Gypsy Engine

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Thanks, yes, I see you checked the dimensions on the main drawing and accounted for it. I did notice that the taper angle was a bit shy of 6 deg with a taper length of 0.625”. Just to add to the fun the prop boss that sits on the taper is drawn as 0.538”. If you add the 0.063” discrepancy to that you get 0.601” which is magically the exact taper length to get 6 deg. Does make you wonder..
Anyway, I will make the taper length 0.562” and use some other angle (like 6.4 deg) to make it work.

A little further down the path with the crank shaft after a fair amount of time at the lathe knocking the corners off. There are a lot of oil ways to drill yet.
Before completing the crank I want to make a start on the top case. Skim the flange flat and mill the bearing supports to size.
Something I noticed on the casting is there appears to be some shrinkage in the casting front bearing support. On one side it looks to have dropped about 1/16th” below the flange.
I’ll see how it goes, maybe it’s Ok.
If the casting is a problem, i.e. it doesn't clean up, you could build it up with Technoweld, Lumiweld or similar antimony based aluminium 'brazing' rods.
I've used these rods over several years, mostly on motorcycle parts, with great success.
The antimony based rods aren't really welding, more like brazing.
I use a propane torch for heating.
The only problem in using those type of rods is that it is almost impossible to do any future TIG welding without removing all traces of it in close proximity.
OK, I skimmed the flange of the upper crankcase and as expected there is an area of the front main bearing support that remains below the level of the flange.
There is still maybe 0.02“ that could be removed but the depression is a good bit deeper than that. As kindly mentioned above this could be filled in relatively easily.
Having stared at this casting for a while and having looked at pictures of the original casting, as well as the drawings it became apparent that there are some significant differences which I think are worth noting.
The big difference is in the bearing supports.
The three central bearing supports in the new casting consist of semi circular sections supported by webs across the case. The front bearing support is a semi circular section that cantilevers out from the front of the casting. There is no material or webbing underneath it for either support or to accommodate the cap screws.
For comparison here is a photo of the ‘original‘ Chenery casting at a similar stage of the build.
Notably, the bearing supports extend right down to the ‘top’ of the casting and the webs are a good bit wider. The front bearing is basically solid. There is plenty of material and depth for the bearing cap screw holes to be drilled into. I’m not sure the same can said for the new casting.
There are some differences on the rear of the casting. Here is the new casting and the original below.

There are two ‘buttresses’ on the right side of the original casting below the flange to hold two studs. The studs are not on the casting but you can see the tapped holes for them. These buttresses are missing from the new casting.
Also the flange on the left side is thicker and joins to the cam gear housing to provide material for the two studs on that side.
In contrast the flange on the new casting is oddly much thinner in this area and does not meet up with the gear housing until much further towards the center, with very little material for robustly securing the studs.
I’d be very interested in hearing from anyone who has built an engine with this casting to hear how they worked around these differences.

Note: The ‘original’ casting photos are borrowed from a really excellent thread on the Model Engineer forum by David K.

Sorry to read of your casting woes.
I have used the aforementioned aluminium 'brazing' rods to add a lug to a casting (motorcycle gearbox) which was then drilled and tapped and has supported the clutch cable loads for over 10 years use.
Thanks, yes, I’ve used those rods once before to fill and re drill a misaligned threaded hole on a previous engine. The brazed metal seemed much harder than the cast aluminum and took the thread far better.
I’d definitely consider that to possibly fix the stud issues at the rear.

My chief concern are the crank cap screws on the front bearing. Based on the drawing dimensions and some measurements they would be breaking out of the underside of the casting after a depth of 0.275”. They need at minimum a depth of 0.438” so they would be fully threaded for only about 60% of their length.
I wondered about over sizing the holes for these screws then try brazing some plugs in long enough to take the screws. What could possibly go wrong !

The webs would help on the other bearings, they are 0.17“ wide and the screws threads are 4 BA so drilled for 0.118”.
The problem here is that there are supposed to be oil ways passing to the bearing shells from one side of the flange through the web. I’d guess that’s why Les made the webs as wide as he did on his own casting to allow for room to get a drill past the cap screws.
In fact due to the novel front bearing support arrangement it’s not going to be easy to get an oil way to the front bearing shell. Maybe a tube to cross the gap….

This casting has the feel of a prototype that hasn’t had the bugs taken out yet. I’ll probably find a shaft passing through the spoke of gear wheel next.

By way of an update on the Gypsy castings.
I contacted the originating foundry to ask if it was possible to modify the patterns and try another run. It’s been about 10 years since the original casting so I was not overly hopeful but they did reply and asked for details on the modifications, which I have passed along to them.
I had hoped that only the upper crankcase casting needed rework but there are issues with the sump as well.
Anyway, I await their response and quote for the work.

Your casting looks similar to the Hemingway casting that I have (pictured). There is insufficient meat on the casting to drill the oil holes for the two end bearings. However, I'll only ever run the engine for short periods (if I ever get it to run) and I'll rely on enough splash cooling for those two bearings.



  • Hemingway Gypsy casting.jpg
    Hemingway Gypsy casting.jpg
    186.2 KB · Views: 83
Thats interesting, I didn’t know Hemingway ever sold the Gypsy castings. I built the Aero V twin from them.
I wonder if they also tried to recreate them from the drawings and got similar results.
Hemingways don't list the Gypsy on their website but supplied a set of three castings to me as a special order about 18 months ago. The castings are fairly good in themselves but there are some simplifications/omissions in the patterns - no showstoppers though. If anyone's interested I'll post a summary.

I really liked your Aero V twin - must finish mine someday.

Spent some time recently preparing an attempt at a home casting of the upper gypsy crankcase. The first step was to modify the original casting to avoid the external feature issues.
I filled in some areas at the rear of the casting and blocked up the holes for the cylinders, leaving them slighty raised so that their positions could be located in the sand impression. I will use these to help locate the cores.
After a trial packing of sand and a few modifications I could lift it out of the packed sand without trouble.

I have also milled the sides of the lower crankcase casting. I was not hopeful that the flanges for bolting through to the upper case would be wide enough to allow clearance for the bolt heads, but it turned out it was OK They obviously built in a generous cross sectional taper to allow for the removal of the mold.
I’m glad because it’s a really nice quality casting.
The top case requires four cores.
The middle two are identical but there are some differences in the two end ‘bays‘. Notably there is an undercut to the front crank bearing housing on the right that the oil fill passage connects to from above.
The pattern shown above is based on the original casting with a lot modifications involving jb weld epoxy and pieces of wood. So this represents what I’m aiming for with the interior of the new casting.

I built two core boxes. One is for the rear core only, the second, shown below, can be used with alternate parts to make either the two center cores or the front core.

As shown the scheme to produce the cores involves a number of fitted pieces of wood that can be assembled using screws, then packed with sodium silicate impregnated sand, and then disassembled without putting stress on the cores. The core can easily be broken so the core molds were designed so that each section contacted a minimal number of sides.9C6E63AF-7464-4A13-BB14-C6BFE40DEA7E.jpeg
Filled with sand.

Some additional notes on the core making….
Even with these core boxes there was still problems with the sand sticking in the molds and breaking. Sodium silicate sticks to pretty much everything ! Success was eventually achieved by coating the interior surfaces of the box sections with a low viscosity epoxy resin. Epoxy is possibly the only thing the sodium silicate will not attack. I also rub carnauba wax onto the mold surfaces just for good measure.
I found that the sand needed to be thoroughly wetted through with the sodium silicate or the cores ended up too delicate to use, but maybe my sand was just coarse. Sieved play sand from HD.
After packing the molds with SS sand I placed them in a ziplock bag which was then filled with CO2.
Initially the CO2 was produced using vinegar and baking soda. I would place a 1/2 filled cup of baking soda in the bag and add vinegar through a tube, just enough at a time to get a reaction that came up to the top of the cup. This technique worked well and is very cheap. If time isn’t an issue you can just leave the cores in the mold and they will harden up just from the regular atmosphere after a few days.
I ended up making quite a lot of cores at the time as I debugged my core making process and eventually I bought a 0.6 L cylinder of co2 plus a valve. These cylinders are meant for carbonated drink makers. You can get a filled cylinder for $15 plus shipping, it’s ten times the expense of vinegar. You can fill a large zip lock bag about 50 times I found.

So with sand cores to hand, here is the sand casting plan of attack!

Step 1. Press a plywood template of the crankcase outline into the sand. The underside of the template includes indents to position the four cores.


  • DD03C52B-7429-443B-BEFD-99290950719F.jpeg
    323.4 KB · Views: 28
Step 2. Place the crankcase pattern into the impression in the sand.D236D747-63E9-4992-8217-058EE64BAD96.jpeg

Step 3. Pack the sand around the pattern with sprue and vent. Separate the two halves of the mold , pull out the crankcase pattern and place the cores in position using the indents from the step 1 template.

Check the core positioning using another template for the cylinder positions.

Step 5. Figure out how much aluminum this is going to take and realize you need a larger crucible.
Step 6. Realize the larger crucible won’t fit in your furnace.
Step 7. Build a larger furnace…….. sigh.
This is one of the more confusing threads I have ever read, but none the less very interesting.
Someone named "floppyearedmule" (LOL) started this thread, and then 8 years later ringsnapper picked it up again, with all sorts of other folks apparently doing parallel builds, popping in randomly ?

After a quick read, I have a few comments, as follows:

1. I have followed numerous backyard "lost foam" casting projects, and so I researched lost foam castings (but have not done them myself).
The commercial foundries use polystyrene beads that are expanded into a metal mold via heating/steam.
The coating on the foam pattern has to be permeable enough to pass both liquids from the melting foam, and gas produced by the vaporizing foam.
The defects I have seen in backyard lost foam castings seem to be mainly due to using foam that is not polystyrene (rigid foam so that it can be machined), and perhaps from using a coating that is not appropriately permeable.
If the wrong foam is used, it may produce so much liquid and gas that it won't work regardless of the coating.

2. One guy showed me some lost foam castings, and he did not use any coating on the foam at all, and the castings turned out quite decent.
He just packed the foam in sand.

3. I experimented with sodium silicate cores, and learned what not to do.
I found that 3-4% of sodium silicate is all you need, and any more than that makes the core very difficult to break out after casting.
And putting a sodium silicate core in a bag with CO2 destroys it, which is why many people jack up the percentage of sodium silicate very high.
Sodium silicate sand should be gassed for 5 seconds only, and then no more CO2.
The SS cores I made and put into ziplock bags lasted less than a day on the shelf, regardless of SS percentage.
The SS cores I made at 4% with 5 seconds of CO2 have sat on the shelf and been perfectly usable after a year, and break out easily after casting.

4. Overheating aluminum causes all sorts of defects, such as porosity, and bad surface finish.
I heat my aluminum (356) as fast as possible to pour temperature (which is 1350 F), and then pour immediately.
Never stir aluminum; just skim and pour immediately. I don't use a degasser, and don't need one the way I do it.
1350 F is gererally hot enough to fill most mold cavities, and cool enough to give a smooth surface finish, especially in Petrobond.

5. I would probably bake those cores to be sure to drive off all moisture; perhaps at 250 F?
I generally lightly flame my cores with a gentle propane flame, to do the same.

6. I generally use a runner down the side of thin parts like that, with several gates.
Thin parts seem to fill better with knife gates, which are wider gates with fill the mold faster.

7. I have had molds like that trap air, and ruin the casting with a large air bubble in the casting.
I have since started venting the high points of the mold with small holes poked with a wire.

8. And the best way I have found to stop ss sand from sticking to a corebox (ss sand is very bad about sticking) is to apply blue painter's tape to the interior surface, and then apply a little Johnson's paste wax to the tape every time before you make a core.
There is not much that will stick to 3M blue painter's tape.

These are a few things I have discovered over the years.

Great thread whoever is posting on it !!!
I must say that floppyearedmule wins the prize for best screen name though.
And one last trick, if you need a larger crucible, and it won't fit into your furnace, stack hard fire bricks around the top, and temporarily extend the height of the furnace (assuming the diameter is ok).

You can use a pottery shelf as a temporary lid.

This is just hard fire bricks banded together.
You can make a new furnace in about 30 minutes, for aluminum use.
Not pretty, but functional.


Latest posts