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.
I'm by no means an expert on these things but I have done a fair bit of Aluminium casting in the past so here are my thoughts, right or wrong (more experienced members should feel free to disagree it won't cause offence).

The "micropores" in the picture do look like gas porosity and whilst in many cases these are not a problem, in this case, as we are talking about engine parts they could cause difficulty some degassing would improve matters.

As would an "active flux" which not only makes drossing off much easier, but actually refines your melt by actively drawing impurities out. The products I used were "Degasser 190 and "Coveral 11" respectively both from Foseco limited in the U.K. You may not be able to get them in the States but I am sure there are suitable alternatives

The surface "craters" are a little more difficult to diagnose, As you say they could be residual products from the foam, but could equally be due to some kind of sand penetration from the mould, they could even be caused by steam if the moulds were at all damp,

I think the pouring temperature might be a little low as shown by the fact that some of the gates have not filled completely, and ideally you should provide a "riser" opposite the pouring sprue this keeps a "head" of metal to allow the casting to "draw" from and also provides extra exit for the gasses.

That all said these are good castings for a first attempt at this type of casting. I wish some of my first ones looked that good.

Best Regards Mark
I agree with what has been shared so far, entrained gases are normally pinholes under magnification through out the casting
,normally remelting ingots releases the entrained gases. My processing furnace is direct fired with the scrap in the flame, this does entrain gases, the degassing recipe below does solve the problem .

I provide quality aluminum ingots from automotive engines, mostly to Hobby casters and a few small foundries and Galleries. I've processed over 14 tons in the last 5 years,

That said I have had a number of customers ask casting questions, so I started providing the following information in each box sold.

I'm based in the US, so some of these ingredients maybe known by different names in other areas. I also have a process for heat treating with a ceramic Kiln if any one would like that.

Art B

Flux recipes

Drossing Flux (used in our processing)

50% sodium chloride “normal salt, preferably without Iodine)

50% calcium chloride “NoSalt brand”

One teaspoon per 10 -12 lbs.

Degassing and Drossing Flux Caution releases chlorine gas, you must have adequate ventilation

50% Drossing Flux

50% calcium hypochlorite “Pool Shock”

One tablespoon or three teaspoon per 12 lbs.
The third attempt on the oil pump casting produced better results as far as the surface finish was concerned and I think this technique will be good enough. The patina is now essentially the finish on the foam pattern.
So this time round I coated the pattern with plaster. Not plaster of Paris but a metal casting version of it called HydroPerm. The main difference is there is more talc in it plus some Portland cement, not sure what else.
My theory from the previous attempt was that the joint compound had simply not been strong enough to stay together in the presence of all the violent off gassing of the foam etc. I think joint compound could still work but it needs to be applied very thick.
The problem with plaster is that the foam gas does not get out easily even using the modified stuff.
Very impressive, well done !

With changes in section and some very awkward angles these are not easy castings to do but you seem to have the technique well worked out now.

I doubt a commercial foundry with extensive (and expensive) resources could do any better, If you could find one willing to take it on at a reasonable price.

This new capability brings your Gypsy project one step closer and I shall be following this thread with much interest.

Best Regards Mark
Thanks. I think the coating was definitely the answer to the surface finish issues of the previous attempts. To that I’ll add some rambling thoughts and learnings.
There are commercially available coatings specifically for this but they are only available in large quantity and are very expensive. I think these coatings (and joint compound) must micro fracture enough to allow the pressurized gas to escape into the sand but retain enough structural integrity to keep the surface finish.
The plaster obviously sets hard and does not let the gas escape, so if the casting is to be machined, drilled, tapped etc. then the potential for problems with gas voids is much greater.
Just my two cents but I think limiting the plaster use to only visible surfaces that are not machined and using JC ( or possibly nothing) everywhere else might be a good compromise. That said I am experimenting with a thicker more even JC coating as I just feel it’s the best solution for foam if it can be applied correctly.
Any trapped gas bubbles are likely to occur near the top of the casting so orienting the pattern in the sand with the ‘void’ sensitive areas low down in the flask seems advisable (not what I did ). In this case the oil pump gallery. Also avoid large flat horizontal surfaces by tilting the pattern in the sand.
Vents certainly clear a lot of the initial gas envelope out of the pattern if it can’t get through to the sand but don’t guarantee that all the later bubbles will get out as the metal cools, particularly as this is occurring at the fast cooling surface of the casting. My worry about vents, other than this, is that it may be that the initial expansion of gas that fractures the coating and allows the gas to permeate the sand. So dissipating it through vents may not be what is wanted. Though clearly if the pattern is encased in plaster then this is the only option.
Anyway this Is an art and my hat is off and head bowed low to those that have mastered it, I’ll continue along the curve hopefully in a generally upward direction.

Rgds, RS.
Whilst most of my casting was greensand I have done a little investment casting and as I understand it the coatings used are designed to be permeable enough to allow the escape of gasses but restrict metal penetration of the mould much in the same way as greensand does.

Just as with greensand, venting, especially at the uppermost points of the mould or where gas pockets might form is helpful and as you point out orientation can be important.

I would be interested in some more details about this HydroPerm casting plaster it certainly seems to be doing the job well, If it is the same sort of stuff I think it is then your micro-fractures idea is fairly close, It partially decomposes at casting temperature and forms a sort of matrix which lets the gasses escape.

Here is an old idea from greensand practice which may or may not work but might be worth a try. What about powdered graphite to improve surface finish, traditionally used for Cast Iron but I have used it with Aluminium with good results. It is traditionally mixed with Alcohol or a suitable solvent, sprayed on the mould cavity then either ignited or simply allowed to evaporate. If you can find a suitable solvent that doesn't melt your casting foam a quick spray on your pattern and allow it to evaporate before coating could be beneficial.

In any case I don't think you have a problem with surface finish anymore these castings look fine, the acid test will be machining them of course but you are definitely going in an upward direction so keep at it.

The cylinder heads look pretty tricky are you going to cast them or machine them from solid

Best Regards Mark
That is an intersting approach looks like it turned out ok. Attached picture is my attempt made from the solid on mill. I have the original casting but I figure getting the distributor drive gears to mesh is going to be trickey better to have a spare. I could never figure how to drill through that elbow for the in and out on the pump. I made the replacement without the elbows figuring to make them seperately.


  • chenery.jpg
    146 KB · Views: 250
That looks great, there‘s a good bit of machining and fitting on that part so doing a trial run is definitely the way to go. I‘m planning on making two casts just in case.
I’ve had a bench mill on order from Taiwan for four months so I’m using the time to work on these castings.
Since the last attempt shown above I’ve done one other casting. It turned out that the longer of the magneto supports shrank by 50 thou and the shorter about 15 thou after the metal cooled so I
adjusted for that in the pattern.
Also on this attempt I went back to drywall compound as the refractory investment and the surface finish was a big improvement on the earlier attempt. The two big mistakes I think I made previously were to overheat the metal and I did not get enough of the residual moisture out of the cast plaster pouring cup.
The Hydroperm plaster coated version looked good because it did not disintegrate with the excess moisture and heat but I don’t think it is as permeable as the joint compound. So in my case the Hydroperm seemed to be the answer but it was basically because I was doing other things wrong.
After machining and polishing the latest casting there is evidence of porosity in the metal still. I think you can see it in the photo. I solicited some suggestions from experienced lost foam casting folks on another forum and I think I need to reduce the metal temp some more (to 1400-1450 F) and also change the pouring arrangement to include a fill basin and weir into the sprue to eliminate drawing air into the metal during the pour.
Just waiting on a new thermocouple temp sensor and I will try these things and report back.

Regards, RS.
First go at casting the crankshaft bearing end caps.
Lost foam cast also. The center cap has been experimentally milled fairly close to drawing dimensions but still a little oversized, the others are as they came out of the sand.
I think the porosity is a bit more under control now.
These end caps were poured at 1450 F using a pouring cup of Kelly Coffield’s design. Made from moldable ceramic fiber over a foam pattern, I would highly recommend using something like this for lost foam as it greatly reduces air entrapment.
It can be used many times as molten aluminum does not wet it. It’s also a high thermal insulator.
The latest installment (#5) in the long saga of the oil pump housing casting shown below.
I realized as I assembled this foam pattern that I had not allowed clearance for the numerous studs and nuts around the periphery of the cam gear housing to the ‘magneto’ housings so I added 0.2 inch to the width. It’s in the drawings, I just missed seeing it before.
After filing off the gates and cleaning up some areas with a file I have not encountered any of the dreaded gas bubbles of previous attempts but I’ll see how it goes.
It’s tempting when seeing progress to keep trying to make it better with ‘one more go’ but I want to move on to the carburetor.
Chenerys original castings never left much room for error. I screwed up boring the main bearings I will show my "fix" if any interest. I remember asking Les how he did it he said no problem just use a boring bar. I tried but had no luck. Also boring for the cam shaft is very tricky.
Oh, yes please, I was wondering about that. I had seen a boring head being used for a 1/4 scale Cirrus build but it’s quite a different arrangement with only one centre bearing so plenty of room to get a boring head in there.
When I first tried holding it vertical and drilling just did not work well. Thought about mounting on the lathe carriage with boring bar between centers but dismissed that. Anyway with the screwing around the bores were not straight etc. Thought about buying a new casting but found Hemingway too difficult to deal with. So to salvage the block I made a jig plate to locate bore centers and cut away the damaged stuff. Then I made new inserts, bored them undersize split and and secured in the block. Then I reamed them with a long reamer. Made mains again a little undersize mounted and reamed to size. Looks good but have not tried a trial fit with crank.
The carb was the last item on the casting list and proved to be more of a challenge than anticipated (of course), mostly due to how many attempts at lost foam casting I stubbornly attempted.
This was probably the closest I got with foam, and possibly a light sandblast might have made it OK but I was struggling with porosity issues despite controlling the temperature and being careful with the pour.
In the end I drew the part up in fusion360, printed a resin pattern, pulled out the bucket of green sand and got to work. Just needed to remember to taper the fuel inlet pipe under the bowl so it lifts out of the sand cleanly.
I added a tang for holding the part in the chuck to facilitate machining of the conical intake mouth and drill the air passage.
This was the third try. It’s not perfect but it’s good enough for me and I can move on to other parts I think.
I've been waiting for a mill to arrive which it has just done so. The plan is to start roughing the crankshaft from a piece of 1040 flat bar.

Getting the mill set up took forever but the DRO is on. Had to make custom brackets for the read heads as none of the supplied brackets worked out very well. Its finally all trammed in.

The crankshaft starting material was flat bar 9x2x3/4. It was fly cut to thickness and end milled to width. Then put in the four jaw and the ends faced and center drilled on the lathe. The block was then back in the mill to start the roughing out.

Ending up with all the cutouts.
Next job was to fabricate a lathe dog for turning between centers. I made this from some 1/4 in ms plate and scraps I found around the shop.
To minimize distortion during turning I plan to make some steel spacers out of 3/4 in bar stock to fit between the webs.

About now someone shouts “Did you even look at the drawing before you started cutting metal !”
So the dimension from the tip of the front of the crank to the first crank web is shown as 1.987“ in the rough machining drawing. In the detailed finished drawing the same dimension totals as 2.05”. Ruh row. It’s not clear where the additional 0.063“ came from.
The finished drawing specifies the tapered portion of the shaft as 0.625” plus a threaded portion of 0.25”. (The actual length of the threaded portion on the drawing however is less than 0.2”). My inclination is to shorten the tapered portion to 0.562”. Maybe the numbers got swapped around. Then it all adds up.
Possibly someone else has seen this and knows the answer or did something differently.


When I made the crank for my Gypsy, I also noticed some issues with the crank drawing. The attached sketch shows my own interpretation of the dimensions. Here, the original dims have been converted directly to metric. In order to get a 3 degree taper, the length of the taper needs to be adjusted (15.17mm or 0.597ins). The dimension from the front of the crank to the first crank web is (here anyway) 51.37mm (2.022 ins) and there is a 40.46mm (1.593ins) dimension between the taper and the centre of the first crankpin.

In the end I actually decided to use a different angle for the taper and a longer screwed section. The crank turned out fine and fits in the crankcase bearings perfectly so I must have done something correct. I'm still someway from completing the whole project though!

Hope this helps.



  • crankshaft orig 3 deg taper jpg.jpg
    crankshaft orig 3 deg taper jpg.jpg
    95 KB · Views: 94

Latest posts