Water Tank Pipe Work Magneto

Part two of three: Restoring a 2 HP IHC Nonpareil


Content Tools

The lack of water tank and cooling pipe work was the next challenge, and part of the problem was the size of the tank. I had some dimensions on an old IHC leaflet, but these were for the Famous and appeared to be some 6 inches taller than the Nonpareil. However, I measured the tank on a friend's 4 HP Nonpareil and it appeared the same as the equivalent Famous. While the old pictures of the 2 HP Nonpareil appeared to show a smaller tank, I decided to stick with the dimensions of the Famous tank.

Sourcing sheet metal and pipe was not a problem, but to be correct in the restoration it would be necessary to use American pipe threads. Fortunately, my son had to go on a business trip to Houston and he travelled with a shopping list of pipe fittings.

I purchased a sheet of 1.5-millimeter (approximately 16-gauge) steel to make the side of the tank and the ring for the screen, plus a smaller piece to make the base.

My rolling bars were only suitable for smaller work so the tank would have to be fashioned by hand. I decided to make the base of the tank first so I would have a template to gauge progress in rolling the sides. The diameter of the tank was 16 inches including the rolled edges of 1/2-inch, so two circles were scribed on the sheet metal of 16 inches and 16-1/2 inches diameter to allow for an overlap seam with the side. After folding the seam, the diameter of the side walls was 15-1/2 inches. I used a nibbler to cut out the base and tidied up the edge with a file.

To form the lip on the base, I had to bend the metal over a former with a small hammer and ground a piece of scrap steel to the correct radius. And to prevent making marks in the metal I ground the face of the hammer smooth, and then polished it with some 200-grit wet-and-dry sandpaper.

I clamped the former to both the sheet and workbench. Then, with a light, continuous tapping action while moving around the base, I gradually formed the lip. I only tapped until there was a slight movement, and then moved the former around the metal. To avoid marks and distortions in the metal, it was a case of tapping lightly and often, working slowly around the base.

Once I had been around the base for the first time and created the first impression in the metal, I clamped the base between two formers to fold the lip - again still working by tapping often and bending slow, moving the formers and continually working around the base until a perfect lip had been formed.

The next step involved cutting the sheet for the tank sides. For a 15-1/2-inch diameter, the sheet needed to be 48.71 inches long (15-1/2" × 22 ÷ 7) plus a joining lip of 1/2-inch (let's say 49-1/2 inches). The height of the tank was 12 inches to which I added 1/4-inch for the jointing seam at the base and 3/8-inch for the rolled edge. I decided I would not be able to form a raised profile in the middle of the tank so no allowance was made for this.

Again I used the nibbler to cut out the sheet and tidied up the edges using a file. I scribed the allowances for the seam and rolled edge on the sheet.

I formed the bottom shoulder for the seam joint with the base first by clamping the sheet between two lengths of steel and then tapping until a 90-degree lip was formed, again by gradual bending.

To form the rolled wire edge, I again used a former. I formed the rolled edge around 3/16-inch diameter galvanized wire and brazed a piece of 3/16-inch diameter steel onto a piece of 1/8-inch sheet metal to use as the former.

Using the same process as the base lip, I clamped the former to the top of the sheet and rolled the edge by light, continuous tapping. I continued forming until the roll was three-quarters formed and the wire inserted. The wire extended 1/2-inch beyond the roll at one end, and at the other a separate 1-inch piece was inserted to act as the final former. It was then removed when completed to leave a hole for the joint. To steady the metal and keep a straight edge, I used a block of wood to support the top edge and tapped in the direction of the wooden block.

I tapped the top edge again and again until the roll was formed, then used a ground tool profiled to the diameter of the roll to complete the edge.

Once I formed the rolled top edge, I created a step for the seam to join the ends of the side piece. I formed this by hammering the last 1/2-inch of the seam over a 1/16-inch-thick piece of steel. I cut off the end of the base lip where the seam would overlap, as well as 1/2-inch of the rolled top edge where the roll had been formed over the 1 inch length of wire.

To roll the side, I set up a length of 3-1/2-inch outer diameter round steel in my portable workbench, supported at the extremity by a length of wood with a V-notch at the top. Then I placed the side piece lengthwise over this bar and exerted light, even pressure on each side with my gloved hands. I slid it around the bar, working from end to end and back. I kept moving my hands to keep them close to the steel bar to prevent any sharp bends or kinks forming in the sheet. Gradually, it started to curve and adopt the required shape. I took care throughout to exert an even, moderate pressure with my hands to prevent ridges forming in the sheet.

The forming of the seamed ends of the side piece required more work than the center section. I accomplished this with some light tapping using a large rubber mallet.

Once the side was formed, the joining seam needed brazing. After cleaning and fluxing, I placed the side in the base and held the top of the side square and in the correct position with a small clamp. I used a long-reach clamp to hold a section of the seam tightly together and brazed the joint up to the clamp, which was then moved to another section. I decided to use a medium temperature solder to braze the seam at the side and low temperature soft solder on the lapped joint at the bottom, so applying the latter would not melt the seam.

Once I had soldered the seam, my attention turned to the fitting of the drain-off point. I made two discs of 1/8-inch-thick steel, 3-1/4-inch diameter and squared off a little at the bottom to place the drain hole low in the tank.

I drilled a 0.850-inch hole through both discs for the pipe and four small 1/8-inch holes for the joining rivets. I then heated these plates and bent them to match the profile of the tank. I brazed a section of threaded pipe to the inside plate after taking careful measurements to ensure it was long enough for the T-joint to be under the water pump when the tank was fitted on the skid. I then drilled a 0.850-inch hole in the side of the tank. I cleaned and fluxed all pieces, then riveted them into position. Finally, I applied solder to complete a waterproof joint.

Next, I needed to form the lap joint at the base of the tank. I cleaned and fluxed the base and side pieces before tapping the lip on the base to fold it over the lip on the tank side, until there was a tight joint. To finish off, I heated it and applied soft solder to ensure a complete seal.

My tank has plain sides, as I could not easily form the raised profile in the middle. To finish the tank, I found a 50-inch length of thin-wall steel pipe, and using a slitting saw on the mill, I went down its length cutting off a thin section, which I then soft soldered to the side of the tank.

Because I had used soft solder in the construction of the tank, it was recommended that I not galvanize the tank in case the solder melted during the galvanizing process. I therefore applied several coast of a high zinc content primer, lightly rubbed it down then sprayed on a silver paint.

Mesh screen

I made the ring for the mesh screen to a diameter of 15-1/2-inches using a strip of 1.5-millimeter steel, 49-1/2 inches long and 2 inches wide. The wired edge top was rolled in the same way as the edge on the tank. The ring itself was rolled (before forming the inward facing lip as I felt it would be difficult to form later), as its narrow section might make it easy for ridges to form. I rounded one edge of a former with a grinder and again used the tapping process to form the 1/2-inch inside lip. In the same manner as the tank base, an overlap seam of 1/2-inch was recessed with a similar amount of the rolled edge and bottom lip being removed. I checked the ring for a sliding fit in the tank and brazed the seam.

The main problem with the mesh screen was finding a source of mesh. Nowadays, everyone appears to use pressed metal and numerous trips to local sheet metal businesses were unsuccessful. Eventually, I managed to track down some mesh although it was made from a thicker gauge wire than I really wanted, and I also had to buy an end of roll which was over 5 feet long.

I had to work out how to make a template for the cone. After studying my reference books, I managed to understand some of the principals involved in drawing the cone and developing it.

The vertical height of the cone was 10-1/2 inches, and to fit inside the ring I had made, the required diameter was 14-1/2 inches, allowing some clearance. Some quick trigonometry resulted in the missing dimensions being calculated.

Known measurements (seen above) were: BC = height of cone = 10-1/2 inches; CD = half width of cone to inside of the lip = 7-1/4 inches; BE = half width of top hole to outside of screen = 1 inch.

In order to calculate the extension AB to a complete triangle, I used the fact that the dimensional ratios were the same. Therefore: AB = BC ÷ CD × BE = 10.5" ÷ 7.25" × 1" = 1.45"

To calculate AD and get the width of mesh required: AD2 = (AB + BC)2 + CD2; AD2 = (1.45 + 10.5)2 + 7.252; AD2 = 195.4; AD = 14.

For the diameter of the cone, the circumference was 14.5" × 22 ÷ 7 = 45.6"

To step out and develop the pattern for the cone, scribe an arc radius 14 inches (AD).

Then step out on this Arc the required circumference, 45.6 inches, by scribing it in 30-degree sectors (1/12 of the circumference), 3.8 inches to end up with the extension seen at the center of this page.

The template needed truncating for the hole in the top. To make allowance for rolling over at the edge, I drew an inner circle of 1/2-inch radius so that after folding over, the inner diameter was 1-1/2-inch and the outer edge diameter was 2 inches.

To join the sides of the cone, I allowed a 5/8-inch seam for an overlap, which is shown with a dashed line. To solder the mesh to the base ring, I made a further allowance of 5/8-inch, which again is shown as a dashed line. This was more than needed, but I could trim to size when fitted.

I then made a template to these measurements out of thin cardboard and formed it into a cone to check my calculations - it fit perfectly. The next step was to clamp the template on the mesh and cut around it.

Once again, I rolled the mesh to shape over the mandrel I used to form the side of the tank.

To join the sides of the mesh I decided to wire the seam rather than try and solder the mesh. I used some short lengths of galvanized wire, which I looped through, then twisted tight inside the cone and flattened with a hammer.

I formed the lip to join the cone to the ring at the top of the tank by bending the mesh with pliers, and similarly the edge in the hole at the top of the cone was turned over with pliers.

The metal disc to spread the water to the cone is located 9 inches high at which point the diameter was 4 inches. I put this in place inside the cone, rested it on temporary wire stays threaded across the cone, soft soldered and removed the stays.

I placed the cone inside the ring and after a little adjustment and checking to ensure it was square, I soldered it in place and tidied the edges of the wire mesh with a grinding wheel and file. I then painted the screen to match the tank, taking care not to apply too much and clog the mesh.

Pipe work

To complete the pipe work, I needed some 135-degree elbows and a T-piece, which I had been unable to buy.

I made the elbows by taking a 3-inch length of some 1-1/8-inch outer diameter steel, boring it out to 0.715-inch and tapping it for 1/2-inch by 14 thread. A 45-degree V was marked out in the center, which I cut out with a hacksaw and tidied with a file. I brazed the two pieces together and cleaned them with a grinder.

Similarly, I made the T-piece by cutting the pipe thread in a 2-inch length and a 1-inch length of 1-1/8-inch steel. I bored the center section of the longer piece to a diameter of 1-1/8-inch and a depth of 9/16-inch so the short piece could be brazed in position to create the T. Finally, I cleaned it by grinding.


The main problem with the magneto was the missing lead-out tower. The magneto, a Bosch DU1, was made in 1911 according to the serial number. A quick search on the Internet failed to reveal any photos, so I resorted to calling magneto specialists. This trail eventually led to a magneto collector, and in addition to providing a photo, he generously lent me a tower to copy.

While waiting for the tower to arrive in the mail, I stripped the magneto for cleaning. I needed penetrant on some of the screws and the careful use of an impact driver to remove two of them.

The spring for the contact breaker was broken, so I bought a piece of spring steel and made new holes with a small grinding point in my Dremel. I understand there were three styles of contact breaker spring in the DU1: a single spring, a double spring and a single spring with short lengths of support spring at both ends. The latter appeared to be the style fitted to my magneto. I removed the contact breaker points, cleaned the square with a small slip stone and set the contact gap at 0.025-inch.

I had to make the lead-out tower without the luxury of stripping the one I borrowed, but at least from the exterior it was identical to the original.

I turned the body of the new tower out of a piece of Delrin plastic - the original appearing to be made out of Bakelite. I was able to measure the tapers at 13 degrees. The most difficult part was the brass locking ring, which had an internal taper of 13 degrees and a 14-by-l-millimeter pitch thread, which was screw-cut on the lathe. To finish the lock ring, I slit one side with a 0.025-inch saw.

The remaining nuts, internal brass tube and the out lead lock screw were simple turning tasks.

I tested the magneto by flicking it by hand and measuring the distance that the spark would jump in open atmosphere. To work effectively under compression the spark needed to jump a gap of 5 millimeters. Unfortunately, the maximum was only 3 millimeters, so I took the magneto to a friend who re-magnetised it, resulting in a big improvement in the spark.

Read part three in the next issue, covering the remainder of the restoration.

Contact Peter Rooke at: Hardigate House, Hardigate Road, Cropwell Butler, Notts, England, NG12 3AH; peter@enginepeter.co.uk • www.enginepeter.co.uk