This post has taken months
to make and way more money than I'd care to admit. I've been tumbling puzzles for about 10 months now and a few months back I decided to take video of the entire tumbling process, from start to finish. I wound up with almost 3 hours of footage. Unfortunately as I was taking the video I was learning more about the process. I've learned so much more since that footage that I no longer feel it is a good representation of the optimal way to tumble. Instead of taking another 3 hours of footage, I decided a detailed post with photos would be a better introduction to the subject.
When I first decided to try to tumble Shapeways WSF (SLS nylon) parts I had a few preconceptions.These are FALSE:
- Nylon is softer than rocks so therefore it wears faster in a tumbler
- Coarse grit is bigger and sharper and therefore wears parts faster
- As with tumbling rocks, you need to use a wet grit slurry
- A soft media like corncob or walnut shell is enough to polish nylon
- You can significantly change the tolerances (fix a very bad print) by tumbling
I will spare you the details of all of the assumptions, preconceptions, and hypotheses I've had with regard to tumbling. Over 10 months of guessing, testing, experimenting, and observing I have come up with a model that seems to work well.TUMBLING MODEL:
In brief, the model I use to understand how parts will wear in the tumbler treats the grit as a STATISTICAL PROCESS
which behaves as a FLUID
that has a TEMPERATURE
and a PRESSURE
. This is just a metaphor and it doesn't perfectly match with the strict physical definition of these terms but I think it's close enough so allow me to explain. I DO NOT MEAN LITERAL TEMPERATURE AND PRESSURE.GRIT FLUID TEMPERATURE:
The "temperature" of the grit is essentially the number of times per second that a given spot on a piece collides with some grit. You can "see" the temperature of the grit by how fast it is moving. The temperature is not uniform either. Ridges (edges) and other structures that stick out on pieces are "hotter" than other places in that they collide with grit at a faster rate. A collision won't necessarily wear down the piece being tumbled.GRIT FLUID PRESSURE:
The "pressure" of the grit is how hard they collide. Big heavy grit (coarse grit) collides with more force and big heavy pieces being tumbled push against grit more force. The harder grit collides with a piece the more likely it is to wear down that piece. The pressure is not uniform. Irregularly shaped pieces that have parts of them that stick out tend to focus their mass onto a small point dramatically increasing the pressure at that point. This means the more massive a piece is and the more its mass can be focused the higher the pressure at the focal point.
What wears pieces down is PRESSURE * TEMPERATURE. You need a lot of high-force collisions to wear down a piece. Unfortunately, it isn't quite this simple. Because neither pressure or temperature are uniform, pieces do not wear evenly. Different pieces and different parts of pieces wear at different rates. From what I have been able to determine, temperature is the biggest source of uneven wear. If you reduce the temperature as much as possible you can get mostly even wear (it just takes a very long time).
So how does this translate into grit?MEDIA:Coarse grit:
Coarse grit is made up of very large pieces. Because there are so few pieces of grit they have a low temperature. Because they are so big and massive they have a high pressure. Through a lot of trial and error testing, I have determined that coarse grit actually wears pieces slower and more evenly than fine grit.Fine grit:
Fine grit has a huge number of very tiny pieces. Because there are so many pieces they collide with parts very rapidly and have a high temperature. Even though they are very small, most of the pressure of the collisions comes from the size / shape / mass of the piece so while fine grit itself has a low pressure, it still wears pieces very fast. Because of this fine grit wears pieces very unevenly. Fine grit does have the advantage of producing more smooth results though.Porcelain spheres:
These spheres have an incredibly high pressure and an incredibly low temperature. Because of this they tend to just wear the already loose portions of the parts and smash down microscopic bumps into a more flat surface. Porcelain wears parts so slowly that it is only good for making them shiny, not for wearing them down. Because the spheres are so big they often can reach some parts of a piece and so yield uneven results. Generally the outer, visible surfaces of the parts are reachable though so they are able to do their job where it counts. You mix the porcelain spheres with enough water to slow them down and allow them to form a toroidal vortex. Without water they bounce around like crazy and the pieces tend to "float" on top. With water the process slows down and the pieces are able to submerge and move inside the media.Other stuff:
You can use other stuff mixed in with the fine grit such as steel shot or glass beads. I've tested this stuff a lot and their effect seems unpredictable. Both seem to increase the pressure by smashing grit in between the piece and something heavy. The actual effect on wear evenness has proven quite unpredictable though so I don not recommend using other stuff in with the grit. The cases where you would want to use this sort of mixed media is when you're tumbling parts that aren't visible at the surface of the puzzle and you're trying to wear them down FAST.Understanding WSF:
The Shapeways "WSF" material is made out of a nylon powder that has been melted together layer-by-layer with a laser. Unfortunately "melted" is not really a binary property of "melted" or "not melted". On the inside of the part the nylon is nearly fully melted an almost completely solid. At the surface though (the boundary between the part and the air) the nylon isn't fully melted or fully bonded to the rest of the nylon part.
This means that the nice "sharp" model you send to Shapeways comes back with a "fuzzy" boundary. There is a certain amount of nylon powder partially stuck to the surface of the part. This cushion of not-fully-sintered powder makes for terrible mechanical interactions. When people talk about "breaking in" a puzzle what they are really doing is wearing down this fuzzy outer shell to the solid part.
The tumbler tears away this outer cushion very quickly. No matter what the temperature and pressure of the tumbling process is, most of this outer, poorly-bonded powder will wear away in the first dozen hours or so. This means that at first there will be a rapid nylon powder buildup in the grit and it can make it look like the grit you start with is wearing the parts down really fast. It is an illusion. If you start with coarse grit it will look like the coarse grit is wearing really fast. The same is true if you start with fine grit.
Wearing down this outer powder does "break in" the puzzle very fast but it does not polish the parts. You need to wear more to achieve a polish.TUMBLING PROCESS:
Everything I've said assumes that you use a vibratory tumbler. Using a vibratory tumbler, you have some control over the temperature of the tumbling process. The higher the amplitude of the vibrations the faster the grit will move and the higher temperature it will be. Also, you can control how fast the grit and pieces move based on the fill level of the tumbler bowl. There is an optimal level that creates a toroidal vortex of grit. Maximizing this vortex maximizes the tumbling temperature.
Ideally you want perfectly even wear. The only way I know how to do that is to tumble with a very low temperature for a very long time. That is, very slow-moving coarse grit. This can take an unreasonable amount of time though. More than a month at least.
A reasonable trade-off is to tumble somewhat quickly with the coarse grit (medium temperature) for a medium amount of time (two weeks or so). Then after you've done most of the wearing down with the coarse grit, you can switch to fine grit for a few days to achieve a bit smoother finish. Then you finish up with the porcelain spheres for a few days.In summary:1)
Start with coarse grit tumbling at a medium speed. Tumble with this as long as possible, checking up on the pieces to make sure they are wearing evenly. This can take 10-30 days depending on the tumbling temperature and the uneven wear on irregular pieces.2)
Once you've worn the pieces down enough or the pieces have worn so unevenly you have to stop, then you can switch to the fine grit. Use the fine grit for only long enough to produce a smooth finish. This can take 1-5 days depending on the irregularity of the pieces and the tumbling temperature.3)
Finish with porcelain spheres to shine up the surface of the parts. You mix the spheres with water. If grit stuck in part's recesses leaks out and gets into the water you must change the water. If you tumble with a combination of porcelain spheres and even a small amount of grit, the pieces will not shine. They will stay matte looking and you will wonder why the porcelain isn't working. This process takes 2-5 days.SUPPLIES:You're going to need a tumbler. Here are mine:
From left to right: 2x Thumler's Tumbler Ultra-Vibe 18 Industrial; 1x Raytech Tumble-Vibe TV-5; 1x Raytech Tumble-Vibe TV-10
The Ultra-Vibe 18 Industrials are far superior to the others. There are quieter, isolate their vibrations better, hold more media, and seem more durable. One of these can hold about 6 Multidodecahedrons at the same time. Note, I'm using the industrial version of this tumbler. As far as I know, the non-industrial version has a weaker motor, less effective vibration dampening springs, and a thinner bowl. I don't know if the non-industrial version of this tumbler works well or not.
I actually own two Raytech TV-10s even though only one is pictured because it now squeaks and doesn't run right. Don't get this tumbler, it isn't worth the money.
The Raytech TV-5 (the small tumbler) is a nice starter tumbler. It doesn't have a very high capacity and it tumbles at a low temperature. Because of the low temperature it gets amazing results but it takes a very long time. One Multidodecahedron completely fills it to capacity.
My tumbling workhorse:
You may have noticed the padding under the tumblers. I'm using 12-inch-square viscoelastic
pads sandwiched between two aluminum plates for the bulk of the vibration isolation. I'm then either using a towel or thick acoustic foam. The big tumblers are too heavy for the thick foam which is why I use the towel. Below that is a rubber mat for grip to prevent the tumblers from "walking" around. Below that is two more thinner, higher density acoustic foam pads. Below that is a high-density foam exercise mat which does a good job of catching grit so it doesn't get in the carpet.
Be aware that if you have two identical tumblers running at the same time and you haven't isolated their vibration you will be very sorry. They will produce constructive and destructive interference (very low frequency acoustic beating) and the rumble will be unbearable. It will sound like an earthquake just getting started or an idling 18-wheeler big-rig. I didn't add all of that padding for fun, that's the minimum amount of padding to reduce the vibration effects to something bearable.Of course, you also need tumbling media:
I use 120/220 for fine grit (this stuff is very fine) and a combination of 46/70 and 60/90 (not shown) for the coarse grit. I'm using silicone carbide. I suspect aluminum oxide would work just as well for the fine grit. Silicone carbide turns the pieces a dark gray color. You will not end up with a white puzzle after tumbling. If you want a white puzzle you will need to experiment with grits other than silicon carbide.You also need porcelain media:
I have both 3mm and 6mm spheres. DO NOT MIX THEM. The 3mm spheres work better but if you have pieces that are able to trap the spheres due to holes they can fit in, use 6mm spheres instead. Trying to get trapped spheres out of pieces sucks. It can take hours of manual labor or even cutting into pieces (yes I've had to cut pieces open to get spheres out).Don't screw around with your health, you need protection:
The tumblers grind grit and nylon down into a mist so fine it's like a smoke. Wear gloves. I've found latex to be more durable than nitrile for handling parts in grit. Wear an organic vapor mask. Use goggles or something that will prevent dust from getting in your eyes. Cover your cloths with a painters coverall. If you don't cover your cloths and skin your cloths will turn your laundry an ugly grey color and you will have to wash them at least twice. The pores on your skin will turn black and you'll have tiny polka-dot skin that takes showering more than once to go away.
I'm serious about this health stuff. I was using a standard "N95 dusk mask" that you see people use for random tasks. It is not adequate. When you separate the pieces from the grit so much fine dust comes up that makes it right through the mask that your spit and mucous will be black. One Saturday I worked on a particularly big and dusty tumble job with a dusk mask and standard eye protection glasses that don't seal. My eyes were red and raw for more than a day and I was still spitting black mucous three days later. My lungs hurt for a week. Get yourself real protection.There are lots of other supplies you need to sort pieces, grit, and spheres:
Top left: a full sink strainer for washing things.
Top right: a mesh to cover your drain. Taking apart a garbage disposal to recover a piece really sucks.
Bottom left: a big pot and a strainer for the pot
Bottom right: tools to dig out stuck grit and wedged porcelain spheres. These are mostly only needed for the porcelain spheres.If you're going to work indoors you need a bag-less vacuum and an air filter:
My air filter is too filthy to even recognize that it's an air filter so I didn't photo it. After every big tumble job I have to take it out and blast it off with an air compressor and change the filter.I've found both an air compressor and a dehydrator handy:
The air compressor really aids in getting wedged porcelain spheres out of parts. The dehydrator is nice for speeding up removal of moisture from grit or drying pieces so that you don't put them into grit while still wet. I also use it after dyeing so accelerate the drying of pieces.You can automate / maximize daily tumble time with a programmable timer:
I now tumble 24-hours-a-day but before I figured out how to isolate vibration, I had to turn of one of my duplicate tumblers at night to prevent the acoustic beating from driving my neighbor crazy. A programmable timer makes sure that you maximize the runtime without any manual work.Part wear analysis:I don't have time to finish this section tonight. I will post all of the pictures and comment on the wear of each piece at a later date.Each of these pieces has been in coarse grit for at least 10 days. A US penny is 19mm in diameter.Dave Pitcher's FTRD corner:TomZ's Helicopter Skewb corner:TomZ's Helicopter Skewb triangle:TomZ's Multidodecahedron Brilic edge:TomZ's Multidodecahedron Megaminx corner:TomZ's Multidodecahedron Brilic edge and Megaminx corner group example:TomZ's Multidodecahedron Pentultimate corner:TomZ's Super-X edge wing:Conclusion:
Tumbling takes a big investment of time and money. It's a pain in the ass to get good results but when you do, they're worth it.