After doing some R&R on one of my Monotype composition moulds I found that when installed on the caster, the type carrier could not reach into the mould without rubbing on the face of the left type block. One of the assembly steps for this style of mould does not have any reference points to replace a part in the exact location required.

I set out to measure all my moulds to determine what the correct distance is between the faces of the left and right type blocks and the read edge of the mould’s base plate. For the 3E style moulds I also measured from the faces of the type blocks to the rear of the squaring plate to see if that was consistent from one mould to another. If it were, a simple spacer could be used to properly install the squaring block in the right position.

The first measurement is an awkward one to make because the two faces are not back-to-back. I eventually found that a metal angle clamped to a flat surface could be used to do this. The back edge of the mould base would be held against the edge of the angle, then measurements taken from the upright leg of the angle to the face of the type block. I used a piece of 2″ aluminum angle bar, so my measurements were 2″ greater than the actual distance I wanted to measure.

Measurement on an English composition mould using a 4″ micrometer.

The distance I found was pretty consistent across the six moulds I measured: the average was 1.940″ and the maximum was 1.946″ whereas I found the distance to be around 1.96″ when I measured the refurbished mould.

I loosened the screws on the underside of this mould (5 holding the squaring plate and 5 more holding the type blocks) so the inner assembly could move around. The type blocks were still held to the squaring plate by 5 other screws in the ends and back. I set two of the screws to provide just a bit of friction, placed the mould in my measuring setup and tapped the parts until they showed an acceptable distance. Then I tightened all the screws again and made a final measurement of 1.938 inches, a bit low, but within the range of my other moulds.

I reinstalled the mould on my caster, and now the type carrier moved freely. I used a feeler gauge to measure how much clearance the type carrier had, and the result was 0.013″ which would imply a maximum distance of 1.951 inches would be acceptable.

The other distance of interest, from the face of the type blocks to the rear of the bosses on the squaring plate, at least initially seems disappointingly inconsistent and so not useful as a reference for mould reassembly. One problem though is that I was making these measurements on assembled moulds and this makes accurate measurement difficult. As well, the rear surface behind the right type block may not be precision machined but the surface behind the left type block is precisely machined, so I will have to take a closer look at this idea next time I take a mould apart.

I also did a check for water flow through the mould and that is still insufficient, as it is only a drop or two per second. I will have to take the mould off again and try blasting fluid through the cooling passages to try to clear out any blockages.

With the crossblock and blade reassembled, the rest of the mould goes together quickly. All of the remaining parts are built onto the base plate.

The first parts to go on are the cam which operates the jet ejector as the crossblock moves sideways, and the squaring plate on which the parts forming the mould cavity are built. The cam is positioned by two dowels which I did not remove from the base, and secured by two screws inserted from the underside of the base plate. The squaring plate is held by five screws from the underside of the base plate and does not have anything to provide positive positioning. I set it in about the middle of the play provided by its screw holes, but as I later found out its position is more critical than I had thought.

The next parts to install are the left and right type blocks. The larger right block (on the left in the photo) also has two plug screws for oil passages and two felt pads to meter out the oil, which I installed into the block beforehand.

Each of these blocks has screws securing it against the squaring plate in three dimensions: one at the outer end pushing the two blocks towards each other, one or two through the boss of the squaring block to pull the type block back, and one or two from below through the base plate to hold the type block down.

As with the screws in the crossblock, these must be tightened gradually in sequence to ensure the type block is seated properly.

If all goes well, at this point, the mould blade can be inserted by sliding it in from the back. In my case, one of the felt pads (the one that oils the side of the blade) was not fully into its hole, so the blade actually had to shear off some of the fibres from the felt. As a result the initial insertion was tight, but it quickly loosened up as the sheared-off fibres worked out of the sliding area.

Three small parts that retain and guide the blade are installed next. The are the right mould blade shoe which holds the main blade down and limits how far it can open, the mould blade stop which limits how far the mould blade can close, and the mould blade carrier guide block which prevents the back end of the blade from moving sideways. The back surface of the mould blade stop has been precision ground at the factory so the mould blade, when closed, is exactly flush with the front faces of the two type blocks. At this point I also dropped the upper blade operating lever into place on its pin on the blade carrier.

The two covers are installed next. The left cover (the one with four screws) is also called the left mould blade shoe because it completes the guide that the blade slides in. It both holds the upper blade down and also prevents it from shifting to the left. The right cover (the one with two screws; the apparent third screw is actually the one holding the right mould blade shoe in place) is also called the mould blade top guide as it prevents the blade from shifting to the right.

The last part to assemble and install is the apron, or front abutment. It includes the nameplate and the crossblock shoe along with its adjusting screws and locknuts. It is held to the base plate by three large screws from below.

I slid the crossblock into place, and as with the mold blade, the felt pad (the one on the crossblock) was not tucked back into its hole neatly. However, in this case I could reach the pad to tuck it into its cavity as the block slid in. I slid the block to the position in the photo and tightened both shoe adjusting screws tight and left them that way for a while to seat the felt.

Later (after taking the photo) I backed off the adjusting screws a tiny bit (less that a tenth of a turn) and tightened their locknuts.

Everything moved freely and was now well oiled so I put the mould in place on the caster (using only the two clamps, not the blots from below). On turning the caster handwheel by hand I found to my dismay that the caster would not operate smoothly. The type carrier appears to be rubbing on the front face of the left type block.

Some thought on which parts mount where leads to the conclusion that this interference is caused by incorrect placement of the squaring block. Evidently on factory assembly these are installed using some sort of fixture to hold the squaring block at the right location on the base plate.

To correct this I don’t have to disassemble everything; I should just be able to loosen the eight screws that hold the squaring block and type blocks to the base plate. Before doing this, though, I will have to measure some other moulds to determine how far the type block fronts should be from the rear edge of the base plate. If they are too far the type carrier will rub the type block as I observed. If they are too close the type carrier will fail to pick up very narrow pieces of type as they eject from the mould cavity.

The next post should include the results of these measurements and the correction of this mould’s squaring block position.

The next major part of the 10-point mould to reassemble is the crossblock. This block forms one side of the casting cavity as well as the bottom of the type and the jet. After casting it slides sideways, cutting off the jet in the process, allowing the type to be ejected from the mould. A cam in the base of the mould moves an ejector to push out the jet, where it drops back into the metal pot to be remelted.

The crossblock consists of four major pieces, a felt pad, and ten fasteners. The main body is in the center of this photo, and the adjusting screw is already installed in the end of it. The surface shown in the photo is the underside of the body, and the surface facing the top of the photo is the one that forms one side of the mould cavity.

The two parts at the top of the photo are the Gate Blocks (“gate” is another term for “jet”, the scrap part of the casting where the molten metal is injected), which attach to the body with a fine adjustment to set the width of the V-groove for the jet ejector. The Left Gate Block (on the right) is positioned by the dowel, essentially a threaded post, which fits into the extra hole visible on one side of the main body. The adjustment screw is then used to position the Right Gate Block so the jet ejector fits properly. Each gate block is held to the body by four screws. The Gate Blocks form the foot of the type and two sides of the jet.

The felt pad fits into the Left Gate Block and serves to spread oil along the lower sliding surfaces of the crossblock. This pad does not have its own oil supply, but just smears around whatever oil reaches this area via gravity.

This shows the crossblock right way up with the Left Gate Block installed including the felt pad. The dowel that positions this block is a bit strange: It is threaded about halfway along its length, and has a slot head with the top edges tapered parallel to the slot. Depending on where it stops rotating when it bottoms out in its hole, it may or may not be a snug fit in the matching groove in the gate block because of its tapered sides. On mine, there was perhaps a hundredth of an inch of play in the gate block position, not enough to worry about. Gross mispositioning of this block would mean the jet opening would not be centered over the pump nozzle, and the metal would inject into the mould cavity with excess turbulence, possibly making more porous type.

The underside of the assembled crossblock shows the jet ejector in position. The triangular end of the ejector (towards the bottom of the photo) forms the third side of the jet.

The vent groove added by Hartzell can be seen running towards the right from the jet ejector.

The Right Gate Block has been adjusted close enough to the left one so that the ejector has no sideways play but not so close that the ejector can’t seat fully into its slot. This adjustment is easier with no oil, or only a thin film of very light oil, on the ejector, as a heavy film of oil will make it act thicker than it really is.

Both gate blocks must be tightened down carefully to ensure they are properly seated against the crossblock body. All four screws in each gate block must be tightened in sequence a bit at a time, starting from when they first start to snug down.

The first parts of my 10-point mould to be cleaned and reassembled are those that form the mould blade. The flat end of this blade forms one side of the cast type, and the caster adjusts the width of the type by the position of this blade. A separate top section of the blade can remain closed to cast a low space rather than the character formed on the mat positioned over the mould. The blade also ejects the finished type from the mould cavity.

The main parts of the blade are the Upper Mould Blade, the Lower Mould Blade, and the Mould Blade Carrier. Below them are a couple of pivot pins, a spring, and the Mould Blade Carrier Latch. On the right is the lever that selects whether the upper blade opens for normal casting or remains closed for a low space.

The caster moves the lower mould blade using the square hole near its right end. The carrier latch normally makes the blade carrier move in unison with the lower blade, and the upper blade always moves with the carrier. Thus the upper and lower blades normally move together.

For low space casting, the lever releases the carrier latch and also pushes the carrier to the closed position, so the lower blade still moves as selected by the caster but the upper blade remains closed.

I suspect that my mould might be made from parts from several other moulds. In particular, the upper blade is wider (by about 0.001″) than the lower blade and also sticks out a bit (I didn’t measure this but it was enough to snag my fingernail). This would imply to me that the upper blade is a newer replacement. The two sides of the mould, between which the blade slides, also show a wear pattern consistent with the upper blade being wider. I’ll have to keep this in mind if there are problems with the cast type. This could also be compensation for differences in thermal expansion when the mould gets hot in use, although the width difference is close to 1% which is quite a bit more than I would expect for any expansion.

I have reassembled the blade, and it is shown on the left with the upper blade projecting beyond the lower blade, as if a very wide low space were to be cast. The lever cannot be installed until the blade is properly set into the mould.

I think this is the last thing to do before trying to cast the ribbon on my Monotype: Cleaning the old 10 point composition mould so it works at top performance.

This particular mould is a 10 point serial number 3E1046H. It is essentially the composition mould design used by Lanston Monotype, but as indicated by the nameplate and the ‘H’ on the serial number it has been reconditioned by Hartzell Machine Works, who for many years repaired Monotype moulds. One visible change Hartzell made was to add an air venting passage to the jet so that at least some of the air could escape the mould cavity rather than having to be compressed into trapped bubbles. This means this mould should eject the tomahawk-shaped jets characteristic of this modification.

The exploded diagram for this mould is shown on page 38 of Plate Book, Monotype Typesetting Machine, the Composition Type-caster. The only difference I could find from what is shown in the manual is that this mould does not have the left mould blade guide 18MC3E1 nor its screw, nor even the hole in the top cover to clear the screw head (so it is not just a case of a missing part). Furthermore, the mould blade latch spring on this mould is a single piece with the end loops formed by bending the spring wire, while the parts manual shows a spring with separate end eyes that the last coil of the spring threads into.

On disassembly I found that all of the cooling water passages were plugged with rust. I would have liked to remove the four plug screws so I would have full access to all the passages to clean them out but these plugs refused to budge, even using an impact screwdriver on them. Although I could drill them out, clean out the scraps, and make replacements, I will first try other means to clean these passages. Perversely, the two plugs in oil passages, which seem perfectly clear, came out easily.

It is not necessary (nor even advisable) to disassemble the mould to this extent for normal cleaning, but I wanted it to be as factory-fresh as it could be. Besides, if I destroy it, Don Black has several other moulds of the same size that I could use. YOLO, as they say.

A 10 point Lanston 3E composition mould, freshly disassembled, dirt and all

As annotated in the photo, a few parts were left to be disassembled when the photo was taken. I have now removed items C and D and two of the plug screws B (the ones in oil ports). The remaining parts are too hard to remove or might be too hard to reinstall.

I was planning on casting the Monotype ribbon I brought home from Monotype University 8 last August, but I just noticed today that the ribbon is punched for 9½-set and I don’t have an S5 wedge for this size. All I have is the wedge for 10-set (and fractions larger).

I can use the 10-set wedge with this ribbon, but the lines will come out with the justification all wrong because the variable space width is specified in the ribbon punches based on the set size being 9½ points. So I will have to baby-sit the machine, with the galley set overly wide (to accommodate the lines that will be about 5% longer than planned), the wrong-line-length trip disabled, and dropping in leads between each line.

There are only about 8 lines of actual text (not counting warm-up lines and such) so this would not be too burdensome.

I have also figured out how I can manually read the justification punches in the ribbon, convert the value to inches, scale it by 20/19 (10/9½), and convert back to punch codes. Then I can paste over the existing justification punches and manually punch new ones, thus giving a ribbon made for 10-set and a line 20/19 times longer than the original line length. Once again this is only feasible because there are so few lines. I had originally thought that to correct the justification punches I would have to count all the unit widths of all the characters in each line. The latter method might be a bit more accurate since it would not suffer from double rounding during the calculations, but that isn’t really worth the trouble.

So for now I’m continuing to plod on, with the next job being to disassemble and clean my 10-point composition mould.

After deciding to replace some of the valves that control the cooling water flow on my caster, I made a visit to a few hardware stores. I came home with a ball valve and a needle valve but as they did not have the correct pipe connections I also bought two adapters.

The valves I purchased. Main row (L-R) are the original supply flare fitting, the ball valve, a ⅜″ compression nut, the body of the needle value, and another ⅜″ compression nut. Above is the valve needle and cap.

The ball valve has ¼″ FPT (Female Pipe Thread) connections, which are fine. The flare fitting that connects to the supply line will attach directly to this valve.

The needle valve was unfortunately only available with ⅜″ compression fittings. In order to fit this to the ball valve and to the line that carries the water to the table and mould, I also bought two adapters from ¼″ MPT (Male Pipe Thread) to ⅜″ compression. I would also have to dig through my supplies for a couple of short lengths of ⅜″ soft copper tubing.

It turned out there were three downsides to this approach. One is that the combination of all these adapters and connections would make the valve unit too long to fit in the original valves’ location. Another is that all these connections are potential points of leakage. Finally, closer measurement revealed that the fitting to on the line to the table does not attach onto ¼″ MPT as the thread pitch is wrong.

The compression ends on the needle valve could not be directly modified to another thread type because there was not enough metal thickness to cut different threads. What I did instead was to modify one of the compression adapters I had bought to slip into one of the compression sockets and to make another adapter from scratch for the exit fitting that also slipped into the compression fitting, and solder both these directly to the needle valve.

Cutting these to fit into the compression fitting was fairly easy, and the solder would fill in plenty of sloppy measurement. The outlet fitting turned out to need 19 threads per inch on an odd diameter. To do this on my lathe I had to replace a 24-tooth gear in the carriage drive with a 60-tooth one, and select 48 threads per inch on the gear box (which had no 19TPI setting). This would make the lathe cut at 19.2 threads per inch (48×24/60) which is close enough. I did test fittings of the matching nut to determine when the threads had the correct diameter.

The needle valve and custom adapters, ready to solder in. The one on the left is the oddball 19TPI straight thread, and on the right is the 1/4″ MPT.

After soldering the adapters onto the needle value, I loosely assembled the two valves and verified that the assembly is about the same length as the original valve pair.

I assembled these using Teflon tape on the pipe thread joints and thread sealing compound where the cone ends screw onto the ends of the copper tube. With everything installed on the caster, the water valves look like this:

The repaired outlet valve is at the upper left and the two new inlet controls on the right. There is no longer any chance of confusing the flow control and the water shutoff. A trial run indicates excellent flow rate and good control using the needle valve.

The next thing to check on my caster was that the cooling system could deliver sufficient water to cool the mould for composition casting. I have a closed-loop system which uses a small fountain pump to force the water up into the cooling system, and cools the drain water in an open coil of copper tubing before emptying it back into the tank.

The fountain pump does not supply much pressure, so the passages in the system must be reasonably clear.

With no mould installed, I turned on the pump and opened all the valves. No water flowed out of the drain pipe. I disconnected the drain fitting on the right side of the caster table near the mould, and no water was coming out there either. Actually, after a few minutes, and small trickle started to appear but nowhere near enough to provide adequate cooling.

I disconnected the inlet to the table cooling port to check if the passages in the table were clogged, but the flow there was just as bad. Finally I disconnected the inlet to the two inflow valves, and got a gusher there. Obviously the valves were barely opening despite having their knobs screwed all the way out.

After shutting off the pump I removed these two valves and their associated piping, disassembled everything and cleaned it. I also removed the cover from the outflow valve whose purpose is to generate controlled back pressure to force cooling water to flow through the mould itself rather than just the caster table.

The three valves and associated piping, in pieces, on the right, and some spares on the left.

The valves are of a type called a weir diaphragm valve. The two valve bodies are at the bottom center of the picture. To their right is the water line which runs from the valves to the table inlet fitting. Above these are the three sets of (top to bottom) knobs with taper pin, stem, lock nut, top cover with screws, piston, and diaphragm. Each piston is split with an internal hole to grab onto the end of the stem and onto a button on the back of the diaphragm. Once the valve is assembled, this piston is contained in the cap, preventing the two halves from separating.

The upper half of this diagram, on the Spirax Sarco web site, shows how these valves work. When closed the diaphragm is pressed against the weir to cut off flow. When open, the valve stem pulls the diaphragm up to provide high flow.

One advantage of this style of valve is that none of the moving parts are exposed to the controlled fluid (cooling water, in this case). They also provide fairly good modulation of the flow rate. Unfortunately on my caster, the diaphragms, which might be 50 to 90 years old, are very hard and set in their closed shape since the casters were generally mothballed with these valves closed. Furthermore, on two of them the button that the stem pulls on to open the valve is breaking off so the valve cannot supply much flow unless the water supply pressure is sufficient to push the diaphragm open.

Torn button on back of diaphragm

In my spares I have two complete valves and one valve cap assembly. On examining the two spare valves, it is clear that the diaphragm button is broken on one of them as well. I feel it is unlikely that I could find any replacements to buy, and although I could make most of a diaphragm from sheet rubber, it is not clear how to make and securely attach the button.

Given the shortage of good diaphragms, I decided to save them for the outflow valve, and replace the inflow valves with something more modern.

The caster is fitted with two identical valves in series, and I have never seen a written explanation of this; I guess this is something one would have learned at Monotype School back in the day. My assumption is that one is supposed to set one of the valves to control the desired water flow, and turn the other valve either shut or wide open. In this way the water can be turned on and off without losing the flow setting. I have been trying to use them this way myself but I could never remember which valve I was using for which purpose so I was always losing the flow setting anyway.

For the replacement valves I will use a ball valve for the on/off control, and a needle valve for the flow control. Time to wander through the plumbing department at the local hardware store…

I made myself a Winding Spool for my Monotype caster, to take up the ribbon as the machine reads it. The internal mechanism is similar but not identical to the standard part. The main difference is that the part know as the Driving Disc 21G7 is closer to the rear of the machine (the open end of the spool) and also does the duty of the Shaft Spring Abutment 21G11, with the spring being held between the Driving Disc and the front end cap (Flange Bush a21G2). The Driving Disc is held in place on the Shaft by a setscrew, and the Shaft Driving Disc Pin 21G8 is threaded directly into the Driving Disc and held by thread locking compound rather than a locknut.

The parts for my home-made Winding Spool, showing the names and numbers of the closest corresponding standard parts.

Most of the parts were scavenged. The Tube was a piece from the photosensitive drum of an old laser toner cartridge, the Flange Spring was a strip cut off some scrap aluminum siding, and the Flange came from some random piece of black anodized aluminum scrap. The spring was hand wound on my lathe. Everything is held together with #2-56 flat head screws in countersunk holes. Getting the spool to work properly required a bit of fine tuning on the lathe adjusting the thickness of the Rear Plug and Flange Bush so the Shaft had the correct range of longitudinal motion.

The Winding Spool all assembled

The Winding Spool installed on the caster with drive engaged

When the spool is off the caster both ends of the Shaft are flush with the ends of the spool. Once on the caster, the big end of the shaft is pushed by the Winding Spool Spring Box Plunger group X25G, causing the small end to project and engage the hole in the center of the drive disk. If the Button 25G2 is turned so its bumps are not in the notches in its housing, this is all that happens, and the spool can rotate freely. If the button is turned so its bumps are in the housing notches, the Shaft is pushed further, causing the Disc Pin to also project from the end of the spool and engage one of the drive holes on the drive disk. In this manner the Winding Spool Driving Ratchet X23G can turn the spool to apply enough tension to the ribbon to cause it to wind onto the spool.

With this spool on the caster, my patched-up ribbon runs flawlessly from start to finish, although I have yet to try this with the air on to see how much leakage there is over the ribbon patches.

I think I’m finally coming to the home stretch to get this sucker working. All that seems to be left is to clean the mould so it operates smoothly and verify that my closed-loop cooling system still works, and then I’m ready to try casting this ribbon.