Category Archives: The Power Plant

Building the Engine and Boiler
Below are posts related to the building of Arthur Leak’s marine compound steam engine as originally described in Model Engineer magazine and more recently provided via Camden Books (drawings, manual and castings). It will also include the construction of a oil-fired Yarrow watertube boiler, to designs from John King from the SBA (see links on right of home page)

This is the engine that will power SY-Befur, it is expected to produce 10-14HP. These post cover all the construction from Camden’s castings and the addition of pump and alternator assemblies for our particular installation.

A Plumber’s Nightmare & a Real McCoy

Over the last few days we have encountered the two items mentioned in the title in real life, in a slightly stressful way.

The Real McCoy

One of Elijah McCoy's displacement lubricators - actually this one was made by the Detroit Lubricator Company.

One of Elijah McCoy’s displacement lubricators – actually this one was made by the Detroit Lubricator Company.

While Wikipedia suggests two origins for the phrase “The Real McCoy”, the most well documented version relates to one of the brass beauties shown here.

It is a displacement lubricator patented by one of Elijah McCoy’ in the 1870s in America.  These devices perform a simple, but vital, role of providing internal lubrication for steam engine cylinders and valve gear, but they do it using an apparently impossible process.

Basically they are attached to a ‘T’ in the steam line supplying the engine. You fill the lubricator with oil, and then via a mechanism which is not at all intuitive, the steam in the steam line to the engine decides it would prefer to be in the lubricator and, as it migrates that way and condenses as it cools, displaces the oil which is forced back into the steam line. So the connection to the engine sees steam heading one way and oil heading the other with nothing obvious causing that to happen!! The lubricator has a number of valves to control the rate that this happens and allow the machine to be shut off and refilled while underway.

Apparently Elijah McCoy’s  lubricators were so good and reliable that companies wishing to purchase steam locomotives were given to checking that it was fitted with a “Real McCoy Lubricator” – hence the phrase 🙂

In my case I purchased this at an auction a few years ago, and prior to fitting it I needed to pressure test this (as with all other pressurised components). As I did this, it revealed a number of leaks and failed seals on the two sight glasses.

Taking it apart revealed seals that may once have been rubber, but in the intervening years (100+?) had turned into something more like wood. I managed to find some replacement ones that just needed shortening, but fitting these is a tense process as cracking the glass would be just too easy.

Anyway, we managed it, and with a little TLC and returning some of the needle valves, we had the Real McCoy: a leak-less displacement lubricator.

The Plumber’s Nightmare

The fwd boiler fittings almost complete.

The fwd boiler fittings almost complete.

The rest of the week has been consumed by attempting to fit the fittings to the forward end of the boiler.

This seemed as though it would be a simple process, but pushed me to the edge of serious depression.

Due to the temperature and pressure of this assembly (250psi and 200+ degrees Centigrade) all of this needs to be in steel pipe with screwed fittings.

The problem is that while one can imagine how it all goes together, the reality is more complex, principally because you can’t actually screw all the parts together as the end points are fixed and so you can’t tighten everything.

Moreover it appears to need a certain sort of brain/thinking to figure this out – my respect for plumbers has been raised considerably! This was born out by half an hour in Penrith at the sales desk of a pipeline and hydraulic supplier while the staff demonstrated considerable fortitude (and difficulty) in “rummaging” through their stock to find the combination of bits we needed. (Actually I found this quite encouraging – if they couldn’t sort it out, perhaps I was not being a complete klutz.)

The net of this story is that what you need are “cone unions” – and I think the ones provided by Bessegers (like these) are by far the best design. These allow you to assemble the screwed bits, and then attach the assembly to the component you are plumbing without needing to move the joints you have made.

The second take away, is that sometimes you need to opt for an indirect path for the piping to allow you to accommodate the offsets in all three dimensions.



Boiler Gauge Glass and Boiler Fittings

Boiler Gauge Glass

Finished Gauge Glass

Finished Gauge Glass

We decided to make the sight glass for the boiler following John King’s design, with slight modifications. Ian Cross of the SBA was very helpful and modified some existing patterns he had for “normal” reflex gauge glasses to suite rear-entry installations, and had 3 sets of these cast for our “boiler making syndicate”.

As it happened he made the castings with larger/longer mounting “lugs” so I decided to try to make the gauge with the cocks integrated into the body (as opposed to separate valves as drawn by John). This was a quite stressful decision as the cocks are not easy to make, and any errors result in a scrapped casting, but none-the-less I am quite pleased with the result.

The design is unusual as the lower cock is made as a three-way cock to allow the two cocks to be set for normal operation, isolation of the glass (in the event of failure of the glass) and to allow blow-down of the glass and drum – a neat and innovative solution.


The manufacture was quite simple:

    • The flat faces of the casting were all faced on the shaper (using stop-pins) to hold it to the shaper table. This approach allows the complete faces to be machined in one pass.
  • Milling the Glass Recess

    Milling the Glass Recess

    The body and cover castings were then attached to the mill table and the recesses were cut for the glass (a B2 Klinger Reflex Glass from Heritage Steam Supplies).

  • At the same time the holes (ports) to connect the drum to the gauge glass space were drilled. This required some careful measurement as I made my usual mistake of not really understanding how the gauge fitted into the casting (it’s a bit like the sculptor finding the subject hidden in the stone) and so had not realised that the pre-cast slot for the glass was not actually arranged to be at the longitudinal centre of the casting – so the ports to the glass slot had to be offset from the threaded holes for mounting the gauge. As all these holes had to also mate with the taper cocks taper part it needed to be done carefully.
  • The next process was to drill and bore the holes for the taper cocks. Firstly, the cocks themselves have to be turned with the correct taper – the important thing is that the cocks and the holes are turned at exactly the same setting, so one sets the top-slide to the 10-degree angle and then locked in place while both parts are made. Here are some pictures of the parts and video of setting up for the the boring using a sharpened rod (located in a dot-punched mark at the centre and a closed up tailstock chuck, which provides a female centre point) and dial gauge to get the casting correctly positioned on the face-plate. This is a very off-set operation so large amounts of junk have to be bolted to the face-plate to attempt to balance everything.
  • Boring the drain port

    Boring the drain port

    The last operation is to drill and centre-bore the drain hole from the bottom of the casting to the taper-cock bore. This needed to be accurate, so I drew it up in ViaCad to ensure I had accurate dimensions and angles. Then a digital angle gauge allowed the head on the Rambaudi Mill to be set over to drill the hole (rather a “tall” set-up).

  • The final assembly showed that the handles on the cocks clashed with the dome nuts on the steam drum ends, this could probably have been avoided by shortening the shafts of the cocks, but this would have required bent handles (as described in the drawings) but I did not have any suitable material. So I made two spacers to lift the gauge clear of the end-plate. these have o-rings installed on both sides to seal the assembly.

Boiler Fittings

The next stage of the process is to attach all the boiler fittings and then assemble the steam/exhaust/feed-water components to the engine so that we can complete the boiler test and check everything operates before installing into the boat.

plumbing diagram

plumbing diagram

plumbing Bill of Materials

plumbing Bill of Materials

En-route I drew up a plan for this to choose the appropriate fittings and a bill of materials to match.

This single sentence hides many hundreds of pounds of components and hours of hand-wringing and frustration!

PTFE – one lesson that has been learnt is that PTFE seals in clacks and valves will not work at 250psi, as their maximum operating temperature is about 190C  (saturated steam at 250psi is at 207C or 406F) and at that pressure the operating pressure limit is very low – so PN32 valves are no-go and PN40 or above is what is required. (So I have quite a stock of ball and globe valves which are going to need to go to eBay 😦 ).

Here are a couple of pictures showing the current state of play…

Boiler Fitting in progress

Boiler Fitting in progress

Boiler Fitting in progress

Boiler Fitting in progress


All Pumped up!

Well a definite milestone was reached today; the boiler passed its official initial hydraulic test at 500psi conducted by our Boiler Inspector.

It will never need to be pressed that hard again, next we have a 375psi test with all the ancillaries fitted (gauges, valves, plumbing etc.) then we put some fire in its belly and prove that the safety valves will stop the pressure going more than 10% higher than its 250psi operating pressure – then we will be allowed to insure it and use it in anger!!!

This might seem like a bit of a palaver, but a boiler failure will typically kill everyone within many feet – so it pays to take care.

When you consider that the pressure trying to force each end cover off during the test is equivalent to a large African Elephant standing on it, you can see why it uses a bunch of M16 and M20 studs to hold everything together!

Boiler sitting at test pressure.

500psi on the clock

500psi on the clock

A Few Notes

It took three attempts to get everything perfect, and here are a few notes for other builders:

  • The dome nuts needed to be skimmed, as the imperfections in their surface were enough to produce a significant shower at 200psi as they prevented the dowty seals from working.
  • One of the tubes had not been expanded into the mud drum – that started leaking without any pressure!
  • The stays on the steam drum needed shortening by about 200-thou to stop them bottoming in the dome nuts (it’s a bit of a fiddly dimension, as you must ensure enough thread is engaged with the nuts even with one end completely screwed up to the limit.)
  • The upper end of some of the tubes were a bit long and clashed with the lower stays in the steam drum, resulting in an hour or so with a Dremmel and air-driven countersink, to provide the necessary clearance.

Either way, an important day on the route to launch day!

Boiler Insulation and Funnel

Just a quick note on recent days’ work. We have been insulating the inner boiler casing and installing the funnel (at a jaunty angle)!

The boiler has two casings- one surrounding the burner, tubes etc. (the hot stuff) and an outer one of wood, with an air-gap in-between to keep passengers safe. This also includes a double skinned chimney (also to keep people safe – Lou has quite a scar from another steamboat where the funnel was not lagged or double skinned).

We are using “ceramic fibre board” from Vitcas This is quite expensive, but robust, capable of standing 2,300F (1,250C)  and capable of being cut with a knife or jig saw.

However, the mental gymnastics required to think about how it all fits together inside the boiler I find quite hard. However, after a deal of cutting and trial and error it’s almost done.

Tube nest

As we needed to remove the casing, I also took some pictures of the “tube nest” which shows the 200+ 12mm tubes all expanded into the steam and mud drums.

This lot totals about 30sq-ft of heating surface which is capable of producing about 300lbs of steam at 250psi per hour!


This is formed from some 7-inch stainless steel liner and enamel outer flue. This is what you would find attached to your average AGA.

Boat aesthetics required that this is attached at a jaunty angle (5-degrees) to provide an impression of dashing speed etc. However, making things at an angle is often more difficult than making things straight.

We have made the casing from 3mm steel and this has provided a solid foundation for mounting the funnel.

The inner liner is  178mm dia, so I bored out the hole in the top plate of the casing to size with the head of the mill set over 5-deg to get the required oval (a bit OTT I think).

I opted for a set of 20mm square blocks with their base machined to the aforementioned 5-degree angle and then M5 screws are used to attach these to the roof of the boiler casing to provide a “socket” for the liner.. The inner liner also has a “bulge” about 15mm above it’s lower end, so I machined notches in these blocks to engage with this bulge and thus provide vertical (or jaunty angle) support to.

The outer face of these blocks was then machined to fit the outer funnel, again with the steps providing the angled support. M5 screws thru the outer into these blocks secure the outer funnel. Lastly, a set of three blocks were made to secure the inner and outer funnel together at the upper end, and then an air-nibbler used to chop the inner to length.

Milling at 5-degrees

Milling machine set over to machine oval hole in cover, and inner funnel resting in said hole at required “jaunty” angle!

Funnel mounted on Boiler Cover

Funnel mounted on Boiler Cover

Mid 2017 Update

Progress since April

Well, it seems like high time I provided an update, as the last one was in April!

At some level it feels like not much has been achieved, but that’s because a lot of the work has been “bitty”, finishing up jobs and tidying up items that had been hanging around for a while – and then there was the distraction of needing to design/build a new garden shed (the last one literally blew down – the joys of living 900ft up in the Pennines!).

So here is a list of the items I can recall completing….

  • Finishing the inner Boiler casing – next job is to “lag” the inside with ceramic board insulation
  • Making a manifold for the feed clacks – basically milling and turning off about 80% of a steel block.
  • Remaking the battery pack for the VHF transceiver – no replacements available.
  • Testing the antique Sailor VHF radio – (using the aforementioned transceiver)
  • Rebuilding and modifying the lubricator pump and plumbing to fix leaks – (correction; most leaks!)
  • Making a sump/oil tray for the engine – expensively made from spare 3mm brass sheet!!!
  • Repainting the condenser – maximum Nitromors, but looks better.
  • Finishing steam re-heater  – making unions, and lagging in “broken bone” plaster bandage.
  • Plumbing in the condenser steam and cooling water circuits – lots of cursing, custom unions and silver soldering.
  • Fixing the pump/alternator assembly to sump – decided the floating design was no good.
  • Craning the engine and boiler around ready for testing.
  • Spent a fine day on Grayling on Windermere – we all need a break sometimes!

Next Steps

  • I think the engine is now effectively ready to install into the boat, but we are going to bench test the whole shbang before we do this.
  • Strip the boiler casing and fit the insulation.
  • Mount inner and outer funnel onto boiler.
  • Screw cut the M20 and M16 stays for the boiler (thanks John for loan of larger lathe).
  • Make water gauge – modified castings arrived (thanks to Ian Cross for the pattern making).
  • Assemble and pressure test the boiler!!!!!!

I have assembled a slideshow of photos to record some of the above items, rather than post them all individually – enjoy!


Finishing the Superheaters & Boiler Casing

Work on the Boiler continues with the finishing and installing of the Economiser (pre-heats the incoming water to the boiler using waste heat from just before the flue) and the Superheater (adds energy to the steam on the way to the engine, again using waste heat from the flue gasses.

Milling and Drilling the Econo/Superheater Headers

As noted in the last post, I decided to mill the recesses in the two halves of the headers, as there is a lot of metal to shift, and with the “ripping” milling cutters this was by far quicker. (some pics)….

Milling Recesses in Headers

Milling Recesses

O-ring milling set up

O-ring milling set up

Some finished headers

Some finished headers





Tapping Machine 😦

As there were such a lot of M8 holes to tap (about 48 per boiler) I thought I would “invest” in a tapping head for the mill, and went for a cheap Chinese one – and you certainly get what you pay for… I was worried when it arrived as it appeared to not have a load-sensitive clutch, so decided to try it gently on a test piece… the pictures show how that worked out…

The £^*kin Tapping Head

The £^*kin Tapping Head

Machine Tapping in Action

Machine Tapping in Action

I can imagine it would have been OK with tapping thru holes, but even being as careful and slow as I could, it was useless on a blind hole -sheering the tap instantly. I was just very happy that I snapped the tap in a test piece, and did not scrap a header that I had spent some time (and money) on making… So it’s back to standard/hand mode… and we have a full set…

Back to the old method

Back to the old method

Some finished headers

Some finished headers

Cutting the O-ring Grooves

The final job on the headers is to cut the 4.5mm wide groves for the o-rings that seal the joint between the two halves. I was a little nervous about cutting these in the lathe, as you need to produce a narrow tool with additional clearance to avoid the outer edge of the groove as it is cut.

O-ring milling set up

O-ring milling set up

So, again I decided to mill these, using a small slot drill and the rotary table. This did work, but was slow and tool wear was a problem – I might revert to lathe cutting these for the drum ends.  The setup for milling the grooves was to fit new clamps in the recess before removing the clamps on the rotary table that were holding the header while recess milling (see pic). This meant we knew that the part was still centred on the rotary table, ensuring the recess and o-ring groove were concentric.

With all this complete we just needed to fit the tubes with the same expanding process as used on the main drums.

Boiler Casing

You will recall I drew the parts for the inner casing of the boiler in ViaCAD and then had them plasma cut directly from the drawing; this meant assembly was quite simple. I used M5 rivnuts and screws to make the assembly, spotting through from the plasma cut holes to locate the rivnuts. This went well. I also purchased an American design of tool for fitting the rivnuts, this is based on a pair of wedges which produce the compression to expand the rivnuts. The more normal “pop-rivet” style of tool requires bigger hands and forearm muscles that I currently own!

Completed Inner Casing

Completed Inner Casing

Starting the casing

Starting the casing

Pressure test

The last job is to pressure test the assembly, as the boiler is rated to 250psi, the hydraulic test (with the assembly full of water to avoid any large bangs!) is 500psi. Once we located the one tube we failed to expand (light shower ensues) the assembly proved pressure-tight – which is more than could be said for the plugs and pump, which took several hours of work to fettle to a point where they would stand the pressure. (In the process I became a complete fan of Dowty Seals, which rendered the plugs in the headers pressure-tight with little more than finger pressure inserting the plugs. Here is a video of one of the completed assemblies under pressure (before I cured the leaking pump!!).

Not a bad months work in total…

Tube Expanding and Economiser Headers

Just an update on progress with the boiler and other (interrupting) activities.

Tube Expanding

Nigel was good enough to make the trip north and assist with the tubing of the first boiler.

This was actually a simple, if repetitive, job.

Fitting the tubes

We firstly supported the three drums in the casing end plates, which were temporarily joined with lengths of timber. When fitting the tubes the game plan is to first locate the “centre” row of straight tubes, then fit the outer rows.

In practice there is not enough clearance to insert the first end of the tube with the drums in their correct orientation, so one firstly turns one of the drums in the casing to allow one end of each of the first row of tubes to be slid into this drum, and then the drum can be rotated back to its correct orientation and the free end of the tube inserted into the other drum.   (Hint: Insert them into the mud drum first, we actually started by inserting them into the steam drum first, but they fall out as you rotate this drum).

Once the straight tubes are in place then the outer rows can be inserted by inserting the tube into one drum and pushing it in as far as the first bend, this then provides the clearance to allow the  tub to be swung into place and then you can slide the free end back into the other drum.

Once the tubes are in place you can commence expanding (in fact we expanded the straight tubes into place before inserting the bent ones. Drilling all the holes with a 12.1mm drill actually worked well and provided just enough clearance to allow assembly – any tighter and it would have been a real struggle.

How much expansion?

Before expanding we did the maths on how much expansion was needed. The idea is that firstly you expand the tube to take up the slack between the tube and the hole (12mm OD, 12.15mm hole) so this was 0.15mm on diameter, then you expand the tube by 5% of its wall thickness. So given a wall thickness of 0.7mm this meant expanding the tubes a further 0.07mm. So the resulting ID of the tube was 10.82mm.  We did this by test expanding one tube in stages and checking the diameter, and when it was right we measured the free-length of the mandrel still showing. This proved to be about 20mm, so we then made a 20mm collar for the mandrel which stopped the expanding at the right point.

This worked very well, the load on the drill driving the expander could be heard to change as the expander reached the collar, thus making it simple to know when each tube was tight, without a lot of squinting down the drums. It’s really a two man job, as one needs to hold the tube so that the required 3mm+ of the tube is extended into the drum, while the other drives the expander and “contraption”. A video of the process in action:

We actually fitted about three-quarters of the tubes that we could reach from one end and then turned the boiler around and did the tubes at the other end. It might have been better to work from both ends, as by the time we had finished the mud drums had been pushed out sideways in the casing by about 1mm – no big issue, but information we will use when expanding the other boilers.

One other issue was that the “contraption” and expander proved too long to be assembled inside the smaller mud drums. We solved this by shortening the “chuck” on the “contraption” and by shortening and regrinding the square drive on the expander.

Here are some pictures showing the process. About 14hrs work to expand the tubes for one boiler.


The next part of the job was to turn up the ends for the drum, and also the economiser and superheater ends.

I decided to start with the economiser and superheater ends. Having first drilled another 12.1mm hole in the centre of the blanks for the inner headers (to take one of the super heater tubes), this was initially a simple turning job with the blanks held on a mandrel pre-turned in a 3-jaw chuck – this lead to Interruption 1, the clutch on the Harrison started slipping when confronted with the loads of highspeed turning with the insert tooling I was using. So an hour of so cleaning and refilling the parts washer followed by getting all the oil off  the clutch plates solved the problem.

Then we needed to cut 70mm diameter 12mm deep recesses in the centre of each part – I started by boring one of these, but it was slow work, so I decided to change over to milling these – here two more interruptions intervened in progress. Firstly the Mill started a rather worrying clicking noise from the gearbox – trying to ignore it didn’t seem to work, so I concluded I needed to strip it and investigate, as any major failure would be catastrophic as there very few parts available if anything broke.

Rambaudi Mill Gearbox Servicing

Rambaudi Mill Gearbox Servicing

This was simple to do, but hard on the nerves  – more parts washing, lots of inspecting and new grease restored the machine to full (quiet) operation. I think small bits of dirt and swarf had migrated into the gears and spindle bearings, and the cleaning sorted things.


Lastly, I spent a while trying to resharpen a bunch of broken and worn milling cutters with pretty limited success – so I bit-the-bullet and purchased some new cutters, including a 16mm “ripping” cutter designed to rough out at higher speed – a good choice.

Recess Milling and PCD drilling

I decided to mount the header blanks on a rotary table and then mill the 70mm recesses with the ripping cutter, then remount the blanks on the table with a stud through the centre hole, located between a set of stops to drill the 12 M8 tapping holes using the DRO’s PCD program – nice and simple and accurate. More pics.