Anzani Engine Project

[David Paule]

This is about a six cylinder Anzani engine from around 1919 or so. This round engine is a twin-row-three, in case you're wondering how an even number of cylinders can work in a radial engine.

It arrived here assembled but heavily preserved. The cylinders and accessories are now off, and the case has been split. An engine test stand is built and waiting, and Chad Wille built us a prop.

It has a dead-loss oil system and one single magneto. There are some interesting quirks, like the facts that there are only two bolts to attach each cylinder, or that it uses slippers at the bottom end of the connecting rods rather than a master rod, or that nearly every single part is numbered and assembled to fit that one location.

Craig and Roger are the engine experts here. I expect that they'll comment here too. I'm more of an airframe guy. All of us have a mixed background of hand's-on mechanical work as well as being engineers.

Today, the cylinders were getting cleaned and were being measured to check their basic condition.

In the photos, Roger has the tan hat, Craig has the blue one, and the cylinder in the photo is one of the spares. The picture of the nose of the old airplane is from 1919, showing one of these engines installed on an airplane. You can clearly see how the second row of cylinders is offset from the first row.

[AAC Cadet Leader]

david paule,
very interesting! where is "here," if that's okay to ask? is that a fleet pictured to the right? what aircraft will you be putting the anzani in when you are finished restoring it?

[David Paule]

"Here" is Colorado.

No, it's not a Fleet in the photo. It's one of the 1918 - 1920 airplanes that initially used the six cylinder Anzani, a Bellanca biplane. The photo was taken in 1919 in Maryland.

The engine is a bit smaller than most of the WW I aircraft engines, and is rated for 55 or 60 hp. It's an evolution of the earlier three-cylinder Anzanis, and the company later brought out larger engines. I know that a ten cylinder Anzani was commercially available, and I've seen an ad for a 20 cylinder one, but given how lightly this one was built, that seems almost outrageous.

Our plans are to get the engine running. At this point we have no plans beyond that. This is simply an engine project.

[David Paule]

Last week the pistons got cleaned up and measured.

The pistons are cast iron with cast iron rings. You'll remember that the cylinders are cast iron, too.

The pistons have a threaded hole in one piston pin boss, for a special bolt. This retaining bolt has threads right under the head, but a smooth shank for the piston pin.

The bore is nominally 90 mm, and the piston pin is about 16 mm diameter.

One of the six pistons was damaged, with a chip in the bottom edge. Fortunately we have a set of new pistons available. Unfortunately some of them have similar damage. There appear to be enough intact pistons to choose from, though.

You'll see that the rings are identical, and that their ends have a smoothly-shaped overlap. The workmanship of the original machining is outstanding throughout the engine. Still, almost every part is marked for its location because the part sizes, one to another are not close enough to be interchangeable - at least not that we've seen yet. Since some parts aren't numbered, and of course the replacement parts are unmarked, we've been tagging and bagging every part.

This engine appears to be a first-run engine. That is, it hasn't had an overhaul before. We were debating how much run time the engine has, and suspect that it's not a whole lot.

One thing that I forgot to mention in an earlier posting is that each cylinder has two spark-plug holes. Since the engine is a single-ignition engine, that wsa at first confusing. Then I noticed that there were blind plugs in all the lower holes, with spark plugs in all the upper holes. In this fixed radial engine, that tends to prevent oil from fouling the plugs when it's not running.

Conveniently, the spark plug holes fit the same size spark plugs as my Continental O-470 engine on my Cessna Skywagon. That suggests that we should be able to find some suitable spark plugs later on.

.

[David Paule]

Rings were the topic of the day. The rings on this engine have some interesting features and if anyone knows how rings are made, please help us out here; it's interesting.

Quick background to this - we have the original pistons and rings, as removed, as well as a set of new but original (NOS) replacement parts. Some of the replacements are damaged but there's enough good stuff to put together this engine.

1. The gap between the ends of the rings is an overlap. The edges of the overlap are rounded, and there's about a millimeter to 1.5 mm between the edges looking at them circumferentially. Looking at them axially, there's almost no gap due to the overlap. The pictures show this.

Question - how were the ends of the rings made?

2. There are some manufacturing marks on the inside face of the rings, pretty much opposite the gap. At first glance these appeared to be a scale of some sort, but we decided that they were probably tooling marks.

Question - why are these marks there? Was it from tooling?

3. The rings are all like the one in the photos. That is, they're all rectangular cast-iron rings, one piece. Well, we thought that they were rectangular. Turns out that the top and bottom faces are slightly beveled, maybe a total of one degree between them. The inside edge is shorter than the outside edge, and the piston grooves match. It's intentional, because the new rings and the new pistons are that way, as well as the original ones.

Question - why was this done?

Incidentally, in the last picture below, the ring is resting on a disposable aluminum pan that's upside down. Just mentioning it because we wouldn't want you to waste effort trying to figure out what the pattern refers to.

Some of the piston pins from the unused new pistons will be used here to replace some worn original pins.

.

[David Paule]

The retaining bolts that hold the piston pins to the pistons are threaded up near the head, with a smooth shank on the bottom half. The bolt threads into the boss in the piston, and the smooth shank into the piston pin.

I was wondering why they did it that way, especially as the piston is cast iron and the pin is steel.

Then I realized that if the piston pin had been threaded, the bolt would have drawn the pin radially towards the piston on one side and would have put a load on the boss on that side. It might have even tended to load the piston pin to one side of the connecting rod.

All those issues are avoided with the bolt threaded to the piston rather than the pin.

[David Paule]

One of the questions that we're facing is how authentic should the rebuild be?

The issue came up when we started debating the clearance between the pistons and the cylinders. Right now, the clearance looks like it'll be about .010, which is somewhat large. We're not going to chrome the cylinders, but replacing the original cast-iron pistons with aluminum ones is a possibility.

There will be other decisions to make down the road regarding safety and reliability versus originality. You can imagine them easily : dual ignition, a return oil system replacing the dead-loss oil system, carburetor heat, things like that.

We'd like to know what you think about the general question of originality versus safety. Care to chime in?



Here's what we're thinking now. We anticipate that we'll run the engine and tear it down more than once. It seems to make sense therefore to at least start with a purely authentic configuration and use the disassemblies and inspections as guides to changes. If it should develop that the engine will be used in a flying airplane, as opposed to one in a museum, we'll reassess what we've learned at that time.

[David Paule]

Kind of a quick thing tonight, remember that piston pin retaining bolt? Here's a not-so-great picture of it, square head and all.
.

[David Paule]

Craig wrote this episode -

Pistons and heat transfer

We spent the afternoon of the 16th looking at pistons and wrist pins. We had previously measured all the pistons and cylinders and found that both had wide variation in dimensions. Of the good pistons we selected the six with the greatest skirt diameter for use. The final selection included four used pistons and two new ones. We then matched the biggest piston with the biggest bore, etc.

The resulting clearance between piston skirt and cylinder wall is about .010 inch (.25 mm). This is terribly large by modern standards (.003 inch today would be normal) but this engine has several differences from a modern one. Most important is the terrible cooling of the pistons. First of all, the pistons are iron and thus are poor conductors of heat so that the piston crowns don’t transfer the heat down to the skirts well.

Secondly, the diameter at the crown end is the same as at the skirt end (unlike modern engines) so this is where seizure would likely happen. To prevent seizure at the crown end, the entire piston is given large clearance. The large clearance at the skirt end results in poor heat transfer between the skirt and cylinder wall.

Finally, the Anzani has a total loss oil system. This necessitates a low oil flow to give a significant range to the aircraft. Modern engines have much higher oil flow and depend on the oil to cool the underside of the piston and especially count on the oil film between the aluminum piston skirt and the cylinder wall to conduct heat away from the piston.

Note also that aluminum has three times the thermal expansion of iron. This makes the .010 clearance even more extreme.

So why did we decide to run with these pistons and bores? First of all, it seems to be how it was designed and it did work (although with poor reliability). Secondly, to change its design to modern practice would entail extensive changes. Finally, we can always change it later after we see how it ran in its time.


Piston pins

We examined the wrist pins of the used pistons and discovered they had about .002 inch (.05 mm) of wear where they ran in the connecting rod bushings. This is a very high amount of wear by modern standards. We have several ideas for what might have caused it. First, the wrist pins are held in position in the pistons by a bolt screwed into one of the bosses in the piston and extending into a hole in the wrist pin. This keeps the pin from turning and concentrates the wear on the top (piston crown) side and bottom sides of the pin where it runs against the rod bushing.

Then there is the marginal oiling. This is exacerbated by three holes in the rod bushing. These holes are on the axial center of the bushing and at about 120 degrees from each other with one hole at the top of the rod. The problem with these holes is that a plain bushing generates the maximum oil pressure at its axial center and these holes are strategically placed to bleed that off, especially the one at the top of the rod. Modern engines have a single angled hole in the rod at the bottom (opposite the piston crown) angled toward, but not reaching, the axial center of the pin.

Another problem with the wrist pins is the tolerances on their diameter. The pins out of the new pistons showed a variance of about .001 inch (.025 mm). This might seem like precision machining, but in a modern engine, a .001 inch clearance at the wrist pin bushing is considered worn out. The reason for such tight clearances is that there is very little rotation and the load is hammering. Significant clearance allows the oil to flow around the pin (and out the holes in the Anzani bushings) and concentrates the load in the top and bottom of the bushing. Another peculiarity of the pins is that they are tapered about .001 inch from one end to the other, with the bolt-hole end being bigger. I suspect this is from holding the pin by one end when grinding, rather than grinding between centers. It could also be from worn-out equipment or even intentional.

When we attempted to put the new pins into the used pistons we discovered another problem. The pin holes in the pistons (even the new ones) are not aligned. We didn’t measure the misalignment, but it appeared to be as much as .010 inch (.25 mm). This lead to some debate as to whether we should hone the holes to a larger and aligned state and make new wrist pins (we could then hone and use the existing rod bushings) or just force the pins as was done with the original equipment and make new rod bushings. We opted for the simpler "force the pins" and new rod bushings because the pins don’t turn in the piston anyway, and the pistons have plenty of clearance so that a little distortion isn’t going to hurt anything.

[David Paule]

Joint Sealer -

The cast aluminum intake risers fit into holes in the aluminum crankcase. There's no retainer at that location, it's simply a tube in a socket. At the upper end of the cylinder, the risers are held on by two bolts. The risers have ears for the intake rocker, you'll see that in one of the pictures below. Since the cylinders are held to the case with only two long bolts each, there's going to be some motion between the case and the riser.

Roger suggested that shellac has been used as a sealer for joints like that.


Rockers -

There are two kind of rocker hinge post used on this Anzani. One is clearly newer than the other. They are the same except for the intake rocker pin support. On the old style, it's built into the aluminum riser. On the new style, it's built into the rocker boss. The first picture shows the new style and the riser.

This engine has five of the old style and one of the new style. All six risers have the old-style ears. As you can see in the photos, the ears conflict with the intake pin support on the new style.

We considered what to do about that. The first option would simply be to cut the ears off the riser for that cylinder. The second option was to cast a brand new riser without the ears. The third option was to make a brand new early-style rocker post. Then Roger had a great idea: the rocker post screws into a boss on the cylinder. Why not simply clock the rocker 180 degrees and use the intake boss on the riser? That way nothing needs to be made and nothing needs to be permanently altered.

It turns out that this will work. There's no interference with anything. The unused intake pin boss on the new-style post will get a free ride. The post will be half a thread farther out from the cam, so we'll need to adjust for that. The threads in the post to the cylinder are about 1.49 mm per turn, so the extension will be about 3/4 mm.

On this cylinder, and I think this is the only one set up this way, a small screw goes radially through the boss on the cylinder into the rocker post to retain it. It is then safety wired.

The new-style post is mounted to the cylinder boss, which is about 5/32 higher than for the old-style cylinders. The cylinders look alike otherwise, so maybe that's simply the tolerance the engine is made to. The exhaust valve pushrod is about that much longer than the others.

We needed to remove the intake rocker hinge pin on the new-style rocker post. This hollow pin was swaged into the post. This is typical for all the rocker pins. The third photo shows the Roger chiseling the end of the pin off.

The pin has three radial oil holes at 120 degrees apart. They are offset axially from each other, but all are under the rocker.

The rocker pins are oiled by squirting oil into the hollow pin until it leaks out past the swaged ends. There's no other lubrication source. I guess it would be a thick oil.

That raises a question about the viscosity of castor oil. Does anyone know what it would be equivalent to when it's at operating temperature?

We anticipate that we'll need to replace at least this one pin and probably all of them.

The final picture shows the pin in the new-style exhaust pushrod that retains the end fitting. With the nut loosened, there's some perceptible movement between the fitting and the pushrod. The little pin (which is probably a rivet of some sort or a pressed pin) carries the pushrod load, but the preload from the nut stabilizes the fitting and converts most of the alternating load to a steady load with a smaller alternating load.

Incidentally, the intake rockers have a trunnion for two intake pushrods. All the intake rockers are of the same design. I was going to say they are "identical," but since they haven't been measured, we don't know how much they actually differ one to another.

I think that the engine designers needed two intake pushrods to clear the intake risers. The cam has two intake valve tracks, so there's some redundancy there.
.