A Design Question (Engine Cylinders)

Engine Cylinders: Inverted-vertical compound; 1.25 and 2 inch bores X 2 inch stroke.

No castings: cut from solid cast-iron bar stock.

Original made rather too long a time ago but with unforeseen design errors that result now in valve-chest studs entering the passages, little room for the plumbing, and very awkward gasket-making.

Solution: make a new block!

Material to hand- continuously-cast C.I. bar long enough and high enough, but barely 60mm wide even before trimming the faces.

Option1: Skim the worst of the lumps and bumps off, sandwich the block between two 10mm / 3/8" thick steel plates secured with countersunk screws, with a layer of filler in between. The screws will be hidden by a thin cladding.

This will give a block capable of taking a 3-inch dia LP cover, with the outer 1s or 2s of its studs in the steel not the iron, depending on the chosen pattern to miss the steam passages.

Drawback: the LP cylinder wall between the bore and joint will be hardly over 1/8" thick at its thinnest, though fully-supported. The steam in there is likely to be little more than hot squashed mist except when the starting-valve feeds it with hot squashed fog. (Boiler pressure 90psi, no superheat.)

.

Any likely problems there, apart from some very careful centering to bore out the cylinders?

Or does consensus think I'd be better going for:

Option 2 : Buy a new block of iron big enough (80 thick) to hold the lot in one lump?

Advantages: inherently stronger and room for the covers without needing side-plates. Less machining.

Disadvantage: cost, mainly.

The original by the way, is rectangular in outwards appearance right across the valve-chests and cylinders (archive photograph) but probably due to cladding. The casting itself was probably a lot more complicated, with lightening recesses etc.

Option 3 separate flanges top and bottom. Held with CSK screws and JBWeld. I usually make my cylinders this way with either JBWeld or silver solder depending on material

Ramon has an album with his marine compound in it which shows the method well, you also get to easily shape the cylinders to reduce weight and have space for lagging

Liquid gasket can solve the last issue with your existing block

Edited By JasonB on 20/10/2022 07:16:56

Thank you Jason.

I see what you mean but it would still leave the problem of two vertical "strips" of wall barely an eighth of an inch thick; although top and bottom plates will hold things together.

It is that thin area on both sides that worries me - unless I am worrying too much.

'

I did think of making the cylinders as two separate units made from round bar, sandwiched by top and bottom plates carrying the cylinder covers, but that gives no joint-free way to connect the valve ports to the cylinders.

I've also contemplated machining the existing block heavily to cut out the passages that are the nub of the problem, but I'd still run into difficulties with the old drillings and various studs.

'

Afore anyone says so, yes I know I should attach a drawing, but I don't have one to attach! That won't come about until I am sure of my best approach.

Fag packet sketch would be a lot more helpful than guessing plus a picture of what you already have.

Nigel, if you study the images of the cylinder build that Jason points you to you can see that the passage ways are all milled in before the final assembly. I packed the passage ways with soft polystyrene packing to ensure the JBW did not get into them. Once all set dribbling cellulose thinner through the ports dissolved the polystyrene. Worked perfectly.

Don't know if that will help your particular situation but this is the second time I have made a cylinder block in this fashion and am very satisfied with the results

Tug

This is a sketch of Option 1, exploded. It is to scale using the candidate block of iron; but I have omitted all details including the ports - which would have been very difficult to draw.

The ports are dug out of the end faces so there is not very much depth left for the LP passages. The 52mm (2" ) LP cylinder wall is only 15mm from the port face at its closest, and on the arrangement above only about 3mm from the sides of the casting. The screwed-on plates are drawn at 10mm thick, the block nominally 60 mm wide, to hold an LP cylinder cover 75mm / 3" diameter.

The ports, the overall length over port-faces, and the valve-spindle and cylinder centres, are all set. I can't change them.

.

The existing block is much wider, but I drilled the passages parallel to the port faces, so quite shallow-lying, and when I drilled the valve-chest studs years (literally) later, some go straight into them.

The passages meet the bores by milled recesses that leave very narrow lands for sealing at the ends of the block, though the gasket would receive some support from the valve-chests - if I had thought to make the chests slightly over-long.

This arrangement was easy to machine but creates disproportionately large clearance volumes. One of Luker's locomotive drawings in ME showed me what I should have done!

I've also now problems trying to find room for the HP steam inlet, and the connecting and starting-valve pipes or passages. The ancient photograph that is my GA "drawing" shows the steam pipe meets the back face of the HP valve-chest via a fitting with a long vertical lever, which I take to be the starting-valve. The regulator is merely a globe-valve on the top of the boiler!

.

The attached drawing, in TurboCAD, is on my 3D CAD limit. I aligned 3 rectangles and 2 circles along the Y-axis, spaced then extruded them with the circles over height and depth, so subtracting the "solid" cylinders from the block left the two holes.

The facets are a result of how TC extrudes circles. It imprints a polygon on the base circle then extrudes that to form a prism it calls a "TC surface", but you can edit the appearance by raising the facet count - I chose 60 here. These are horrors - some basic editing tools break the prism into separate facets you cannot reunite; and I have never discovered what if any pattern there is to it.

I don't have drawings for the original block. To investigate milling out the original passages and fitting inserts to give properly-inclined passages, I found drawing it in 3D CAD far too difficult. If I ever had a paper drawing of it, drawn some 20 years ago, it would be on a sheet too large to scan.

I have just examined Ramon's photographs.

That is very impressive work! My engine is simpler anyway - 2 not 3 cylinders, but also has to be to be within my "designing" and machining abilities.

His methods do though suggest I could look again at modifying my existing engine-block.

One path might be to cut the block right back to the floor of the passages; fit a port-face plate with new passages where they should be; and make new valve-chests nearer the width of the block, with re-arranged studs giving space for the "plumbing" while matching the photos.

The engine is very prominent on this vehicle, as it is between the seats, with the full-sized cylinders next to the driver's chest.

If you take a look at sat the Stuart compound then the chest covers are 1/8" think around the edges, allowing for the cast in "S" then you probably only have 2mm thickness to play with and that's a flat surface and you don't hear of them going pop.

Any reason the cylinder covers have to be round? a lobe or similar could take care of any thin sealing surface that the end notches cause

As for the stud holes into the passages plug them and move the stud positions or just seal the studs well into what metal there is.

Like Ramon I often mill the passages into a block that fits to the side of the cylinder though with 15mm to play with you have plenty

I think I sketched something out a while ago in one of your CAD threads, if you did vertical drilled passages rather than the angled you would need even less metal.

Your images there are more or less as I have made -

- but I drilled the passages parallel to the surface and quite shallow, and the valve-chests are separate so I can machine the port faces flat.

This makes the "keyhole" recesses approach the end of the block very closely, greatly reducing the sealing area. It also gives insufficient depth for the stud holes without entering the passages; and prevents the direct use of round covers.

My plan was to fit a steel plate top and bottom, with the circular covers themselves spigotted into that; but the lands at the ends are too narrow for reliable sealing.

A lobed cylinder cover may work, if there was anywhere sensible for the lobe screws, but would look wrong. The top of the engine is in full view with round, 6-stud covers on a big flat surface. (11 studs as one is a shared central one.) I am hardly Highly Commended league, let alone the Hewson / Seymour-Howell level of quality, but I do want this thing to look somewhere near its prototype.

The valve-chest stud dispositions I have used are much as your drawing... only they give precious little room for the HP steam inlet / starting-valve (on the chest wall itself), and not a great deal for the valve-rod gland.

.

I don't think the valve-chest covers (mine are mild-steel) and the cylinder walls are sufficiently comparable. The problem with this 60mm raw-width block for a 52mm diameter cylinder, is if the wall would be strong enough even if flanked by a plate each side.. The area of concern would be adjacent to the side stud holes, on your image.

.

Taking these points into consideration, and looking at your examples, Option Three is looking more feasible: drastic surgery on the existing block, which is a healthy 110mm (about 4-1/4" ) wide. I have allowed 5 - 6 inches wide for the engine case below it, whose original photographs show a distinct step outwards to be wider, below an area with access holes to the stuffing-boxes.

.

[ I cannot produce drawings like yours, by the way!]

I have started drawing the existing block to investigate modifying that, by machining it down to the floors of the passages and fitting new port block - JB Weld and all.

It will need new valve-chests too.

I have worked out how to secure the blocks and chests by what are essentially double-ended bolts - studs with a cylindrical head at part-height, that fit counterbores to hold the intermediate parts secure while leaving the valve-chest covers removable.

So far it is a 4-elevation drawing in TurboCAD, though halted by a very strange snag in creating the rather involved passage/ port cross-sections. Not an engineering problem but a drawing one I can't identify, in a technique I have used many times previously without trouble. Most odd!