.)

Step 1, of course: totally unpower/unplug the machine, and allow caps to discharge before proceeding.

I was able to tip the machine over on its back (onto a soft surface), and remove the screws holding the feet on the bottom of the enclosure.  The screws holding the feet in place also hold the cube frame to the bottom of the enclosure.  The feet themselves could just be pulled off after the screws were out, though one of them gave me a bit of trouble, because it had been mounted so tightly.  Fortunately, there was no glue involved.

This is the back side of the machine with the back panel removed.

There's a "Back, Inside Acrylic" panel (BIA) that houses clips for keeping sheathed cables and Bowden tubes organized  The BIA is bolted in place with a number of metal clips.  The clips then attach to the back edge of the enclosure.  I removed all those bolts that connected to the BIA, and loosened the bolts at the edge to rotate the clips out of the way.

One aspect of the Cube Pro Duo is that there's a lot of acrylic involved.  The acrylic parts can be broken pretty easily. 

I removed the screws at the ends of the metal bar that serves as a cross-brace for the BIA.

I started to remove the bolts that hold the BIA to the bottom posts of the steel frame, but realized it was too early to do that.

In order to get the back panel off, I would have to unhook all the connecting cables.  I started by pulling off the three plastic clips that held the Y axis cable sheaths in place.  These could just be pulled and wiggled off.  If they get hung up on the sheathing, you can reach in from the top and free them.

 Then, I removed the two screws and nuts that were holding the drag chain to the BIA.

I finally unscrewed the last bolts that were holding the BIA to the frame.  This would allow me to lean  the BIA outward and undo the remaining clips.  Or so I thought.  Be quite careful at this point.  It'd be really easy to get the BIA corner caught up somewhere, lean it outward, and crack the acrylic.

 

There is a small PCB at the top of the BIA.  I think that PCB is there just to hold a thermistor for measuring the enclosure temperature.  I started to remove the nuts/bolts that were holding it in place, but that was a mistake, so don't do that. 

Instead, if you lean the BIA outward, you can reach the clip that holds the signal cable to that PCB.    Unclip that.

The BIA is still not free, because it has clips that hold onto the sheathed cables and Bowden tubes.  I thought I could just pull out those clips in the same way as I had done for the little plastic clips earlier.  However, the Bowden tubes go through holes, not C clips, so to get this apart, I would have to thread those tubes through the holes.

The Bowden tubes were connected to a screwed-on ferrule along the Bottom Acrylic Plate (BAP).  Those would allow a proprietary filament box to feed acrylic into the tube from a known point.  The ferrules were zip-tied to the bottom, so I cut those off.  (Pardon the bad photography here.  The shadows make it look like there's a weird shape to the BAP acrylic.)

 I then threaded the tube+ferrule end through (below) the BAP, and then unscrewed the ferrule from the Bowden tube.  

I then threaded each Bowden tube through and out of the BIA clips.  (Note: even though it's a dual-extruder machine, it's set up to accommodate three extruders.)

There was a lot of bending and kinking of the tubes at this point, which was uncool.  I'm not sure if the bends were so bad that I will need new tubes whenever/however I reassemble it.

With the Bowden tubes out, I could gently pull the sheathed cables out of the clips, remove the clips themselves, and take off the BIA.

I wasn't sure what to to do move the frame next. 

The following things need to be done:

Where possible, unplug the stepper motors.  It helps in these steps to move the printer head carriage and Y gantry out of the way and it's best not to risk having current flowing back to the motherboard.  (In my case, assuming the motherboard truly was blown, I didn't really care, but did it anyway.)  It also helps -- but gets greasy -- to manually turn the Z axis threaded rod, thus lifting the build plate high and out of the way when needed.

Remove the upper bolts holding the power switch shield.  That shield holds to the side frame, and gets in the way of sliding the steel cube frame out from the enclosure.

 

Remove any remaining bolts holding the cube frame to the enclosure sides.  If I recall correctly, there are three on each side.

Remove the two lower bolts (one on either side) that hold the power switch guard to the enclosure side, remove the switch guard itself.

At this point, the frame is mostly moving separately from the enclosure.  There are just a few things in the way.

There's a USB and SD card connection on the left side at the front of the enclosure.  It's tricky to remove these.  They can't just be unclipped, because the upper board (USB) blocks removal of the lower board's (SD card) wiring.

What I did, which may not have been right: first, unclip the upper board cable.  Then, with some gentle nudging plastic, the upper board can be pulled inward from the enclosure side and removed.

This reveals the lower cable, which can be unclipped.  You can leave the SD card PCB in place.

 In retrospect, it appears that the whole assembly could have been unclipped from the side of the enclosure with the wires in place.  It looks like the original assembly would have had the worker push this into place, clipping onto the metal (elongated octagonal) clip. So, maybe by pressing gently on that top tab, the whole thing might have pivoted inward.

Finally, there are a few more cords to be unclipped, the main one being the front panel display.  The wire bundle that goes to it is neatly hidden in a wire management thing.  But to unclip the front display requires disassembly of the display.  If you're looking to use this in the future, beware of ESD.

This is a not-so-great picture of the front display.  It's a little LCD panel and a single button.

On the inside of that, I took off four screws...

and then, with the PCB exposed, removed the screws that held the PCB (and display) to the front plastic

and then I was able to unclip the wiring from the front panel PCB.

All the wiring of the USB/SD Card and front panel displays has to come out from a nice, cable management thing so that the cube frame can be removed.  I can't remember exactly how that part worked, and kind of don't care because I won't be reusing those electronics.

Success! I was able to slide the cube frame out from the back of the enclosure.  It was still quite heavy.  You can see here that I used a few pieces of redwood fence board scrap to support the back of the cube as I slid it out.  Otherwise, the weight might have crushed the bottom edge of the back of the enclosure. 

A note on lifting here, too.  I would typically lift the cube frame using the front and back upper bars.  Avoid lifting with the "Y" axis bars (left and right), because they need to stay lubricated and there are belts and such that you don't want to damage.

In the picture that follows, you're seeing the cube frame in the foreground.  The drag chain still holds all three sheathed cable assemblies and Bowden tubes.  The view is from the rear of the printer.  At the top left you can see the empty enclosure.

With all that done, I'm not sure what the resulting weight was, but it was still heavy.  I slid it onto a furniture dolly, and moved it inside for further work.