Making Patterns for a Rider Engine

[vtsteam]

Pekka just asked me if I'd explain how I make patterns. I'm making some for a Rider type Stirling engine I'd like to build from iron castings, so I'll try to explain as best I can. This isn't going to be a how-to from the perspective of correct traditional practice, but just the way I do them and think about them.

First step is to draw out what you want as a finished machined product -- at least in its outer form. In this case it's two different cylinders. Both are water jacketed, so there are two round outer rims on each, which a tube jacket will fit to in order to make the water space.

At one end also is a square flange to fit the head of the individual cylinder. There will be ports drilled into those flanges, so they are rather thick. The drawing is really an idealized yet incomplete rendering. There are no fillets, chamfering, shrink allowances, machining allowances, or draft here. It's just something to work with as a start towards what I want to do. I design patterns in my head for simpler stuff, on paper with pencil for more complicated stuff, or as a last resort where I'm having a hard time visualizing probably unwise complications I want to inflict on a casting, in the old free Google SketchUp 7. As I have here.

Note that this is not at all what a pattern will look like, and something this shape would likely be a failure to cast if you tried to. But it is a good picture of the intended shape and dimensions of what I want to end up with.

[vtsteam]

I decided to simplify the pattern so that one pattern can be used to make both cylinders, even though they vary in both height and the location of the water chamber rings. The casting can be shortened, and the rings turned in where needed. The new shape is on the right, compared to the smaller cylinder of the pair.

I have also increased slightly the dimensions all around to allow for shrinkage, and a machining allowance. I made the central bore smaller for the same reasons, but also, my cores (which will be placed in the mold) are already made up at 2" diameter, so I drew that size. Actually the central finished bore will be 2.4 inches in diameter, but I didn't have a pipe "corebox" closer to that dimension. So I will need to bore a fair amount on the finished cylinders.

I have not yet added any fillets/ chamfers, etc yet.

[vtsteam]

I now need to add core prints. These are the extensions needed to make a space in the sand mold for the cores, which will in turn create the hole in the casting which is the bore. The core prints are the darker sections below. These will be part of the pattern.

[vtsteam]

This shape is not an ideal one. Right angles in fairly thick castings have a tendency to create shrink cavities right at the inside corners. It helps to fillet those areas. I will, but I'm not going to draw that in here, because it's a pain in SketchUp, plus I already know that I will add those later after turning the cylinder part of the wood pattern, rather than try to turn in a radius.

Also in general one should determine a parting line and add draft to vertically drawn surfaces, like the flange. But I hope that here I may be able to rap the pattern to give me enough clearance to draw it as a straight section.

Maybe, maybe not. I'll probably decide once I start making the pattern. It's easier for me to get the feel of something like this when I'm holding it in my hand, rather than trying to draw it.

[vtsteam]

I forgot to mention that the sharp corners on the squared portion could also benefit from being slightly radiused. Sharp corers are harder to mold, requiring more care in drawing the pattern out, though doable if absolutely necessary.

In this case again, I'd do that on the pattern rather than draw it out. I know what I need to do. Well that wouldn't work if you were sending your pattern out to be made, and cast from -- playing strictly the engineer.

The next decision is whether to make this a split pattern or just a single pattern. Or a half pattern.  If a half pattern it will likely be mounted on a match plate. If a split pattern, each half could be mounted on one side of a match plate, and then two castings could be molded at one time (that only works on a symmetrical shape like this one). Or a split pattern can be doweled together with a slip fit to be molded as a loose pattern.

These considerations also may relate to how you're going to form the pattern. If you have a wood lathe, or wood turning capability, you will likely make either a full loose pattern, or split it. But if you are going to form the shape without a lathe, you may decide to go with a half pattern on a match plate, as easier and quicker to make.

For me the best option would be a split pattern on a match plate -- I could mold both cylinders at once. But since I'm "rusty" at iron molding after several years, and with a new furnace and unfamiliar pattern, it would seem wiser to start with trying to pour just one casting. Pouring and gating for two is less certain of a good result, and also why waste metal for two if you need to adjust something?

If I did do a lathe turned pattern, I'd probably do it as a split pattern, which would be molded loose. At least for the present. Later, if desired it could be mounted to a match plate

Now, while I do have a wood lathe, it would have to be used outdoors, and this winter makes that uncertain. I could use my metal lathe as a wood lathe, but that makes a real mess in my very small metal working space, and wood sawdust and fibers get everywhere on the lathe.

The last option is to make a half pattern, and that I could do indoors fairly easily. It's harder to shape and sand than making a whole or split pattern on a lathe, but it's just a little more time and work.

Hmm.... which to do?

[vtsteam]

Hunting around for some wood of the right size, I found a 4"x4" x 6' of basswood I had cut many years ago from a downed bigleaf linden. I cut off a chunk and roughed it to size on the bandsaw, and then split it. I also could have used two two by's of pine or other softwood without splitting them. But I had the basswood.

[vtsteam]

I had marked the end circles with a compass, and decided to plane the blanks down to about 1/4" from the lines. That will cut down on the mess when I turn them on a lathe. Or I could keep on going with planes and finally sandpaper to bring it to shape.

I drilled and countersunk for long screws to fasten the two halves together.

[vtsteam]

I brought the 20 year old Gingery lathe out of mothballs and set it up in the spare room upstairs. I do have a spur center, but the Gingery spindle wasn't bored for a morse taper to accept it (might be a new project). Anyway, rigged the pattern blank up in the 3 jaw, and made a sawdust mess. But fun, on a cold winter's day to be actually turning something!

I turned the main part of the cylinder as close as I could to the three jaw, then I turned down one end to the core print diameter (2"), reversed the pattern in the chuck and then turned down the other end. I used old fashioned divider type calipers to do all the dimensioning and checking. Worked great. 

[vtsteam]

Here's the finished turning:

[PekkaNF]

I was about to ask about radiuses and drafts...but then I realized that his is a core of a cylinder bore ..... and ends will abut to sand...

Good writeup!

[Country Bubba]

Instead of making a new spindle for the Gingery, borrow a trick from the metal guys and make a spur center that mounts in the chuck.
Quick easy way to switch back and forth.

Country

[vtsteam]

Pekka, yup. But remember, this pattern is also not finished yet, this is just the turned part. For most patterns, fillets and rounding can be added later where needed. I'm not a wood turner....most patterns I make don't even get turned on a lathe.

Country Bubby, yup, many ways to skin a cat. But then I'd have an even longer overhang from the headstock bearing, which on a light lathe like the Gingery is a minus.

I don't need to make another spindle, I just need to bore my present spindle of choice to fit a Morse taper, which would allow not only mounting the spur center I have, but also any other Morse taper center, collet, chuck, or spindle as well.

I'll just bore that out the same way I did the MT on the tailstock. 

[vtsteam]

Next step is to cut out the 2"x 2" x 3/4" square flange piece. To mark it out, I am not going to draw a square of those dimensions, with a circle in the middle for a hole. Why? That hole needs to fit the turned cylinder pattern,and the square needs to be split, just like the cylinder. Otherwise the whole can't come apart, and it would have to be moulded as a single pattern and coped down.

If I drew out my 2" x 2 piece, and cut it out, and then made the hole, I would then have to split the flange, and it would no longer fit the pattern. You always lose something in a saw cut. If you hollowed out the hole larger (always much harder to do than cut it the right size to start with), the outer 2"x 2" dimensions would also be short on one side.

So what I'm going to do is just make a hole first, just slightly larger than finish size. Then cut the pieces out oversize. Split them. And then put them tight together and draw out the actual 2"x"2 finished size for cutting. That way the pieces will all fit together and yet the outer dimensions will be correct.

Old fashioned outside calipers are VERY handy for checking both the pattern and the diameter of the circle drawn by a pencil compass. Notice I've also kept the circle more than 2" away from an edge -- even though it was tempting to use the planed edge as one of the sides.

[vtsteam]

I cut the hole out with my electric hand jigsaw. Unfortunately the curve was just a hair too tight for the blade width I had on hand so the cut was a little uneven. No problem....because I have an oversize piece in it's outer dimensions, I can simply fit the hole better with some sandpaper using the cylinder pattern itself as a sanding block. And, If need be, after sanding both flange pieces to shape, I can trim their split edges further to meet again.

I can do this because I didn't try to trim the outside of the flanges to shape yet. You could make a general principle of this and say, always fit your most difficult shapes in working with wood, and then trim the easier stuff afterwards. It's a mistake to try to draw everything out first, and then cut to that. Or not a mistake necessarily if you're a faultless mechanic. But if you're a regular person, you'll go through a lot more wood and frustration trying to cut to a layout, than laying out in relation to an already completed difficult cut.

If worst comes to worst you can also build areas back up with cardboard or even Bondo (auto body filler). In fact I use both quite a lot when I need a small amount more thickness in some pattern than standard lumber sizes, and also for fillets, etc.

ps. if you have a mill and a boring head that would be a great way to do something like this flange. A hand jigsaw is what I have available, so not ideal but still workable....

[vtsteam]

Traditionally the best tool for doing inside curves would be to use a very sharp in-cannel gouge to pare away material. I would say that's probably the best way, too. Unfortunately I just have one out-cannel gouge in this size range. I may convert that one some day, as it's not very useful as-is.

And here, to depart from tradition, I have used dots of hot glue as welding tacks to position the two flange pieces on the cylinder pattern, and then Bondo-ed across the joint face, as well as some grain tear out at a wood flaw in the core print. This will all be sanded back.

I like using hot glue for tacks, as it makes positioning easy while it's still hot and somewhat flexible, yet grabs fast enough to position things without clamps, just holding the pieces in your hand. If you don't quite get it right, just pick off the glue tack and reposition.

I've finally also drawn the cutting lines for the actual flange outline.

[Sea.dog]

Incannel and outcannel, now there are two new terms. I have one of each but never knew that they had a specific name, other than gouge, that is 

[vtsteam]

Sea.Dog, seems like 95% of gouges you find are out-cannel. I'm definitely going to convert mine some time. Probably need hardening and tempering again. I think with some creative blacksmithing one could also convert a flat chisel to an in-cannel gouge. They seem to command a premium on Ebay.

Back to the pattern....here I'm trimming the flange down:

[vtsteam]

And then I added a Bondo fillet on the cylinder side of the pattern. I did have to remove one of the screws. They all get removed and plugged when the pattern is split.

[vtsteam]

Well, speaking of splitting, why not now?

Here they are, the two half patterns....

[vtsteam]

I cut off two short lengths of 1/8" brazing rod, rounded one end of each and epoxied them into two of the holes left by the screws. They were a tight slip fit.

I "reamed" the holes opposite by wiggling in them another bit of brazing rod, just to make sure there was an easy slip fit for the pegs. You want the halves to separate naturally when you lift the cope, so one half stays in each part of the flask. If the cope pattern falls out after the lift, it can damage the sand cavity, and then you have to re-make it.

For the same reason, I don't make the protruding pegs very long, The pegs only have to locate the pattern when placed together lying flat. With the pattern lying on a table, you should be able to lift the top half off without the bottom half moving at all.

[vtsteam]

Time for sanding sealer. I make my own. Lacquer, talcum powder, thinned with a small amount of lacquer thinner.

I need the real talcum powder with talc in it. Not cornstarch. Important to check the label on that. For amounts, I just mix it in a small jar -- I put maybe a heaping tablespoon in about 8 oz of lacquer -- not critical, because the talcum powder will settle out so you have to shake it once in awhile to mix it in again. Whatever amount it wants to suspend is fine.

I add a small amount of thinner sometimes so it brushes easily, not too thick or drippy, and dries quickly. We're talking like 15 minutes. BTW I do this outdoors in summer or in an unoccupied shop or garage with plenty of proper ventilation. and no open flames or hot elements. The vapors are both harmful and very flammable.

[vtsteam]

I've brushed two coats of sanding sealer on, and sanded lightly with 320 grit paper, then sprayed on a light coat of Krylon (tm) yellow over. This gives me a good chance to look over the pattern for flaws and make any modifications before final painting.

After I brush on each sanding sealer coat, I wrap my cheap disposable bristle brush in aluminum foil -- that way it can be used again without having to clean it out in lacquer thinner -- avoiding both fumes and expense. Generally you can paint, then sand, and then re-coat all within an hour with lacquer.

I use Krylon for my color coat (not strictly necessary) because the standard Krylon is lacquer based and compatible with the lacquer sanding sealer. It also dries in about 15 minutes. If I tried to paint with an oil based enamel like Rustoleum (tm) I'd be waiting a long time, and may have compatibility problems if the sealer is not absolutely hard and dry. Nothing like ruining a finish with orange peel after you've worked hard to make the pattern and undercoat as smooth as you can.

Now, looking at the pattern halves in real life I'm feeling that I should add some draft to the flange edges. There's plenty of extra material for a machining allowance there. I could also taper the flange thickness, but not sure I want t do that because then it makes it difficult to mount on a faceplate. It would have to be faced first with a 3 jaw on the cylinder and a steady rest, etc.

I dunno, we'll see. Anyway, tapering the edges for now.

[vtsteam]

Okay, 1/32" of draft taper (1 degree) was added to the edges of the flange. This was done easily by setting the table of my little disk sander to that angle and just running the edges of the flange against it until it started to take paint off the bottom of the bevel.

A belt sander (linisher) or disk sander is indispensable for adding draft to patterns, and for forming shapes like disks after rough cutting them with a saw. You just mark out the shape on your rough cut blank and sand to your outline. If you have set the sanding table to the draft angle, then draft is added automatically as you shape the pattern, even on curves.

[vtsteam]

The finished pattern. I had a little trouble with the yellow Krylon I'd sprayed it with before. It wouldn't dry fully over the sanding sealer. Not sure why, it was an old can and I dunno, they seem to keep changing their formulations. That one was labeled "Indoor/Outdoor".

I had a can of Krylon "Fusion" in Navy blue, so I tried that and it dried well, so that's the color we've ended up with. In my earlier casting days, I sometimes would paint patterns for aluminum castings yellow, and those for iron black. But I never was very consistent about it. I had lots of John Deere green patterns, too. Besides simple patterns get re--used, often with a variety of metals so coloring for type doesn't make sense.

I did once look up what "real" foundries used, and interestingly, though there was a US joint standard published in 1932 where all unfinished surfaces were to be painted black, machined surfaces red, core prints and seats for loose core prints yellow. Loose pieces and their seats were to be painted yellow with red stripes. And stop-offs were yellow with black stripes.

On the other hand I've seen commercial patterns using those same colors, but switched around, so that unfinished surfaces were red, core prints black, etc.

Blue apparently is way out in left field, but there you are.......

Now, is this a good pattern?
Well, in order not to mislead, um, not totally....: it doesn't have draft on all vertical surfaces, so I'll be depending on it separating properly and then on my rapping and flask lifting skills (not unquestionable!) Also being of a light hardwood, it's kind of heavy compared to a pine pattern, so the cope half of the pattern might drop out.

Finally, the design, with a very heavy flange section at right angles to the main cylinder looks like it's asking for trouble with a shrink cavity, or tear on the inside corner. In aluminum, it would be pretty likely unless you find some way to feed it as it cools. I'm not sure about iron, since it shrinks about half as much as aluminum. It might be okay. Zinc shrinks even more than Al. This one is intended for iron, of course. Anyway we'll find out.

A smarter or at least easier design would just be to cast a plain cylinder and machine a fitted flange after. But there's always the temptation in casting to include shaped features and reduce machining time and numbers of parts in an assembly. It will be interesting to see how this casts in iron.

[awemawson]

Looking smart in blue Steve

I'd be very tempted to feed that flange, once you deduct the core volume it's a significant proportion of the whole casting.