A Ming-Inspired Cabinet (80)

Click here to see the previous post.


During the interim in which I constructed and installed the Japanese ceiling, I moved the cabinets along in a couple of minor ways. I have been applying patination to the bifold door hinges, still in progress, and contracted with a local machine shop for the fabrication of hardware for the bifold door main hinging.

One thing did crop up while the woodwork for this cabinet sat, and that is a problem with the four plank shelves, three of which cupped. I guess that’s one advantage of proceeding slowly with a build, you get to see certain outcomes that otherwise would have occurred after the piece had been delivered to the client. Given the choice, I’d much rather the problem occur while the piece was still in my possession of course.

I took precautions, as usual, when dimensioning stock down to make those shelves, jointing and planing over several days, balancing stock removal on both sides of the blank, etc., and after applying 4 or 5 coats of finish to both sides, they were flat. Several weeks go by, and they had moved out of flat to an unacceptable degree. The cause, given that bubinga is generally a decently stable material, particularly when in quartered-to-rift grain orientation, as these shelves were, lays either in how the material was originally dried, or possibly in growth stresses in the tree that were slow to manifest. I tend to think it was mostly attributable to the drying, and therefore I won’t be buying bubinga from that supplier any longer.

It goes to show as well that solid plank construction is going to allow for a greater amount of movement than frame and panel construction. There’s a reason you don’t see much of it in the classic Chinese work. This aspect – the advantage of frame and panel over other methods -has always been quite clear to me, however I had thought that maybe for shelves it would be okay. In fact, making 3/4″ (19mm) thick shelves at 17.5″ width out of single boards might in most circles be considered quite luxurious as a use of material.

No such luck though, so I am fabricating new shelves and these ones will be, you guessed it, frame and panel.

I sliced up the original shelves, jointed and planed them down, so as to form panels:

That took next to no time at all, as they are ‘merely’ quartersawn bubinga.

For the frames, I had but little extra material from which to choose. In fact, all I had was a remnant flatsawn chunk from the middle of the original 16’x 52″ x 3″ slab that had jump-started this project many moons ago. While I have little use for flatsawn stock generally, by resawing this piece lengthwise I was able to obtain 4 sticks of perfectly quartersawn curly bubinga.

Curly bubinga was a harmonious match for the material used for the two principal horizontal frame and panel assemblies in this cabinet, which lie above and below the shelves themselves, so from that perspective curly bubinga was an excellent choice. What made it a less than scintillating choice however, was the fact that as curly material, normal jointing and planing was only going to get me so close to dimension, lest I risk too much tear out, and that meant that my milling machine was going to be the route for getting to finish dimension.

That meant that basic dimensioning was going to take a couple of days, not 15~20 minutes.

There’s something about having made well-fitted shelves, and then applied many coats of finish to a pleasing result, only to have to start all over again, and to boot be employing a different material and form of construction which was going to add many days of work. I had thought I would be moving into final stages with the door framing  – instead I take a step back and work on shelves again. I thought it over long and hard before committing to this course. I was initially a little grumpy about the extra work, however I am more than happy too knowing that the result will be better than previous. And that is what it is all about for me and my woodworking: a continuous striving for improvement.

So, on with the stock prep, day one comprising the resawing, jointing, planing and that left me with decently clean stock, albeit a little tear out here and there:

The stock is sufficiently oversize that the tear out is of no concern.

Then first pass with the milling machine on all the long sides of each of the 16 sticks:

I’m using a CMT top bearing bit with negative rake teeth, which all but ensures zero tear out.

Now working the edges:


In my humble opinion, so long as you are producing chips and not dust, all is good with the world:

After round 1, I had the sticks good and square – you can always get a sense of that by how well they stack next to one another:

Notice on the far right of the pile that there is a defect there. I had to patch that area, as I simply have gone through absolutely all of my curly stock. As with many other stages of this job, i have no spare sticks to work with, so layout and cut-out mistakes must be avoided at all costs.

During the first day of action with the milling machine, cutting left to right to left to right…, I noticed towards the afternoon a certain odd noise start to develop with the machine. This noise grew worse over time and it associated to the power feed. Every time I shifted direction, the noise intensified. I started to get this feeling of dread come over me, akin to the feeling you get when you start to apprehend something is decidedly wrong with the internals of your car’s engine. As anyone with experience knows, you don’t pull an engine out of a car and strip it all apart just to replace a single valve, or pushrod seal, or piston ring. When you take an engine apart, it is generally wise to rebuild the whole thing. It’s cheaper in the long run.

I was worried that the gearbox on my Zimmermann mill was developing problems. That’s sure what it sounded like, and I could see that to get at the gear box was not a simple job, and may even mean stripping the mill down just to get at it. Besides the timing being lousy for this, I am not in a position to have the mill down for long periods, let alone pour money into it.

I didn’t sleep well last night, and was feeling a bit on the mopey side when I went to the shop this morning. I was fearing the worst, but I hoped, with another round of dimensioning to do on the mill that the gearbox would at least do me a favor and hang in there until the task was done.

The next day I started milling, and sure enough the noise was there right away. I couldn’t seemingly wish it away after all.

But this time I got in a little closer with my head to the gearbox to see if I could more precisely localize the sound, and discovered, to my considerable delight, that the noise wasn’t coming from the gearbox after all, but from the motor and pulley drive set up on the outside of the gearbox. I removed some allen cap screws and got the motor cover off, and sure enough I found the culprit: the motor’s drive pulley was coming loose. Whew! That was good news indeed.

Curiously, the square key which ties the pulley and motor output shaft together was absent. I’m not sure if it was missing from previous repair or whether it had simply come out at some point and worked its way past the cover and to the floor (somehow). Whatever the case, a new square key and set screw were needed, so off to the local hardware store I went. Now, the correct size key was 5mm x 5mm, 25mm, however, this is the US after all and selection of metric parts remains less than optimal at most hardware stores. As expected, they had no metric square keys. I was able to find 3/16″ (4.76mm) square, 1″ (25.4mm) long key stock, so that would have to do in the meantime. Better than nothing.

I fitted the square key and new set screw to the pulley, and put things back in order. It worked, but not silently, and after a while started getting noisy again. I took it apart again and realized that after the old key had fallen out or been lost, and the pulley gradually worked its way loose by grinding the set screw away, that the inner bore of the pulley had been sufficiently worn to now be a poor fit. Correcting the problem, however was, at best, a job for a machine shop. I wasn’t interested in yet another delay however, so I improvised, taking a plastic dado set shim, wrapping it around the motor output shaft, and then slipping the pulley back on over top:

Surprisingly, this worked like a charm! The pulley spins without wobble or any noise, stays put, and all is well.

Now, I’m not the sort who leaves a ‘spit and bubblegum’ type of repair, as this one surely is, until it next fails. I’ll be taking that pulley to the machine shop soon enough and get it properly repaired, but for the interim, my mill will be usable and I can get the job done.

Round 2 of milling could then  proceed, and a whole lot more quietly I might add. Here, after having worked all the broad long faces, I am gang-cutting the edges of half of the short pieces:


They were done after not too long:

The stock is about as square as can be, and within a few thousands of dimension in every axis. I’m always pleased to obtain results like that – it’s like putting the dart into the bullseye when you actually meant to do so….

The frame stock now prepared, it was time to turn to the joinery work. Options abound for corner joints. I’m going with a form of mitered, tongued, and tenoned corner joint with shachi sen. I last did this variation of that joint on a Walnut Vanity I made about 10 years ago (link), in that case with both a tabletop frame containing a tongued panel, as well as the support frame below it.

I’ve cut this type of joint in myriad ways, and this time I decided to have a go with my mill. I thought it might offer some advantages, although if I had certain (other) tooling on hand, a preferable choice might be to process the joints partly on the shaper’s sliding table. You work with what you got, be it hand tools, or sliding saws, or sliding shapers, or milling machines. I lucky enough to have all of those pieces of gear and decided try the mill this time, and meanwhile dream about what sort of shaper tooling I’d like to acquire sometime down the line….

Here’s the jig I came up with on the milling table:

A couple of screw adjusters allow for fine control of position, and the slotted table and beefy hold downs mean that the material is securely held:

So here’s a rough step-by step accounting of how I processed the cuts on the tenoned sticks. First, I ripped the cheeks using a quick-built MDF jig of scraps on the table saw:


That left the following result on both ends of the short sticks:

Then the chop saw, with depth stop engaged, to trim the waste off the miters:


Though both table saw and miter saw could have been used to go to a finial surface with those same cuts, I don’t rely upon those tools for that in this case, I merely rough out with them.

After the chop saw work is done, the mitered tenons are defined, but not yet to their lines:

Then into the mill with the same pieces.

After having taken some time to calibrate the cut on a test stick, and I deck the tenon cheek and trim the mitered shoulder in one set of passes:


The stick is then flipped around to the other side of the jig and the process repeated on the other side, with the slight difference of using a climb cut on the abutment.

Result is a tenon within a few thou of dimension for thickness and with cleanly cut mitered abutments in a common plane:

The process was repeated until both ends of all the short sticks was brought to the same stage.

The next step, using the same fixturing set up in the mill but with a different cutter, was to cut dadoes on the mitered abutments. That required a change of tool holder, to a slimmer unit which accepts ER32 collets:

I also trimmed down the positioning blocks to clear the collet nut. Again, a few cuts were necessary to calibrate the cut depth, and after that the processing could proceed fairly rapidly.

That said, at the end of today’s session, I had just the one stick through the dado-ing stage, and the rest lined up and ready to go:


It shouldn’t take more than 15 minutes to complete the rest of the dadoes on all those sticks.

A last look at the dados and mitered abutments, all cut fairly cleanly it seems to me:



So far this method of cutting with the mill seems to be working well. The proof of the pudding, of course, will come when the mating piece is done, whereupon I am looking to obtain fits without having to spend time fettling the joints. We’ll see….

Thanks for visiting the Carpentry Way. The 81st post is next.

12 Replies to “A Ming-Inspired Cabinet (80)”

  1. Chris,

    Is the hole in the motor axle threaded? Because I would expect a washer held down by a bolt to prevent axial movement.

    I wonder if repairing the pulley is actually worth the effort. It might be cheaper to have a machine shop measure it and turn a new one if you can't get a spare part.

  2. Roland,

    No, there's no threaded hole, but there's a 6mm set screw that keeps it from sliding off the shaft.

    You're right, a new pulley copied from old might be cheaper- I'll see what the machine shop recommends.


  3. Chris,
    It is not unusual to use shim stock as you did, (metal as opposed to plastic). Provided the proper key is used. and the hole in the pulley is “round”. If the “hole” in the pulley is not then it can be enlarged, and a sleeve/bushing made of thicker shim stock, cut to allow for the key. Much, much cheaper than making a new pulley.

  4. Joe,

    that's how I envisioned the repair being done as well. One day I hope to acquire a lathe so as to be able to take care of jobs like this, instead of having to go to a machine shop. The first piece of shop equipment I gained any significant experience on, when i was around 12, was a metal lathe.


  5. Wow, I know that feeling of dread when you are sure there is a serious problem developing but you cant stop it. More than once I have been through that and then, when I looked more closely, happily discovered a very simple fix. Catastrophizing is brutal, but hard to get around sometimes.

  6. Matthew,

    thanks – you see it exactly. In 'preparing for the worst while hoping for the best' I find it easy to imagine the worst outcomes too. Lucky that it was a minor issue this time. I know that the clock is ticking however, and the mill has various issues which can only be addressed by a strip down and rebuild, including scraping, quill and spindle rebuild, etc. It's not surprising given that the mill was made in 1971. I have no idea if it has ever been rebuilt. Rebuilding can be expensive, so I am not exactly chomping at the bit to get going, but at some point there will be no choice. The Wadkin saw needs similar attention, but will be less costly to deal with so likely it will be done first.


  7. Given that this mill was built for milling steel, I'd say that your work milling wood and occasionally brass should not be very stressfull for it. And seeing the pictures of the parts you make on it, it doesn't look like there's much wrong with it for the work that you do! Sure, if it was rebuilt you could probably make deeper and more aggresive cuts with it. But is worth the cost in your case?

    OTOH, it *is* possible to restore mills like these to factory new condition. You can find video's on YouTube from firms specialing in that kind of work. I don't think people will even bother rebuilding the current generation of CNC mills even after 20 years.

  8. Roland,

    thanks for the follow up. Rebuilding is less about making the machine able to take deeper cuts – not really an issue at all, as you note, given that I primarily work wood – it's more about restoring factory accuracy to the knee and saddle travel in the three axes. After 45 years of use, there is wear in the machine ways which make the cutting a little less perfect than it could be. This show up more on longer and wider pieces.

    There are several other issues though. The quill on the machine is very sticky and the quill lock does not work, and I would very much like to get the main spindle working as new. This means a bit of work by a spindle rebuild specialist.

    Also, the z-travel motor sometimes gets stuck in raising (requiring a sprint over to the wall to shut the main power off) and there is some sort of wiring fault in there, not in the switch unfortunately. It's a scary thing when that happens. Probably the machine electrics need a thorough going through.

    Also the table limit switches do not seem to work in the z-axis – again, electrics.

    Also, the machine leaks more way oil than seems acceptable, and I would rather not keep pouring more oil into only to see it come out and onto the floor. I suspect some seals and wipers need replacing.

    The engagement of x-travel is starting to slip at the hand wheel sometimes.

    Also, I would like to add DRO for knee Z-travel, for rotation if possible, and for quill travel. And place the readout on a proper articulated arm instead of the plywood upon which it is currently mounted. I also would like to re-establish the machine's lamp.

    Also, the high speed spindle is a bit useless to me as it uses a proprietary Zimmermann spindle interface and Zimmermann no longer makes this interface so my tooling options are very limited. Again, a visit to a spindle specialist is in order to change to something more modern, or else I may as well pull the high speed head off the machine altogether so I can push it closer to the wall and save floor space. I'd really like to be able to use the high speed head though….

    In short, there are so many little issues that need attending to with this machine that I do contemplate a full rebuild at some point. Better the machine be down just once, than several times.

    I am now learning about the use of Moglice to restore ways, as opposed to full scraping, so that is certainly a less costly option. I'll be trying that out on the Wadkin first.


  9. Chris,

    If the ways have too much play, didn't adjusting the gibbs help?

    From the sound of it, a rebuild is definitely in order. The insulation on the wires is probably past it, especially on wires that move. Hanging power-feeds and inoperative limit switches do certainly not make for peaceful milling!

    Repairing with moglice is an interesting concept. It seems to be a filled epoxy resin. I have some experience with casting filled epoxies, and it is not as easy as it seems. It can be hard to avoid air bubbles, especially in higher viscosity resins. We tend to place mixed resin in a vacuum chamber (less than 20 mbar) to degas it. This can make the mix foam up to five times its original volume! Making epoxy stick to cast iron that has been soaked in oil for decades seems like a tough job in itself. You might end up having to rough up (basically destroy) the way surface to make the resin stick. Such repairs should be done in the summer, since resins often won't cure properly when it is below room temperature.

    My biggest question mark is how to get a nice flat surface? You'd probably have to use a pretty flat and smooth surface as a “mold” to press into the epoxy after it has been put on the way surfaces. To make matters worse, you have different surfaces in one way that need pretty small geometric tolerances in flatness and parallelism and angle. That in itself makes such a counter-mold a very expensive thing to make. Unless one of the two way pieces is in good enough shape to be used as the mold.

    Over the years I've made epoxy putties filled with minerals, metal powders and even ceramics such as SiC. For sure they are harder and more wear resistant than unfilled epoxy, but not in the same league as cast iron or even brass. I would expect epoxy-coated ways to be pretty vulnerable to metal chips and impact damage.

    All in all, I wonder if it is really a cheaper alternative. A comment I found on a machinists forum might be of interest.

  10. Appreciate the detailed reply Roland.

    I just checked the main table rise/fall in x-direction travel yesterday, and the amount of deviation over 750mm was a bit over the recommended (Connelly's Machine Tool Reconditioning book suggests a maximum of 0.001″ over 24″). However that is but one test for alignment, and doesn't tell me a whole lot frankly. I was pleased to find it wasn't as out of whack I had feared.

    The z-travel remains a little bit stiff. I have played around with the gibs a bit, but am apprehensive about introducing too much play be gib loosening. The assembly of kneed and transverse tables is so heavy that I can't readily test to see if there is significant plat in the assembly.

    As for the Moglice, from what I understand it has excellent wear characteristics. Your question about how to get a nice plat surface is one I had as well. it seems that you still need to grind or scrape one of the surfaces into flatness, parallelism and freedom from twist, before Moglicing the mating surface. Of course, one prepares the larger surface in that way and applies the Moglice to the smaller part which mates to it. So, using Moglice does not eliminate scraping entirely, but it does cut that work about in half.

    I had come across that thread on Practical Machinist forum, of which I am a member, and yes, duly noted. One factor that the fellow does not mention, and this is natural enough given he is a scraper hand and all, is the cost of transporting the machine to a facility where it can be dealt with properly. It's not going to be easy or cheap for me to move a 2200kg machine back and forth. Probably looking at $1700 minimum to have it rigged and moved back and forth.

    There is a company down in Avon Connecticut which comes to the machine owner's facility, assesses the machine and teaches you how to do the Moglice work (http://www.mogorehab.com/repair.php) so that is a possible direction in which I may investigate.

    Rebuilding a milling machine by scraping is a costly procedure, and as soon as I take my machine out of my shop and down to a machine rebuilding place, I am best to have them do as much as possible. I just can't afford that right now. I am hoping to find some sort of in-between solution, where i find a way to bring the machine back to spec by having some scraping and or Moglice work done, and I handle the bearing replacements, re-painting, and re-assembly.

    None of the above is a pressing need, but it is on my mind.


  11. There might be an easier way to make patterns for applying the moglice. Have them milled out of tooling board. Suitable grades are dimensionally stable, easily machined and can be milled to an excellent surface finish on a high-speed machine. We've used SikaBlock M940 (density 1.2 times that of water) for building models to laminate molds on. Of course you'd need to apply and polish a couple of layers of release wax before using it with epoxy! But then you have almost a mirror finish.

    The material costs of tooling board are higher than the cheapest aluminium alloys, but this is offset by it requiring much less machining time and a much lower weight.

    A milling shop that makes foundry patters or models for composite moulds should know these materials and would be able to give you an accurate quote.

Anything to add?

error: Content is protected !!
%d bloggers like this: