Editor's note: Last issue, we began a three-part feature on Peter Rooke's latest restoration project. We continue below with Part 2.
Like most fittings on this rusted engine, the fuel mixer needed considerable attention.
It was necessary to give the body of the mixer a good clean with a wire brush and all brass fittings were also removed. The fuel reservoir was cleaned using kerosene and a small toothbrush. The screw to hold the choke cover broke when I tried to remove it, so I drilled the remainder then use a tap to re-cut the thread.
The lever on the tap to control/empty the fuel reservoir had long since rusted away, making it difficult to remove the tapered shaft. While it is possible to heat the shaft to pop it out, I do not like using heat on cast iron unless I have no other option. An alternative is to drill a hole in the base of the casting, using a punch to press the taper out of its seat, then tap the hole and fit a screwed plug to seal it. However, I had to make a new lever for the tap and rather than peen the brass of the body over the new lever to hold it in place, I decided to drill and tap the body for a large head screw. The threaded hole could then be used for a temporary bolt to help pull the shaft out.
Once the body of the tap had been extracted, its seating area was cleaned. The top of the tap was faced off, the remains of the old lever removed, and a new lever was drilled and filed to shape. A brass screw was turned to hold the lever in place and the head made over sized for a very shallow slot that could be filed when the screw was firmly in place.
The needle valves had to be replaced and the state of the valve seats was unknown. Some 5/16-inch brass was turned, then threaded 5/16-by-18-inch and tapered to match the originals. At the same time that the lathe was set up to turn the taper, a piece of steel was machined to copy the brass stems. Later, the steel was used to lap the valve seats with very fine compound so the new valves would be a good fit.
It was necessary to cut the rusted steel nuts from the brass rod, which could be reused after cleaning the threads. The moving baffle had to be replaced and a new one was fabricated by machining some steel of different diameters then drilling holes before brazing together. To complete this part of the repair a new spring, shaped choke back plate and choke front plate were purchased.
There was no sign of any float valve moving in the water inlet, despite prodding with a piece of wire to clear any dirt, so it was necessary to remove the water pressure valve seat. Unfortunately, it would not budge, despite resorting to the impact driver, so it had to be drilled in order that all the accumulated dirt could be cleaned out. A new alloy valve was made from some aluminium and brass turned to copy the original seat.
To complete the mixer, new caps were made for the fuel reservoirs and pointer for the needle valves.
Like all other components, the fuel pump was seized solid and the spring had virtually rusted away. The ram resisted all efforts to remove it and, as the casting was badly rusted and undoubtedly weakened, care had to be taken. The use of hydraulic pressure through the out pipe to push up the ram, might have caused more problems than it solved.
The top cap on the ram was removed by filing the peening holding it in place, in order to remove it and the remains of the spring. Fortunately, the check valve and the inlet pipe connector were unscrewed without too much effort.
The pump was then rested upright and penetrant applied, and allowed to work for a couple of days before gripping the top of the ram with a pipe wrench, and gently twisting out while holding the body of the pump in a vice.
Once the ram was out, the check ball in the bottom of the pump was removed and stored safely until needed. The bore of the pump was cleaned, and a length of brass, 7/16-inch, was inserted to test the fit. It was slightly sloppy, so an adjustable reamer was used to true up the bore, which ended up over-sized by 0.0001 before a piece of brass was turned to a sliding fit to replace the old ram. In any event, the ram needed replacement in view of the high amount of uneven wear around the packing area. For the packing to seal properly, the ram had to be parallel.
The brass collar to hold the packing in place had seen better days, so while at the lathe a replacement was made from a piece of scrap brass.
All the check valve needed was to be cleaned and dirt removed from the check ball and its seat.
As the body of the pump was covered in rust pit marks, filler was applied and sanded down so that the pump looked as good as new. Before re-assembly, a steel ball 3/8-inch diameter, was fixed to a piece of brass tube using an epoxy glue. Fine lapping compound was applied and the ball used to polish up the check ball seat in the pump. Kerosene and strips of clean rag on thin rods were used to thoroughly clean any remaining lapping compound.
The original cap was reamed to fit the new ram, the top of which was pin punched and peened over to hold it in place.
To pack the pump ram, to prevent leaks, soaped mop string or graphited yarn should be used. While I used some leftover graphited yarn, mop string probably lasts longer as it is not so susceptible to attack by gasoline. While this is a kerosene engine, I intended to run it on gasoline, which is cheaper in the United Kingdom.
The brass collar was fitted on the ram with the cone towards the packing, so it compressed the packing around the ram. The check-ball was inserted in the pump and the remaining parts assembled.
Fuel pipe and fittings
The majority of the fuel pipes were damaged, so new pipe was acquired ready for bending. Most of the original pipe was bent in more than one place, so to save trial and error a piece of wire coat hanger was bent by hand to the required shape for each section of pipe and used as a template.
Before starting to bend the pipe, the new pipe was annealed by heating to a dull red and allowed to cool slowly. It is best to anneal pipe just before bending. If you leave it for several days, it will harden again and require further annealing.
A cheap bending rig is normally adequate for brass pipe. It is best to take your time, frequently stopping to check the amount of bend to both the template and engine, as it can be difficult to correct too much curve. Also allow a little extra at each end so that it can be trimmed off in final fitting.
A couple of brass pipe fittings were missing, one of which had an unusual thread (0.540 inches by 20 tpi.
Without taps for this non-standard thread, some round brass was turned to the shape of the original female end. The internal hole was drilled then bored before screw cutting the internal thread on the lathe. Using the same settings, a piece of hexagon brass was bored and screw cut for the new locking nut.
The next step was to turn brass to size for the male section and cut the NPT thread with a die. The non-threaded ends of both pieces were sawn with a 45-degree mitre and brazed together to form the fitting, which after finish filing should look like the original.
While brass fittings can be obtained some do not look like the original, so to improve the appearance, a file can be used to re-shape them. Filing down the threaded section of an oversized street fitting, drilling and tapping an internal thread, it was possible to make a close match to the original.
There was little left of the original fuel tank, and for the low cost involved, it was easier to order a replacement from Hit & Miss Enterprises in the U.S. rather than make one. Apart from the drain plug, fitting of everything was fairly straight forward after making new steel pipes including a delivery tube.
The website of the late John Hammink (http://www.oldengine.org/members/hammink/web/) provided instructions on how to adapt a pipe fitting so the drain pipe from the tank would line up with the hole in the body of the engine casting. As a first step, a 35/64-inch hole was drilled in the fitting and a piece of pipe threaded 1/4-inch-by-18 inches and shaped to allow fuel to pass. The fitting needed to be trial fitted to the drain pipe and when all looked satisfactory, it was brazed together. I found that after brazing, I needed to grind a little off one side of the fitting to get the necessary clearance.
To repair the take-off pipe, a length of brass pipe 1/4-inch diameter, was soft-soldered to the male end of an elbow, with the pipe being long enough so that it nearly reached the bottom of the tank when the elbow was fully screwed into position.
When fitting the fuel tank, I found it easier to fit the take-off pipes and elbow first, then the special elbow for the drain pipe, wrapping gas proof PTFE tape around the threads for leak proof joints.
The inlet pipe and the final length of the drain pipe were fitted once the tank was bolted in place. The tank and underside of the casting were given a good coating of anti-rust paint before assembly.
Fuel filler pipe
A new fuel filler pipe was made out of 3/4-inch nominal pipe. Again, a piece of wire was used as a template to assess the length and the degree of the required bend. While the pipe was still straight, the threads at each end were started by screw-cutting on the lathe, then finished using dies. An end cap was screwed on one end and the pipe packed with dry sand, which was compacted in place.
The pipe was held vertical in the vice and a section of tube was found so it could be used to slip over the pipe to bend it. The pipe was then heated with a propane torch at the precise point where it was going to be bent, ensuring the outside of the bend was the hottest at the point where it would need to stretch the most. When red hot, the larger tube was slid over the top and used to bend the hot pipe to the same profile as the template. This heating process was repeated a couple of times until the pipe was exactly the right profile. When cool, the end cap was removed and the sand was thoroughly cleaned.
Water drain tap
The water drain tap presented a problem as it did not appear complete, and it took some time to find someone who knew this style of fitting. Luckily, I noticed a photo of a similar fitting on SmokStak, and a message to Bryan Storey elicited a very helpful reply with some detailed photos, not only of the drain fitting but also of the correct style of cart, which will be covered later.
Brian's photos showed the drain screw from which a copy was made, scaling the photos based on the known measurements of my fitting. To complete the tap, a silicon ring was used for the seal rather than leather, which would have been originally used.
The only unanswered question concerned a hole in the bottom of the round part, which I sealed with solder as it had no threads for a screw plug. To date no one has offered a reason for this hole, and if any reader knows why and the correct method to seal it please let me know.
The igniter presented a major challenge and the simplest course would have been to purchase a replacement. Everything was rusted solid and the main part of the arm, to support the shaft, had been broken off at some stage. Despite lashings of penetrant, nothing shifted, and I had to resort to the hacksaw and drill to disassemble the igniter.
Fortunately, the hole for the shaft was not badly damaged and a clean up with a reamer was all that was needed. The remains of the shaft were drilled out from the contact head, which was brazed onto a new shaft of 7/16-inch drill rod. The holes for the locking pin and spring retaining pin were only drilled in this shaft when the other parts were made and trial fitted.
The screw thread at the top of the tapered shaft for the fixed contact broke off as soon as I tried to move it, and the tapered Mica sleeve was also badly damaged. While it is possible to wind a new tapered sleeve using Mica sheet, I decided to take the easy option and bought a parallel one. While the tapered profile was copied from the old post, using a parallel mica tube made it necessary to braze a thin steel disc just under the contact to hold the mica washers and shaft in place. A new bushing was made to the size of the new parallel mica tube, to replace the old bushing with a tapered hole.
A new contact was made from a piece of iron nail, although this would not last as long as a special tungsten point.
The major problem with the igniter was the lack of a support arm, but photos and the remains of the original gave me enough information to fabricate a replacement.
As a first step, the igniter was set up square on the milling table and the broken area milled true. The new section was to be shaped from a piece of 1-1/2-by-1/2-inch steel but there was little margin for error. A thin section of similar profile steel was Super glued in the required position in the igniter and the long edge set square to the axis of the mill. The coordinates of the mill were set to zero at one of the corners of the section of steel before it was broken off, and 2-1/8-inch diameter holes were drilled in the remaining stub of the old arm, the measurements from the corner being noted. The igniter was removed and the steel to be used was set square and again the coordinates of the same corner set to zero. Copying the coordinates of the drilled holes, 2-1/2-inch holes were drilled in the end of this steel, to perfectly match the holes in the igniter.
After a trial assembly to check the fit was correct, 2-1/8-inch steel pins and the mating surfaces were covered in flux, the steel piece was brazed to the igniter.
The tube to support the end of the shaft was turned from a piece of scrap steel, then using a greased length of 7/16-inch rod to keep it square with the main hole in the igniter, it was brazed in position.
Once the support arm was finished, the remaining parts were fashioned, and the igniter was completed with new springs, mica washers and fahnstock clip.
All that was needed to finish was to drill the holes in the moving stem for the pins, and a hole in the end of the new support arm to tension the end spring. While drilling this latter hole, several further holes were drilled so it would be possible to adjust the spring tension at a later date. The igniter was set up so that the points gap after snapping was between 0.035 to 0.070 inches.
The repaired igniter was tested by connecting it to a battery and coil and snapping it by hand. After some fine adjustments to the spring tension, it worked fine.
In part three next issue, Peter puts the finishing touches on this amazing restoration.
Contact Peter Rooke at: Hardigate House, Hardigate Road, Cropwell Butler, Notts, England, NG12 3 AH; firstname.lastname@example.org • www.enginepeter.co.uk