18 Cylinders Isotta Fraschini (straight six-cylinder x3 )

[Foketry]

Foketry:

It looks like the casting for the top half of the crankcase cleaned up fairly well when you machined it. There's a couple of spots that look a little iffy, but some JB Weld would fix them right up. In that last photo, it almost looks like the right-hand bank of cylinders has sleeves in it, or are my eyes playing tricks on me?

Don

You have seen correctly, both the right and left banks have sleeves, 7075 aluminum to avoid permanent deformation on the 2 most stressed banks. The length of the coupling between cylinder and cylinder seat is only 9 mm, I could not increase it, the cylinders inclined 40 degrees touch the central cylinder.
Furthermore the cylinders are fixed to the engine block by means of a separate flange and a retaining ring, a not very stable solution, but to assemble the water jacket inserted from underneath I have not found a better solution .

 

[Foketry]

3-way T-fittings for inlet and outlet of cooling water for cylinders
I tried 3 possible alternatives to build 40 of these T : aluminum casting, but when I printed the model with 3D printer I realized that the dimensions are really small, impossible to do these with my poor equipment, a jewelers centrifuge would be needed.
Welding of 2 small tubes 4 mm external diameter, I tried with aluminum tubes, but it is really difficult to weld aluminum of small size, the welding temperature is high and a few more seconds make the tubes melt.
The final solution was to obtain them from a rectangular aluminum bar by milling.

3D printed for casting mold

aluminum welded tubes

milled pieces

 

[CFLBob]

The milled pieces definitely look like the way to make those.

One question. How did you do the third tube, the one that gets threaded? Are they threaded so that the start and end of the thread always leaves them aligned your sample?

 

[Foketry]

The milled pieces definitely look like the way to make those.

One question. How did you do the third tube, the one that gets threaded? Are they threaded so that the start and end of the thread always leaves them aligned your sample?

I milled the T in 4 subsequent steps, in the photo you see the first and second step, deep drilling and contouring milling of the 40 pieces, rotation of the bar by 180 degrees, contouring milling, third step drilling and contouring of the threaded side. In this third step the pieces are separated from the bar, fourth step manual threading.

 

[ajoeiam]

I milled the T in 4 subsequent steps, in the photo you see the first and second step, deep drilling and contouring milling of the 40 pieces, rotation of the bar by 180 degrees, contouring milling, third step drilling and contouring of the threaded side. In this third step the pieces are separated from the bar, fourth step manual threading.

Hmmmmmmmm - - - you didn't consider thread milling the threads?
(Not that your parts don't look great - - -grin!!!)

 

[Foketry]

the cylinder cooling water liners are finished and painted black like the original engine

Primer and epoxy paint

 

[Foketry]

cylinder liners assembly test, without water jacket, with Oring
Upper oring is 1.78 mm, lower oring 1mm

 

[Foketry]

The cylinders painted and assembled to the engine block
Now I have to check the compression ratio with the crankshaft, connecting rods and at least 3 pistons assembled. The theoretical calculation was done during the design, but it's better to do a practical check as well .
Then I have to flatten the entire bank of 6 cylinders exactly at the same height as there is a single head for 6 cylinders,3 heads for 18 cylinders.

to check the compression ratio I must also take into account the geometry of the piston with this central niche due to the position of the spark plug

 

[petertha]

Now I have to check the compression ratio with the crankshaft, connecting rods and at least 3 pistons assembled. The theoretical calculation was done during the design, but it's better to do a practical check as well .
Then I have to flatten the entire bank of 6 cylinders exactly at the same height as there is a single head for 6 cylinders,3 heads for 18 cylinders.
to check the compression ratio I must also take into account the geometry of the piston with this central niche due to the position of the spark plug

Nice work. Silly question, but does the master/link rod assembly deliver the piston to the same height at TDC & are compression ratios therefore equal between center & side cylinder bank? I know on radial engines some parameter needs to be slightly compensated for either equal CR or position, either the link rod assembly or piston/top end. It may not be enough to fret about, but just curious.

 

[Foketry]

Nice work. Silly question, but does the master/link rod assembly deliver the piston to the same height at TDC & are compression ratios therefore equal between center & side cylinder bank? I know on radial engines some parameter needs to be slightly compensated for either equal CR or position, either the link rod assembly or piston/top end. It may not be enough to fret about, but just curious.

In this engine the axes of all the 3 cylinders that are part of the same master rod converge at the center of the rotation axis of the crankshaft so that the stroke is the same for all 18 cylinders and therefore the compression ratio is the same. All pistons reach TDC at the same height.

 

[Peter Twissell]

I think that's not quite true, Foketry.
Assuming you have compensated for the master rod angle when positioning the slave rod eyes, at bottom centre on either the left or right cylinder, the master rod angle is different, so those cylinders have slightly different stroke to the centre cylinder.
I created a spreadsheet to calculate balance for my large 7 cylinder radial.
The spreadsheet calculates the travel of each piston and has options for compensated or non compensated slave rod eye angles.
I'm happy to share it of anyone is interested.

 

[Steamchick]

I agree with Peter. I once found (somewhere?) on utube an explanation of why longer con-rods produce more power than shorter con-rods (for the same stroke), and a few other things that I had "forgotten" from my education in the Japanese design office back in 1989. Many things have moved on over the decades, but not geometry. I suggest you simply check dimensionally from the top deck to piston at BDC and TDC for 3 pistons connected to the same journal and compare to crank angle. It will give you the information for compression ratio per bank, and you may then decide to develop any variation of other parameters to compensate for anything you deem worthwhile. You can even plot stroke in intervals of 10 degrees of crank and produce some nice curves on a graph - if you feel inclined (Pardon the pun?).
There are many articles on the web about engine design aspects. e.g. The Long and Short of Connecting Rods


https://www.enginebuildermag.com/2016/08/understanding-rod-ratios
But none of these really explains the secondary big-end centre motion of the radial - or 3-bank Vee - engines, as found for the outer cylinders of this engine.
peter, does you model produce a graph of stroke versus crank rotation angle? - That may be interesting for us "observers and hangers-on" to this thread.
Ta,
K2

 

[Peter Twissell]

Hi k2,
Yes, my spreadsheet plots the location of all pistons through a full crack rotation at 1 degree intervals.
It also calculates all the mass moments of the moving parts and the resulting balance of the engine for a given counterweight.

Pete.

 

[Steamchick]

Hi Peter, That sounds like a proper job!
Well done!
K2

 

[Peter Twissell]

Spreadsheet here:
https://www.homemodelenginemachinist.com/threads/radial-engine-balance-calculator.34677/

 

[Foketry]

I think that's not quite true, Foketry.
Assuming you have compensated for the master rod angle when positioning the slave rod eyes, at bottom centre on either the left or right cylinder, the master rod angle is different, so those cylinders have slightly different stroke to the centre cylinder.
I created a spreadsheet to calculate balance for my large 7 cylinder radial.
The spreadsheet calculates the travel of each piston and has options for compensated or non compensated slave rod eye angles.
I'm happy to share it of anyone is interested.

The CAD images below demonstrate that the piston stroke is the same for all 3 cylinders, right, center and left, stroke 30 mm for all 18 cylinders, the piston bore is 24 mm for all and therefore the compression ratio is the same for everyone. Am I doing something wrong?

Central rod BDC

Central rod TDC

Right rod DBC

Central rod TDC

 

[petertha]

I've been meaning to work out some kind of simplified CAD tool to evaluate these sorts of engine kinematic issues anyways. (I've also done something similar in Excel but starts to get a bit complicated).

Can you provide me more dimensional details of the master rod assembly so it is fully defined? For example layout of the red cross (defines position of link rod bottom ends & crank throw rotation center etc.). Is the vertical rod dimension (pink) exact same as the left/right link rod length (blue). What is the crankshaft throw? What is angle of cylinders viewed from front?

The way I'm trying to solve is: make the master rod a fixed block element, insert into engine layout background, adjust alignment & angular constraints to some position of interest, then evaluate resultant position of link rod wristpin point which is indicative of TDC. We can add pistons in too, but but for purposes of CR evaluation, same thing. Example (with random guess measurements) shows TDC of right cylinder bank.

 

[minh-thanh]

In my opinion, if 3 cylinders are all on a circle and the connecting rods of the left and right cylinders are also on a circle with the center of the connecting rod in the middle, the length of the connecting rod and the compression ratio is same

 

[petertha]

Maybe this sketch will help (assuming I understand the engine layout correctly). I tried to superimpose the snapshot position where the right cylinder link rod is assumed at maximum throw (TDC for right cylinder). But sketch also illustrates the problem viewing the rod assembly in isolation. The master rod (purple) must also be simultaneously constrained by the middle cylinder centerline (dash red line), which it is not. If it were rotated like blue annotation, the link rod would no longer be at maximum reach, aligned to right cylinder. In other words I don't think this layout can exist as shown. The master rod geometry + constraints collectively 'drives' the link rods.

Not to detract form this wonderful post but I asked the question because it reminds me of the top 3 cylinders of a radial engine. The Edwards radial for example has an equal (uncompensated) master rod layout. Each axle is 360/5=72 deg apart. The distance between master rod (crankpin) center to each wristpin are also identical. The cylinders are also 72 deg apart. But similarly, when you superimpose the MR onto the cylinder layout, constrain the wrist pin centers to occur along their cylinder axis & compare resultant distance at TDC to a fixed datum like cylinder top, they measure slightly different. Therefore slightly different CR. So to achieve equal CR, you must either alter the cylinder length (as Edwards specifies) or compensate the angular position of the master rod axles. How much of either depends on the physical dimensions.

I hope this makes sense. I thought it worth questioning because its kind of important aspect of the engine.
Left TDC comparison for #1 & #2 cylinders Right = identical distances wristpin to master rod center

 

[Mechanicboy]

To understand better why we has difference angles on the master rod connected to other connecting rod..

With 2 different lengths from the crank pin to the piston pin in the connecting rod, the piston does not have the same speed and side pressure in the same stroke length, therefore a specific angle must be set on the master rod in order for the other pistons with short connecting rods to have approximately the same piston speed and side pressure on the cylinder the wall.