My recent post about setting up to cast 18 point display type got up and down mixed up. When casting 18 point on an English mould with an American display mat holder, the mat must be lowered, not raised as I had stated. There is a mat holder insert kit to accomplish this, but it is the one insert kit that I don’t have.
I have added corrections and an addendum to the original post explaining the oopsie.
On Tuesday I was driving along highway 401 where it crosses the Grand river in south Kitchener when I saw the silhouette of a very large bird flying south of the highway.
I only saw it for a second or two before the view was cut off by the top of the windshield (and besides, I was driving), and saw it mostly in silhouette against dark clouds. Just before it slipped out of sight, though, I saw that it had a white head.
So I’m pretty sure it was a bald eagle, the first time I’ve seen one not in captivity. I was mildly surprised to see one so close to the city, but there have been other sightings elsewhere along the Grand river in the Kitchener area.
Today I used my pycnometer and precision scale to measure the specific gravity of the glycerin which we sell and which I use when I’m trying out recipes for home-made composition rollers.
I started taking a video of this procedure but my camera battery ran out of juice before I was done, and the spare batteries were hiding in some Wii controllers and also pretty much drained. I didn’t want to wash up the equipment and restart another day, and I didn’t want to leave everything out either, so I will have an incomplete video of the process posted later. It turns out that you can only watch someone filling a pycnometer, drying its outside, and weighing it two or three times before it gets tiresome anyway.
The only thing missing from the video is the extra step, when filling the pycnometer with glycerin, of rinsing the overflow off the outside of the bottle before trying to wipe it dry. This ensures there is no film of glycerin on the outside of the bottle adding to its weight.
In any case, the pycnometer held 9.934g of water or 12.505g of glycerin (these are weights, not mass, because there has been no compensation for the buoyancy from the atmosphere), yielding an apparent specific gravity of 1.2588 for the glycerin. Interpolating into a table of apparent specific gravities of water/glycerin mixtures gives the value of 98.1% glycerin, and 1.9% water by weight. Contrary to what I had previously stated, the glycerin bottle does state that it is 99.7% pure. We’ve had this bottle for a while, and it is now only half full, so some water has been absorbed from the air over the years, explaining why I found more than 0.3% water. Further evidence of this is that the bottle was slightly collapsed from the water being drawn out of the air in the bottle.
I have some display casting to do in 18 point on my Monotype Composition Caster, but I’m having a tough time finding a set of accessories that will work together to accomplish this.
My original intent was to use my 18 point display mould made by Monotype Corp. in England, and my display matrix holder for using English display moulds with American-style display matrices (part code 49A). This was my preferred route because this mould is in excellent condition. But, try as I might, I’m utterly unable to find this mat holder amongst my casting equipment.
Have you seen this mat holder?
Failing that approach, one alternative would be to use my American-made display mat holder with the English mould. The problem here is that for all display sizes other than 12 and 36 point, the position of the mould cavity is different on American and English display moulds. In particular, for 18 point, the English mould cavity is 3 points higher lower (in terms of matrix orientation) than the American mould cavity. The American mat holder has 4 sets of interchangeable inserts to modify matrix alignment, but none of these is designed to raise the mat—they all lower the mat (typically for casting all caps on a size body smaller that the face’s size) by varying amounts.
American mat holder insert sets. This is not a full selection of all 4 sets as there is one duplicate here.
It is plausible that I might be able to modify one of these sets to raise the mat by 3 points (or, to match the motif of the genuine sets, 0.050″, about 3.6 points), but one piece, in the upper left of each set in the photo, would have to be made longer so I might have to make this from scratch. This is not something I would look forward to because this part has a fairly complex shape. Maybe the lugs underneath can be trimmed to the same effect, maybe not. But that’s not really relevant because I got up and down confused.
The third approach would be to go completely with the American system, using an American mould and American mat holder. This would seem like a sensible way to have started, but my American display moulds are in sad shape: they are dirty, heavily used, and have unknown amounts of wear.
The American moulds work with interchangeable mould blades for the various body sizes. One basic mould (type U) is used for 24 to 36 point type, while the other (type T) is used for 12 to 18 point. I have never encountered any mention of display sizes between 18 and 24 so if such exists, I don’t know whether the T or U mould is used. The type T and U moulds actually only differ in one part, the right type block, which determines the position of the nick-side of the mould cavity (to the operator’s left, defining the bottom face of the type). All other size changes are done by positioning the right type block (defining the top face of the type). The actual casting size is determined by a mould blade kit which consists of 6 parts:
These parts are custom fitted to a particular mould, and you can see the serial number of the mould marked on the pieces. The spacer holds the left and right type blocks the correct distance apart and is the one kit part that is not tied to a particular kit or mould serial number—it is only specific to the body size. The spacer also limits how far the blade can open (to widen the type).
The blade stop, top blade, and main blade are fitted together to the mould so that when the blade is closed, its end (both main and top) are just flush with the front surfaces of the type blocks. This exactly ejects the type without the mould crossblock rubbing on the blade. The top and main blades are also fitted to each other so their ends are flush when the blade is partway open for casting type.
The top and main blade with blade stop, positioned as if installed in the mould, to show which surfaces determine the blade position. The back of the blade stop can be ground to adjust the overall blade position in the mould.
I have quite a few kits:
…but only two of these are complete and have the matching mould. Worse yet, here is my only 18-point mould blade kit:
The spacer and back cover are missing! As it turns out the back cover for an 18 point T mould and 30 point U mould are almost identical. There does not appear to be any close fitting involved (in a mould assembled with no front cover the blade can be visibly shifted up and down) so in a pinch I could use the back cover from a 30-point kit. The spacer is the critical part and since this is the only 18-point kit I have I can’t borrow the spacer from another kit.
So it looks like I might be making an 18-point spacer for an American display mould. The shape of the part is fairly simple and does not require extraordinarily precise machining work, except for its thickness, which must be accurate (and the faces parallel) to 0.0001″ to get properly-sized type.
In the meantime, I can only hope that my 49A display mat holder will pop up, saving me all this trouble!
As often happens concerning the alignment of American display mats, I had my up and down reversed. The 36- and 12-point mats have the same alignment in both systems. Sizes smaller than 36 that use the Type U mould (e.g. 30 and 24) are lower in the American mould than in the English one, while sizes larger than 12 that use the Type T mould (e.g. 14 and 18) are higher in the American mould than in the English one. As a result, the 18-point mats I want to use in the American holder must be lowered by 3 points to fit the English mould. This could be accomplished by using the first replacement mat holder insert set (41A19, 20, 21) which lowers the mat by 0.050″ except that is the one set I don’t have. Even though this set is a standard accessory to the mould it would be just as much trouble to make as a custom set to raise the mat.
This still pretty much leads me into the third option.
Several years ago I purchased at an auction a laboratory scale for doing precision weighing. Up until now it has sat in a corner collecting dust, covered with auction stickers and missing one cover.
I recently undertook to clean and fix it up. The scale is a Precisa 125A which has a range of 125g (about 4 ounces avdp) and a readability of 0.1mg (from later tests, that’s about the weight of a speck of dandruff).
The weighting platform is enclosed in a glass cabinet to keep it clean and also so air currents don’t disturb the weighing. This cabinet was missing its top cover although the handle for the cover was included. The remaining glass and cabinet frame had residue of masking tape adhesive, dust, and a sticker from the auction.
Some work with lighter fluid and a plastic scraper blade, followed by glass cleaner, gave me 3 clean panes.
I cut a piece of 1/16″ glass to fit the top. The rough cut was a bit too wide so I used a diamond cutting disk in my Dremel to remove a bit of the width. Then I polished all the edges smooth and round using a coarse diamond sharpening stone with soapy water to lubricate the work and wash away the glass dust.
I set up my Dremel drill press with a ⅛″ diamond grinding bit to drill the holes.
The diamond sharpening stone also turned out to be just the right height to support the scrap board that would hold the glass for drilling. The handle for the top cover is sitting on the glass, and needs two 4mm holes 4cm apart. I have a video of me drilling the holes and enlarging them to 4mm.
Here is the top cover with the handle attached. I used two tiny screws and washers recovered from dead computer equipment to hold it on. I also placed a narrow strip cut from PVC electrical tape along one edge to act as a bumper and dust seal where this cover meets the front glass of the cabinet. I use my heat gun to eliminate any stretch in the tape (from unrolling it) before applying it to the glass in the hopes that this will reduce the tendency to peel. There are already similar strips on the two side covers.
Here is the scale all cleaned up with the cabinet assembled.
And here it is weighing a short length of clear plastic (likely polypropylene) filament, weighing in at 7.2 milligrams. I couldn’t find anything in my shop that I could see and pick up that weighed less than about 5mg. I suppose I should have tried a shorter length of the filament.
Now that I have a precision scale I can try using my pycnometer to measure the water content of our glycerin.
Another project has now suggested itself: a dust cover for this scale, including pockets for the power cord and manual. While I’m at it I should also make covers for my microscope and the binocular scope mounted on my lathe.
That’s ‘pyc’ sounding like an ice ‘pick,’ ‘n’ as you would expect, and the rest rhymes with ‘barometer’ with the ‘nom’ being the stressed syllable.
This is a fancy name for a container which can be filled to a very reproducible volume of liquid (there are also pycnometers for powder). Even though the actual volume is only roughly known, and varies from one pycnometer to another apparently identical one, the fixed volume is still sufficient to measure the specific gravity of liquids.
This used one I bought holds about 10ml (⅓ of a fluid ounce). The stopper has a fine hole drilled through it up to the top. You fill the bottle to the brim, and insert the stopper. As the ground glass joint of the stopper seats, the last bit of excess liquid is forced out through the hole, leaving you with exactly the same volume of liquid in the bottle each time. Of course you have to clean and dry the outside of the bottle before weighing it.
Essentially, you weigh the empty pycnometer, weigh it filled with a reference liquid (almost always pure water), and weight it a third time filled with the liquid whose relative density you want to determine. A fairly simple calculation, which includes the current atmospheric pressure since the buoyancy from the air affects the apparent weight, yields the relative density of the tested liquid as compared to the reference liquid. When the reference liquid is pure water, this relative density is called the specific gravity. Because the result is actually the ratio of two densities, the facts that the volume of the pycnometer is not known accurately, and that the local gravitational force might not be known exactly (it varies from location to location) don’t affect the result. Details on the calculations can be found in this Wikipedia article.
One liquid I want to test with this is the glycerin we sell. Glycerin almost invariably contains at least some water since it is so hygroscopic, but the stuff we sell is not labelled as to water content. It would be useful to know this information, not only to keep our customers informed, but also since this can affect the recipe used for making composition (a rubbery material made of gelatin+glycerin+sugar+water) rollers for printing presses. Given the specific gravity of a mixture of glycerin and water, the fraction of water can be determined by consulting the appropriate reference table.
I’m just fixing up the precision weigh scale I have (which weighs to the tenth of a milligram), and once that it ready I can test our glycerin.
Someone has asked me to cast some type to fill out a font they already have. For reasons unknown, he has almost no capital S and is also short on capital C and T. He mailed me a sample type of each to ensure that the new type I cast would match the alignment of his existing type. Although there is a standard alignment for most fonts, this is not always followed and it is essential for new type to duplicate the alignment of the old type it will be mixed with, even if this alignment is non-standard.
As a preliminary to doing this casting work, I sat down for a few minutes to measure the size of the samples he had sent—something else I should also match. The result is this absolutely riveting video of me sitting at a table with a micrometer and slide rule. The video has been sped up at times where I have nothing to say about the process.
I found to my surprise that the type was cast 1 to 2 points narrower that the standard size. According to the markings on the matrices, they should have been 11, 10.5, and 11 points (0.1521″, 0.1452″, and 0.1521″) but were instead 0.1315″, 0.1187″, and 0.1422″ for the S, C, and T respectively. The faces of the type overhung to such an extent that the combinations “CT” and “ST” could not be set without a copper or brass space between the letters, or perhaps filing off some of the beard on the overhang.
The face in question is 18 point Swing Bold (Lanston Monotype #217). This is a script-style face with the lowercase letters all joining to each other as an imitation of handwritten text. The capital letters don’t join to the lowercase, though. The face also includes a lead-in stroke to use at the start of lowercase words to clean up what would otherwise be an abrupt start to the stroke (where it would normally join the previous letter). The specimen pages don’t show these lead-in strokes in use, though. Although they provide nice finish to the words, they can’t be used before the second letter of a capitalized word. Doing so would generate too much visual space between the capital and next letters. Perhaps these narrow-cast capitals are intended to be used with the lead-in stroke.
The following examples are approximations made from a scan from the specimen page for this face. The crowding of the S does not necessarily correspond to the width reduction of the type samples I have, but this shows the general effect.
This is a word taken from the text sample on the specimen page. Note how the lowercase letters join, but the start of the t is rather abrupt since it is right at the edge of the type body.
This is the effect you would get with the S cast narrow, so it can be closer to the t. Although some of the space is gone, there is still a rough transition from the S to the t.
This is what the word would look like if the lead-in stroke were added to the t and using a normal-width S. The start of the stroke looks much better, but the large space between the S and t make this look like two words (what is an “s tyle”?).
This is a combination of using the lead-in stroke and a narrow-cast S.
Whatever the reason these are cast narrow, I’ll have to verify that it wasn’t just a fluke that the samples I received were some rare narrow-cast letters from a case of otherwise standard width ones.
Someone had asked me for some parts for their Monotype Compositon Caster. Like me, he has the pump latch mechanism but he’s missing the modified pump spring rod and spacer sleeve, so he asked me to make a set for him, following the model from when I made my own a year and a half ago (see here and here).
New parts are in general not available. They are not being made, and it is a very long shot to get some new-old-stock items from places like the Type Archive in London. As a result, caster parts are often obtained by stripping down surplus casters which have somehow been saved from a trip to the scrapper. The pump latch mechanism, as well as most of the other parts for display casting, are distinctively recognizable for what they are, and generally take up little storage space. The spacer sleeve and extended pump rod are less distinctive, and the rod is bulky to store, and people don’t think of these as parts that break and need replacement, so they are more likely to be scrapped or at least lost in a jumble of miscellaneous parts.
So it is not uncommon to be in the situation of having the pump latch but not the pump spring rod modifications required to use it. To get around this problem when I ran into it, I made my own parts. Now, I’ve made another set of the same parts for someone else in the same predicament.
I have posted a video of me explaining their purpose and test-fitting them to my own caster. It was a good thing I did this because at first, the rod extension would not screw all the way onto the standard spring rod. The female threaded hole on the extension was too shallow, and also was not counterbored to allow for the unthreaded area on the end of the rod. The video wasn’t turning out well either as I found myself babbling aimlessly without some notes to work from.
The video is less than ideal… For part of it I’m showing the relevant page from the parts manual and things are pretty hard to see because the video is only 320×240 resolution, and by the time I noticed how the page would look I was done editing the video. The editor I am using (VSDC video editor) does not allow you to change the basic resolution of a project; you have to choose it when you create the project and live with it. There was quite a bit of cutting and splicing of the video clips and there was no good way to copy all those editing decisions to a new higher-resolution project.
There is one magic spot where my hand vanishes from the picture, only to come in immediately from the side again.
That strange facial expression where I seem to be peering just to the side of the camera is because I had my laptop there acting as cue cards.
One other note about the video: I mention using a temporary thread locking compound when joining the extension to the spring rod. This is so that you don’t later find the extension left behind if you remove the rod (the extension by itself would be difficult to remove from the caster because it would be almost completely inaccessible with any form of wrench). Something like Loctite 248 (the blue stuff) would to the trick. I don’t recommend a permanent thread locking compound (like Loctite 263 red) because then you would never get the parts apart. Also, only use the thread locking compound to join the rod to the rod extension. Do not use it anywhere else, particularly where the rod screws into the rod eye below the upper crosshead!
In a post about two years ago concerning our then-newly-acquired Kelsey press I mentioned that I had made my own composition rollers for it. Someone had inquired about how I made these, and it has taken me this long to get my notes and references together. This post discusses some of the references I had to work from.
My starting point was a document I found online, La Fabbricazione dei Rulli, which appears to be chapter 25 of a book entitled IL libro e L’Arte della Stampa, published in 1914 and reprinted in 1925, by libreria Editrice Internazionale di Torino. The extracted document was prepared by Gianolio Dalmazzo, director, regia scuola tipograficae di arti e affini in Torino. The document is in Italian, and most of the information presented here (including the bibliographical information) is from a mix of Google Translate and my own knowledge of Romance languages. So the document is Roller Fabrication, chapter 25 of The Book and the Art of Printing.
This document gives several recipes for rollers:
All these recipes include some form of gelatin, some form of sugar, and glycerine. Some of them add to this rubbery or resinous materials, probably for more strength and heat resistance. Such materials might make it difficult to melt down an re-use old rollers. I’m particularly wary of the recipe that uses boiled linseed oil, which will gradually and irreversibly harden with age as it oxidizes.
Within these recipes, there are unknown variations because the purity of the materials is uncertain. Molasses may vary as to sugar contents (both in terms of types of sugar and also other non-sugar solids), animal glues come in different grades, and glycerin is difficult to produce completely free of water. Furthermore, the recipes did not come with cooking directions.
For my purposes I wanted to limit myself to three basic ingredients: regular white cooking sugar (sucrose), woodworker’s hide glue (Lee Valley Pearl Hide Glue 56K50.05), and food grade glycerine (water content unknown). The gelatin in the glue provides the rubbery texture, the glycerine holds water to prevent the composition from drying out, but the sugar’s purpose isn’t obvious. Perhaps it makes the composition tacky. I would have to try making a mix with no sugar to see what it is like.
According to Water vapour sorption isotherms and the caking of food powders [Mathlouthi, M., Rogé, B. Food Chemistry Volume 82, Issue 1, July 2003, Pages 61–71], below 86% relative humidity sugar (sucrose) loses all its water, although it can also remain as a metastable glass (think peanut brittle) with higher water content. Stable water content depends on crystal size but is in the general range of 200ppm, essentially zero for our purposes. Above 86%RH sugar absorbs water and dissolves in it to form a syrup.
So if the room is going to be below 86%RH, some humectant must be included to prevent the rollers from drying out. On the other hand, above 86%RH, the rollers will absorb humidity uncontrollably and will melt into a gooey syrup.
The Soap and Detergent Association (New York) published a booklet in 1990 entitled Glycerine: an overview. One thing this clears up is that glycerin, glycerine, and glycerol are various names used inconsistently and often interchangeably for essentially the same compound, although in some circles, the choice of name depends on water content. I suppose “clears up” isn’t such a good choice of words, but this means it is not necessary to fret over the distinction.
Glycerine left in the open will absorb water from the air depending on the relative humidity. The following values were taken from a graph in this booklet (the glycerine and water percentages are by weight):
| %RH | % glycerine | % water | water/glycerine |
| 20% | 97 | 3 | 0.03 |
| 30% | 93 | 7 | 0.08 |
| 40% | 89 | 11 | 0.12 |
| 50% | 84 | 16 | 0.19 |
| 60% | 76 | 24 | 0.32 |
| 70% | 67 | 33 | 0.49 |
The last column is significant because it gives some hint as to how much swelling or shrinkage a roller can undergo between a heated room in the winter (20-30%RH) and summer conditions (60-70%RH). This also shows how it is important that the rollers be cast with the correct water content so they don’t end up the wrong size once they have reached an equilibrium with the current room air.
Gelatin is a family of proteins extracted from the collagen in animal bones, skin, and connective tissue by boiling then in water. It has the property that above a specific temperature, it dissolves in water to form a liquid (called the sol phase), and below that temperature, the sol gradually sets into a rubbery solid, the gel phase. The transition temperature depends on how the gelatin is produced but is typically a few degrees below body temperature so gelatin-based foods dissolve in the mouth. The water concentration only affects the strength of the gel (both in terms of elasticity and also rupture strength), not the transition temperature.
Grades of gelatin are rated for stiffness using the Bloom test, measuring the elasticity of a 6.67% gelatin solution stabilized at 10°C. The hide glue I use is rated at 150g Bloom strength, but there is another grade rated at 260g which I would expect would make much stiffer rollers.
When the composition is cooked, the gelatine gradually degrades, which reduces its strength and probably its transition temperature as well.
When made with pure gelatin and white sugar, the rollers should end up translucent white. If molasses and/or hide glue are used, the result will be a shade of brown. Sometimes powders are added to the composition (as the Paris White in one of the quoted recipes). These are chemically inert but the interfere with the elastic deformation of the composition so they make the roller harder without increasing its tear strength. They also make the roller opaque, so composition that would have been dark translucent brown might instead have a caramel colour. A roller made with pure gelatin and white sugar could have almost any colour one wants. Of course, the colour of the roller itself it pretty much immaterial, other that having a good contrast with the ink makes it easier to clean then thoroughly.
To use such pigments, the main difficulty would be to disperse them in the molten composition without trapping any air. Fine pigment particles have a strong tendency to form small clumps around microscopic air bubbles, and such bubbles would ultimately lead to a porous hard-to-clean surface on the rollers. The pigments are formed into a paste with one of the liquid ingredients (probably the glycerine) so no additional water is added to the recipe. The clumps and bubbles can be worked out by mulling the pigment paste or using an ink grinder, or it might be possible to extract them by applying a vacuum to the paste.
The composition is so low on water content (and what water is present is bound to other ingredients) that decay of the composition is unlikely. A moistened surface, say, due to bringing a cold roller into a damp room, might encourage growth of mould on the surface. The best way to avoid this is to keep the rollers dry so the mould doesn’t have a chance to establish itself in the first place. However, copper compounds have fungicidal properties, so a small amount of copper sulphate can be added for protection. This will add a bit of blue to the roller colour, so ones that would be brown tend to become closer to black instead.
In order to follow a recipe for composition rollers, you need detailed descriptions of the ingredients. Some variables, like water content of glycerine and molasses, can be compensated for by adjusting cooking time. On the other hand, variations in the impurities between different varieties of molasses or hide glue becomes a source of unpredictability.
Ideally you want the ingredients to start with the correct water content since extended cooking degrades the gelatin. As well, casting the rollers with too much water and letting them acclimatize later results in rollers that are too small and with their surface under tension making them vulnerable to splits and tears.
Composition rollers require care: they have to be kept below 86%RH and also below the gel transition temperature of the gelatin, or they will melt. They also have to be used in the range of humidity they were designed for; otherwise they will be the wrong size and hardness. In extreme cases a roller can tear because it is fragile from softness, or it can crack because the shrinkage around its core has puts it under a lot of tension.
One question already raised is: what does the sugar do? To test this I would make test batches with widely varied sugar content to see how they differ. My expectation is that the surface texture will be affected.
How does the stable water content vary with changes to the ratio of the three major ingredients? Again, I can try test batches poured out onto foil and weighed regularly over a few days to see how much humidity they ultimately absorb. The main uncertainty here is the water content of the glycerine.
How does the hardness vary with ingredient ratio changes? For this test I would have to find myself a Shore A Durometer for measuring the resilience of the composition.
We’ve had some people asking us about our Introductory Papermaking Workshop this February. They’re flexible about when they can be here, but we will probably hold it on a Saturday. If you might be interested in this workshop, please get in touch with us soon with your date preferences so we can pick a date get a big group together!