I wired up the motor controller to the vibratory deburrer today.  Since I had already bench tested the motor and controller, I had no real concerns other than making sure I was using the right terminals on the controller.  I put a tub of ceramic media with some aluminum parts into the unit, and gave it a shot.   The media swirled around the tub much like it had in the suspended version, which was something I was hoping to eliminate.  After realigning the weights to 90 degrees apart, I found that the swirling was eliminated if I ran the motor at about 60% of full speed.  I covered the tub and left it run – thankfully the noise is a little less than the suspended version (but just a little).  About 45 minutes later, I heard what sounded like a muffled crash, and ran downstairs to see what had happened.  The bucket frame was leaning way over, and after powering off the motor controller, I had a look at the damage.

One of the 4 springs that support the tub frame had snapped from the vibration.  As the frame keeled over to one side, the hose that coupled the motor to the weight shaft was twisted and sheared off.

At this point, spending the money for a proper industrial unit is looking more and more attractive.

I stopped at the hardware store hoping to find some suitable rod that would fit the undersized bearings of the previous post. Naturally, the rod sizes they stock jump from 1/2″ to 3/4″ with nothing in between. While pondering the unsavory prospect of shaving down the diameter of a stainless shaft on the lathe, I wandered around the store aisles aimlessly. This is something I tend to do more often than not when I enter any hardware or mechanically inclined store – seeing the blue building insulation makes me think that I should build a hot wire foam cutter for R/C airplane wings, while the plumbing section reminds me that I’d like to build a pneumatic launcher with a Rainbird sprinkler valve as the core. Thank heavens the Boeing Surplus Store is now closed – if I ever would have entered their doors, I never would have left.

In my meanderings, I found a nice large 5/8″ diameter bolt that slipped through the bearings nicely (there’s a little bit of play, but I don’t think it will be an issue).  Unfortunately, I found that I had been a bit enthusiastic in welding the weight plates:

Fortunately, I was able to use an endmill to break through the weld and free the screw.  I did a test assembly of the bearing unit, and things looked pretty good:

I then shaved down the threads to about 1/2″ diameter so that I could use a piece of 1/2″ ID hose to couple the bolt to the motor shaft.  I also had to round the outer corners of the top weight to keep it from hitting the frame members.  After attaching the weight and bearing unit to the old motor mount plate, I reassembled the bucket frame and installed it on top of the springs.  I did have to reduce the spring height by 2″, but once assembled, the whole unit ‘felt’ about right.  I finished by adding the hose coupling and adjusting the motor height.

I managed to locate a suitable DC motor and controller for the vibratory deburrer, and then set about designing a proper lower ‘chassis’ for the system. I’ve been trying to make all equipment wheeled as much as possible for easy movement around the garage or basement, but I also needed leveling pads on the unit (I want it to shake and rattle, but not roll). The bottom chassis looks rather small, and I hope that it won’t have a tendency to fall over as a result.

I welded pipe nipples to the chassis and stuck what I hope will be appropriately sized springs onto them.  The legs of the bucket frame now have pipe nipples and washers welded to their bottoms – they should slip into the springs from the top.  The motor location is adjustable back and forth by a little bit so that it can be aligned with the weight shaft (which will be bolted to the previous motor mounting plate on the bucket frame).  The weight shaft mounting has been a bit of an annoyance – I started with a hunk of aluminum bar stock and bored a pocket on each side to hold a 5/8″ ID bearing.  A short shaft then passes through the bearings and will have a weight plate on either end of the shaft:

I’ll be able to adjust the vibration amplitude by simply aligning the plates – 180 degrees apart for ‘purr like a kitten’ all the way up to 0 degrees apart for ‘funny, I don’t recall there being any fault lines around here’.

Although the bearings are 5/8″ ID, I’ve had a heck of a time trying to coax any 5/8″ rod through them.  I finally decided to see if thermal expansion could assist, so I threw the rod in the freezer and stuck the bearing bar in the oven on low.  Half an hour later, I pulled out both parts, tried to slide the rod through and…

The rod managed to get perhaps 1/16″ into the bearing before thermal conductivity stepped in and ruined the fun.  It turns out that perhaps I should have been a little more careful when selecting bearings – these are +0″/-0.0003″ on the inside diameter.  Either I’ll need to find some undersized rod (generally $$$) or I’ll have to make a stab at turning down this existing rod by just a couple thou.

After a brief vibratory stint on some Roundhead bodies to test out the efficacy of my thread insert plugs during the ceramic deburring cycle, I tried turning on the motor once more only to be greeted with a loud angry hum from the motor but no movement (and an ominous dimming of the lights).  A bit of investigation showed that a wire nut had fallen off of a pair of wire ends, and they had probably contacted the steel tubing that served as the frame of the unit.  I replaced the wire nut after checking for any other signs of damage, and tried plugging in the motor once more.  Again, the lights dimmed and an angry hum flowed forth.

So I started disassembling the deburrer frame so I could pull the motor off of its mounting plate.  Once I had the motor entirely free, I tried plugging it in again, and it spun up like a champ.  I figured I must have knocked some bit of debris loose or something, and re-assembled the frame.  Wouldn’t you know it, I tried plugging it in once more, and again got the ‘growling of extreme displeasure’ that unhappy motors are so very good at making.  At this point I surmised that the eccentric weight that I had bolted to the motor shaft (which is what provides the vibration) was the culprit, and that the motor would spin up only when entirely unloaded.

I asked one of our EEs at work about the issue, and he indicated that single phase motors are generally an annoyance.  I mentioned that the motor had a bulged cover on one side that probably housed a capacitor, and he said that this would be the start capacitor.  Furthermore, toasting this cap as a result of the dangling wires was certainly a possibility, and that a slighty loaded shaft could be just enough to keep it from starting without the cap, giving the angry hum.  “Pop off that cover, and you may find the gooey innards of the cap coating the inside”, he suggested.  However, removing the cover revealed a fully intact, apparently functional capacitor (crude tests with a multimter indicated that yes, it was indeed holding a charge).  I brought the motor in to work and disassembled it on a bench (with the help of the EE, who had once worked for Leeson Electric, and hence knows a thing or three about motors).  Everything inside looked just fine – we couldn’t see any signs of trouble.  I re-assembled the unit and tried powering it up.  As before, it would spin up happily when unloaded, but when I held my foot against the shaft and tried turning it on, I got the angry hum.  The EE surmised that the motor may very well have a bad winding – using a clamp-on ammeter seemed to indicate that the motor was drawing quite a bit of current at idle.  To fix the winding (if even possible for one of these cheap overseas made motors) would almost certainly be more than simply buying a new motor.  Off to Ebay I went, and found a beauty of a 1/2HP Marathon for much less than I paid for the 1/3HP Worldwide Electric.

Note the internal plate indicated by the red arrow – I’ll mention this in a bit.  The Marathon motor (manufactured for Graymills for use in a pump) was pleasant to hook up, as the connections are all spade terminals inside a cover plate on the tail end.  Despite being more powerful, the motor is actually smaller in diameter than the 1/3HP I had used.  I assembled the vibratory deburrer with the Marathon, and set the whole thing going on a Saturday.  A few hours later, the noise in the basement ceased, and I wondered what the heck had happened.  The motor case was really really hot, so I let it cool off for the night.  The next day, I tried spinning the counterweight by hand to see if perhaps it had seized or something, and I got a nasty shrieking squeal in return.  Great, “another blown motor”, I thought.

I disassembled the deburrer and had a look at the motor.  Thankfully the squeaking wasn’t a failed bearing as I had feared, but was the internal fan scraping against a sheetmetal plate of some sort (the thing pointed out by the red arrow in the above pic).  I tried using a screwdriver to push the plate back down towards the windings (away from the fan), but this was near impossible to do correctly, as there were vents on only one side of the housing, and I managed to tip the plate even further.  I disassembled the motor at work and carefully pressed the plate down into place (our EE said I could probably remove the plate altogether, as it was probably just for splash protection, but it’s captive due to a pressed bearing that I didn’t care to try removing).  I’m wondering if perhaps the excessive temperature walked the plate out of place and up into the fan.  Sticking a desk fan next to the motor seemed like a good idea, though I really needed to keep the motor from overheating in the first place.  Excessive slip was probably the cause of the heat according to the EE, which meant I was simply trying to use too heavy a weight.  Why I didn’t have this issue on the 1/3HP motor is a mystery to me, but at any rate I removed the hunk of steel from the weight.  I immediately noticed that the unit was a good bit quieter, and the media hardly moved anymore.  Well, at least the motor wouldn’t be turning itself into a puddle of slag while I come up with a better system.

I also noticed something slightly disturbing – some nasty chain wear.  I guess this is the cause of the sprinkling of metal dust that appears on the tub cover, and indicates that a top suspended deburring system simply isn’t going to work very well.

I then have two big changes that I intend to make to the deburrer – switching to a DC motor and going to a standard bottom supported tub rather than a suspended tub.  Using a DC motor will allow me to vary the speed with a motor controller, and I also won’t have to worry about excessive slip causing increased current consumption and massive heat buildup.  Even better, I can vary the motor speed in order to best match the load’s characteristics and find a sweet spot of maximum amplitude.  In order to support the unit from beneath, I bought some beefy spring stock and will have to come up with a base of some sort.  My plan is to also detach the motor from the tub plates, and have the motor attach to the new base instead.  Then I’ll use a flexible coupling (a piece of hose, actually) between the motor shaft and the shaft of a weight system.  Not only will this be kinder on the motor’s bearings, but the motor will no longer be useless dead weight in the vibrational system.

My homebuilt vibratory deburrer is still a far cry from a commercial unit – I went to the local machine tool tradeshow this past week and gazed longingly at the big industrial units on display there…  until I found out the prices.  I’ve got maybe $200 into my homemade version (a really good scrounger would have been able to do it for far less, I’m sure), which is still a good order of magnitude cheaper than one of the biggies.  Still, it leaves a lot to be desired – among its deficiencies are poor media movement, extremely long cycle times (probably related issues), splashing water in ‘wet’ cycles, and acoustic exuberance (“noisy as all hell”).

I also don’t get much ‘scrubbing’ action on part surfaces – while burrs and sharp edges do get knocked down admirably (though needing a bit of time), surface scratches and tiny dings still remain and I can still see the tool marks of the lathe tool on Roundhead bodies.  I’m seeing if using a Scotch-Brite wheel on the bodies will help remove the dings and allow for a better surface finish in the deburrer.  Well, calling it a Scotch-Brite wheel is pretty generous at the moment – since I wanted to try something ‘right now’, I took a pair of 3M abrasive nylon pads, tore a hole through their centers, slapped them onto the shaft of my grinder/polisher and had a go at it.  Not too bad, though I’ll have to get a proper wheel soon, as the pads wore down quickly.  It remains to be seen if this Scotch-Brite pre-treatment will improve post-deburred surface finish.

As mentioned, one issue with the vibe deburrer is that it flings droplets of water upwards and outwards every so often when running wet deburring media.  I had naively thought “eh, it’s just water, and the basement floor drain is like a foot away” when originally making the deburrer.  I had intended to put a cover on it (honest), but there were no suitable covers for the laundry tubs I used.  So I ran it as is, which wasn’t too bad at first.  Sure, the supporting beam, chains and basement floor just got a light sprinkling, but the problem is with the suspended microscopic grit and removed metal suspended in the water droplets.  When the water evaporates, it leaves a gray coating of grime that leaves your hands with that sensation of “eyuugh” when you rub it between your fingers while running for the nearest sink.  I’m assuming it’s like dried slip, but worse.  So it was with this in mind that I created a cover.

Nothing fancy, but it turned out nicely.  I used some cheapo plywood and 2″ thick foam insulation from Home Depot, cut into 24″ sqaures and then ‘assembled’ with some 3M spray adhesive (I doubled up the foam for 4″ of thickness).  I then drew a 20″ circle on the sandwich and ran it through the bandsaw at work (blade must have a massive weld on it – note the vertical lines around the perimeter of the foam caused by the weld that closes the loop of the bandsaw blade).  Topped off with a door handle, and beveled the edge of the foam a little to fit the tub.  Now the nasty grimy droplets will stay put (note my attempts to wipe away the gray grime from the edge of the tub).  I had hoped that perhaps the cover might abate the noise a bit, but no such luck.

The second improvement made regards the media flow in the tub.  The ceramic media that I deburr with mostly just spins around the tub, like the water in a draining sink (which really doesn’t provide any deburring action, and is wasted movement).  Beyond this, the media moves in a toroidal fashion – media moves downward around the outside edge of the tub, and also downward in the very center (naturally the media must come upwards between these two locations).  I mentioned in a previous post that parts tended to pool in the center, and long parts like the Roundhead bodies would stick halfway out of the center, contacting media on only one end.  To counter this, I figured it was a good idea to try to fill that center area with something else.  I filled a Gatorade bottle up with water, and placed it in the center.

Incredibly, it worked!  I no longer had parts pooling in the center – the movement of the media kept the bottle in the center, and the actual parts were free to circulate.  This explains why all circular commercial deburrers I’ve seen have a cone or cylinder in the center of the tub, or a donut shaped bowl – the center is simply a dead zone in all circular tubs.  I’ll have to see about modifying a laundry tub with a bottle or PVC pipe in the center, and perhaps some ribs on the outside edges to impede media from swirling when it should be bouncing.

The third improvement for this post regards not the deburrer itself, but preparing the parts for deburring.  A problem that I had with the first batch of Roundhead bodies was with the threaded holes for the thread inserts that need to be screwed into place (think Heli-Coils, but better).  The outer end of the threaded hole was getting peened over by the ceramic media, and I had to use a tap to chase the threads.  This was less than ideal, as the threads are fine and very shallow, so I had to use a bottoming tap.  Trying to get a tap straight on a cylinder (even with existing threads) is not exactly easy, and I’ve ruined several bodies in the attempt.  One thing I discovered was that my thread milled threads weren’t actually large enough in diameter to accept the thread inserts, so I changed the the CNC program so that the thread milling operation was technically 0.006″ over the nominal 1/4″.  At least now the inserts fit in place without needing the threads chased, but the peening issue remained.

The above is an admittedly poor photo, but you can sort of see how the top edge has been beaten down a little bit, squashing the thread.  I figured the best way to prevent this would be to use a plug of some sort in the threads.  Unfortunately, there are no set screws available in the 1/4-36 thread needed (that I could find, at any rate).  I guessed using the thread inserts themselves as a plug would be the best method, but the centers of the inserts (which have an 8-32 thread) needed a plug of some sort themselves.  An 8-32 set screw was the logical choice, but the trick was in securing the two together.  I considered using Loctite, but I wasn’t sure that even red Loctite would hold up to the torque of really cranking an insert into place.  Plus, I didn’t have any on hand.  I asked Frankie about soldering the set screws into place, and he said that for stainless, Stay-Silv 45 and black flux would be good to use.  However, he noted that on stainless, silver solder doesn’t ‘wick’ into joints very well.  This worried me, as the only way to really get a solder joint between a set screw and a thread insert was by getting the solder to wick between the two.  I tried heating a set screw and thread insert with a propane torch and applying some rosin core electronics solder to the joint, but the solder simply balled up and made no attempt at capillary action.  I figured I wouldn’t have much improved luck with the silver solder.  While idly torching the metal to a red hot state, I suddenly had a thought – what if I could forge the two together somehow?

The idea is simple enough – I threaded the insert onto the ‘head’ end of the set screw so that perhaps .010″ of set screw stuck out past the end of the thread insert.  I then heated the unit up to red hot while it sat on the anvil of my bench vise.  Once sufficiently heated, I smacked the tip of the set screw a few times with a hammer, squishing it slightly and binding it to the insert (while hopefully keeping the insert itself dimensionally unchanged).

The resulting plugs seem to do the trick – I’ve inserted them into several bodies, though I still have to run them through the ceramic media.  The set screws on a few plugs will spin with enough torque, but a little more heat and hammer should fix that.