How Not to Make Matrix Blanks

Posted on by kpmartin Posted in Kevin, Matrix Making — No Comments ↓

The Lanston Monotype display matrices are essentially rectangles of brass (or aluminum) approximately 1⅛×¾×0.094″ with two corners cut off at an angle. It would seem that the best material to start with would be 1/8×¾″ brass strip, and the first step would be to mill a total thickness of 0.031″ off both faces to produce smooth surfaces and the desired thickness. Both faces have to be milled because the surfaces of the strip as purchased are a bit uneven, and both surfaces must be smooth and parallel to be able to make engraved matrices.

When I tried to do this, the work setup I had was to hold the brass down on the mill table with two metal bars bolted to the table. I would be able to work the length between the bars, then loosen the bars and move the strip along to work on the next section.

IMG_8286

This shows the bar set up for cutting the second section. The circular edge towards the left clamping bar is the end of the previous section’s cut.

Although this setup allowed me to mill off the rough factory finish, it proved to be unsuitable for getting any sort of accurate thickness. Although the bar is straight, and tightly clamped to the table, it was slightly bowed up between the clamps.

A 0.004″ feeler gauge easily slips under the brass strip near the center.

A 0.004″ feeler gauge easily slips under the brass strip near the center.

I’m not quite sure what causes this. It could be microscopic dirt on the table near the clamping bars, or it might be that the table of the mill compresses a bit under the clamp bars, and this compressed material pushes upwards on either side of the bars. It is not caused by the extra length of strip overhanging the end of the mill table, as supporting that made no difference.

Regardless of the cause, this produces uncertain variations in the thickness of the milled strip, since the raised area can spring down somewhat under the force of the cutter. Even if the thickness variation were of a definite amount, having the thickness of the strip vary by several thousandths of an inch would make the matrix blanks essentially useless. They would have to be milled individually to the correct thickness and parallelism (as Jim Rimmer did using his Ludlow Supersurfacer).

I was hoping to avoid one step of piecework, but if the blanks have to be milled individually, there is little point in trying to do this thickness milling before cutting the individual mats.

Perhaps some other way of clamping the strip would work better. This would involve clamping inward against the edges of the strip. The clamp jaws would have to be lower than the finished thickness of 0094″ and would have to be designed to prevent the strip from lifting. Clamping this way would also increase the length that could be milled in a single pass, before shifting the strip along the table.

Tagged with:

Tramming the Mill

Posted on by kpmartin Posted in Basement Workshop, Equipment Acquisition, Repair, and Maintenance, Kevin, Matrix Making — No Comments ↓

In order to get a good finish when milling the top surface of a part, as I will be doing to prepare matrix blanks, it is necessary that the axis of the mill spindle be perpendicular to the surface of the milling table. If this is not the case, one side of the endmill will cut a bit deeper than the other side, and depending on the relationship between the tilt and the motion, the result will either be a sloped surface or a dished cut, slightly deeper in the center. If multiple passes are made, you get a shallow sawtooth pattern that shows as a ridge at each pass, or sort of a scalloped surface.

The process of adjusting this relationship between the spindle axis and the table is called “tramming” the mill.

On my Sherline mill, I have a spacer block installed between the head and the saddle to allow the mill to cut a bit further from the back edge of the workpiece. The attachment from the head to the spacer and from the spacer to the saddle are the same: A central post has a beveled groove cut into it, and a bevel-tipped locking screw presses sideways against this, acting as a wedge and pulling the post tight. Tramming the mill consists of making a minute rotational adjustment of the head about the post so the spindle axis is perpendicular (side-to-side) to the table. The joints also have keys to hold the head roughly vertical or horizontal, but the keyways have enough play in them that they can’t assure enough accuracy.

Head AttachmentThis joint, in my opinion, could benefit from more holding power. As it is currently designed, vibration and/or unexpected (or expected!) high cutting forces can knock the mill out of tram. Having the weight of the motor all on the right side of the pivot post tends to mean that the head eventually ends up out of tram, high on the left, by as much as the keys in the joints allow.

To test the tram of the mill, a test indicator is mounted sideways on the spindle. On my mill I use the drill chuck and a right-angle adapter to do this. The feeler of the indicator is about 3″ from the spindle, and the indicator reading shows how much the feeler is being pushed up by the block placed under it.

Tramming Setup I adjust the Z position until I have a non-zero reading on one side, then manually rotate the spindle to the other side, and using the same block, take another reading. Ideally the readings should be the same. The adjustment is made by very slightly loosening the locking screw on the head mount, twisting the head in the appropriate direction, re-tightening the locking screw, and measuring again. It is somewhat of a hit-or-miss procedure: sometime you don’t adjust far enough, and sometimes you go too far, even ending up more out of tram but in the opposite direction.

This evening I got it to this condition:

Good ReadingsThe readings differ over a 6″ span by 0.0005″ (half a thousandth of an inch), or about 1 part in 10,000, which is about as good as I can hope for with the hit-or-miss adjustment. To put this value in perspective, if I were to use a ⅜″ endmill to finish the top of a part in multiple passes, the ridges from the “sawtooth” surface would be 0.375/10,000 i.e. 0.0000375″ high, certainly not noticeable by feel, and perhaps not even visually.

When the mill is as out of tram as the keyway play allows, there is about 0.010″ difference, or about 1 part in 600, and multiple passes with the ⅜″ endmill would produce ridges 0.000625″ high, easily found by feel. I had already noticed this when I was making a point block for an American Monotype display mould, although further hand-finishing steps made the ridges unimportant for that project.

For finishing matrices, I would use a fly cutter, which cuts like a 1.5″ endmill would. It would be done in a single pass, so the main concern would be the dishing which would leave the centre of the matrix low. As it turns out the depth of this dishing would be about the same as the ridge height just calculated. A matrix 0.000625″ low in the centre would be unacceptable; one that is 0.0000375″ low would be excellent.

With the mill set up as it is, and provided that I don’t try any heavy cuts that might knock it out of tram again, I should be able to mill the ⅛×¾” brass bar stock I have to the proper thickness for making the matrices and be assured of flat and parallel surfaces.

An Interem Solution to the Pneumatic Valves Problems

Posted on by kpmartin Posted in Caster Computer Control, Equipment Acquisition, Repair, and Maintenance, Kevin, Matrix Making, Monotype Composition Caster — 3 Comments ↓

This week I’ve been occupied with (finally) putting away all the display mats I bought last year at the Anderson and Skyline matrix auctions. They have been cluttering up the shop since last summer with their variety of packaging. I finally took some time to make more dividers so I could file the mats in the storage boxes I use. Making the dividers involved some dreary repetitive work at the metal shear, then stacking the strips and cutting notches in them on the metal-cutting band saw. Most of these mats are now either in their boxes or in my own surplus section. Unfortunately when making the dividers I miscounted so I have to make a few more before I can finish, but at least now I have the procedure down pat.

The problem will then be reduced to finding a place to store all the boxes because the cabinet I was using does not have enough room. I might have to just make my own cabinet. Later on I can go through the boxes one at a time, cast everything and make proofs, and verify how my holdings match the font contents listed in the specimen books.

Also, preparatory to making some display matrix blanks, I’ve cleaned and lubricated my mill. I will also have to check some of its adjustments before I proceed, to ensure I get flat mats. Jim Rimmer used a Ludlow Supersurfacer to finish his mat blanks but that might have been a case of everything looking like a nail when all you have is a hammer.

Tonight I revisited the problem of my pneumatic solenoid valves. I need four valve units that work properly in Normally Closed (N/C) mode and give good airflow when energized. My valve collection turned out to have 12 Normally Open (N/O) valves of which 6 are electrically different (in what manner, I’m not sure yet) and 6 N/C valves. I determined that I could modify the internal springs to convert an N/O valve to an N/C one, and also that good or bad airflow seemed to be a property of the shutter in the valve. I was stalled on the problem of modifying the shutters to improve the air flow.

Tonight, though, I decided to avoid wasting time trying that. I only need 4 working valves. I have 6 valves that work in N/C mode so I don’t have to fool with springs; all I had to do was collect 32 shutters that gave good airflow and swap them around so they are all in 4 valve bodies. I tested all the shutters on the 6 N/O valves and noted which gave good airflow. I disassembled four of the N/C valves and the 6 donor valves and put the good shutters into the N/C valves. Swapping the shutters only took about half an hour (and would have taken less had I not jostled one of the disassembled valves and spilled all the shutters on the floor).

The result is four valves that give reasonable air flow in all eight ports. There are still noticeable differences in air flow so once I have a prototype assembled and connected to the caster I may find that some of the ports still need better air flow. There are still some donor valves available though, and I can raid those if necessary.

Tagged with:

Converting the Outline to a Pantograph Template

Posted on by kpmartin Posted in Kevin, Matrix Making — No Comments ↓

In the process of cutting a matrix for the 18-point @ I made up to go with Swing Bold, I need to make a template to run the pantograph tracer against. The edge of the template has to be high enough that the tracer is unlikely to lift up and go over the edge. This can happen because the tracer tip will be slightly rounded and while tracing the pattern I might accidentally lift the tracer a bit. The top edge of the template might have a bit of a chamfer or radius as well. If the tip is raised enough the rounded edges will conspire to lift it more until it pops out.

This can be avoided if I constantly press down on the tracer but that is tiring and wears the tracer. Ideally the tracer is not touching anything unless you are against the edge of the template.

One other consideration is the strength of the template material. In order to get the spindle to cut the matrix, some lateral force must be applied to it, and this is generated from the lateral forces the user applies to the tracer tip. The force the user applies is multiplied by the reduction ratio of the pantograph, so for instance when doing a 10:1 reduction, a 1-pound (-ounce, -Newton, -whatever) push on the tracer will produce a 10-pound (-whatever) force on the spindle, barring friction in the pantograph mechanism.

Depending on the matrix material, cutting depth, spindle RPM, and cutter geometry this sets a lower bound on how hard one must push the tracer to get anything cut. When the tracer runs into the edge of the template it stops, and cutting stops, so all the force must be resisted by the tracer against the template. If the template is too soft the tracer will dent it. Of course, the tracer is blunt compared to the cutter thus giving a lot of leeway, but generally there are limits on how much softer a material can be used for the template relative to the matrix being cut.

In the film Making Faces the late Jim Rimmer showed his methods for cutting a matrix. His overall workflow from concept to matrix started with sketches on paper. Using the Ikarus font design program for the Mac, he would enter points from his sketches with a digitizer puck, and clean up the contours on the computer. The outline would be printed at 30× enlargement directly onto Bristol board or heavy card stock and cut out with a sharp knife. He would glue this to a board using PVA glue, and dry the assembly in a nipping press. Using this template he would do a 2× reduction cutting the outline into a slab of type metal, and from that he would to a 15× reduction onto the actual brass matrix. He used two different pantographs for the two reduction steps.

The explanation for this two-step process isn’t really delved into, but I suspect that the idea is that the hardness of the type metal is intermediate between that of the brass and of the Bristol board. Trying to cut a brass mat directly from the Bristol board template would probably destroy the template, especially if one wants to make multiple passes (e.g. a rough cut with a large bit not quite to full depth, then a finish cut with a fine bit to finish depth). On the other hand, cutting the type metal from Bristol board would be a single-pass operation and would require less force on the tracer so the template would last. The type metal template then is hard enough for use in cutting the brass matrix. If you want more than one matrix (one for a friend, or one scaled to a different size) the durable type-metal would be re-used, rather than the fragile cardboard one. Even when doing this two-step process Jim said in the film that one had to be careful to avoid “bruising” the cardboard.

I have a stereo caster which I could use to cast slabs of type metal like Jim used, but with the mould spacers I have the slabs would be type high and very heavy. Audrey suggested using the laser cutter at KwartzLab to cut a template from something like 1/8″ MDF. If I do that I also have to account for the thickness of the laser kerf but again that can be done using an oversize tracer once I measure the kerf width. What I might start with is Bristol board reinforced with a non-water-based varnish (water-based products would distort the Bristol board) or perhaps epoxy cement, mounted on a spacer to provide positive engagement for the tracer.

In any case I can think about templates for a while because before I have need for one, I have to make some cutting bits more suitable than the ones that came with the pantograph, make a tracer tip of a suitable diameter, and make some blank matrices.

Tagged with: , ,

Matrix Making: Handling Ink Gain

Posted on by kpmartin Posted in Gorton 3U Pantograph, Kevin, Matrix Making — No Comments ↓

Letterpress printing has a property called “ink gain” where the inked area left on the printed sheet is a tiny bit larger than the type used to do the printing. Because of this, when designing type (which includes designing the matrices) one has to make the type slightly lighter that the desired printing.

The actual amount of ink gain depends on several factors which come into play when printing, including the amount of ink applied to the type, the ink’s consistency, and the pressure on the form rollers. In theory the ink is only applied to the face of the type and so should produce print of the same size, but in practice, some ink ends up on the shoulders of the type, and even with clean shoulders, the pressure of printing causes the ink to squish outwards a bit.

This shows the effect of ink gain on a capital C in 18 point Swing Bold, taken from Monotype’s own specimen book:

A magnified image of the C. The ink gain is actually visible because the edge of the type is marked by a faint lighter line. All the black area beyond that is the ink gain.

A magnified image of the C. The ink gain is actually visible because the edge of the type is marked by the outer edge of a faint lighter stripe where the ink was squeezed out. All the black area beyond that is the ink gain.

This is a superposition of two tracings. The outer pink outline is a tracing of the C from the specimen sheet. The inner darker contour is a tracing from a photo of the face of the actual type.

This is a superposition of two tracings. The outer pink outline is a tracing of the C from the specimen sheet. The inner darker contour is a tracing from a photo of the face of the actual type.

One thing that is, however, consistent is that for any given printing technique, the extra area caused by ink gain is a constant width, except around hairlines where there is less ink to squeeze out. This in turn means that you can’t make various sizes of a letter by scaling its image if the compensation for the ink gain is part of the letter outline. A letter properly compensated at a large size will have too little compensation when shrunk, and vice versa. The compensation has to be applied after scaling the letter to its correct size.

Based on theses tracings, it looks like the ink gain is about 0.0017″ which does not sound like much, but even on a relatively large (18 point) and fat face like Swing Bold this is about 7% of the stroke width and failing to compensate will make the cast type print too bold to match the rest of the font.

When using the pantograph, one way to get a fixed offset like this is to use an oversize tracer to trace out the template. Normally the diameter of the tracer should be equal to the scaled-up diameter of the tip of the cutter bit. By using an oversize tracer, the cutter is kept a fixed distance away from the scaled outline. So for instance if I were scaling down by a factor of 10 with the pantograph, I would want the radius of the tracer to be 10×0.0017 = 0.017″ oversize. This makes the diameter 0.034″ or about 1/32 of an inch larger than 10 times the cutter diameter.

This method will not work for letters with widely varying stroke widths or fine serifs, since once lines get fine enough the ink gain effect reduces. How fine? Consider how wide the light stripe is in the above image; as the stroke becomes narrow enough that the light stripes from either side meet each other, the supply of ink to squeeze out into the “gain” becomes limited. For strokes finer than that the gain compensation becomes some fraction of the width of the stroke. A fixed compensation by fudging the tracer size will no longer work and the compensation for the ink gain has to be part of the outline.

One other effect of the ink gain is that sharp corners on the type become rounded. This can be seen in the superimposed tracings where the lower end of the C has a sharp corner but the printed outline is rounded. The rounding isn’t even just a radius centered on the corner of the type but it is centered somewhere inside the corner. This is because the ink on the corner of the type is squeezed out in two directions, and so there is less ink available to cover more perimeter. There is a similar effect for inside corners, except that there is an excess of ink to squeeze out and it fills the corner more than would otherwise be expected. One can compensate for these effects, albeit imperfectly, and that is part of the art of type design.

Designing Swing Bold @ 18 point

Posted on by kpmartin Posted in Kevin, Matrix Making — No Comments ↓

Having decided to use my pantograph to try making a display matrix for my Monotype caster, and having chosen to make an at-sign (@) to match 18-point Swing Bold, the next task was to actually examine other letters in this font whose elements could be adapted to form the @.

Obviously, the lowercase ‘a’ is the starting point. To form the circle I also had a close look at uppercase ‘O’ and ‘C’ with the latter in particular providing the shape for the end of the tail.

Rather than designing the character on paper and having to fool around with the inevitable odd scaling to reduce it to the needed size, I used font-design software to make the outline, using photomicrographs from the specimen sheet as shape references.

The software I chose to use is the MS Windows version of FontForge, which is an open-source font editor available online.

Although the actual font design work is generally straightforward, there are several aspects of this program that are causing me some grief. Some of these may be due to the fact that the program does not use native windowing, but is actually an X Windows program, which launches an X Server to communicate with my screen, mouse, and keyboard.

One minor thing is that menus don’t cancel under the same circumstances as they do with native MSWindows programs, so when I decide I don’t want a menu after all, or want a different menu, the mouse clicks I would naturally use to do this don’t work here.

The Print dialog does not seem to contain the buttons required to actually do the printing. This could perhaps be something I could fix by playing with the X Windows configuration, but I haven’t had time to delve into that at all.

One big problem I’m having is that the various file open and file save dialogs don’t work well with drive letters and, as far as I can tell, don’t get along at all with UNC paths (\\computer\sharename\etc). I have at times been able to switch such dialogs to a different drive letter but even that is unreliable. I might be able to work around this by adding junction points (similar to Unix hard or symbolic links) to my file system so I can reach other places, but I would have to place such links on all the hard drives to ensure I could get anywhere regardless of where I started.

Another problem I find is that, when drawing the glyph outlines, the points that define the outline do not stand out enough. They are drawn in different shapes depending on what type of point they are (the point type constrains the reverse and forward tangents in various ways), some points (the curve start point, selected points) are specially marked by colour, and some points don’t actually define the outline but are annotations added by the program to mark inflection points and extrema. The actual point symbols are small, and weakly contrast with the background making them hard to distinguish. If you have a background image as a reference, the outline itself can also blend into that making it invisible. Having the ability to change the colurs and size of the points and colour of the contour would improve this dramatically.

I have also often found that I unexpectedly have more than one point selected, so when I move the point I wanted to move, several others move with it. If I notice this I can Undo to put things back and repeat the move with the correct selection, but if I were zoomed in enough that some of the selected points were off-screen I might not notice the mistake, continue with my work and later have to fix the points that “mysteriously” moved. I haven’t actually caught what causes these multiple selections, though.

One feature I would like to have (perhaps it is there but I just haven’t found it) is keyboard commands to select the next or previous point in the outline, and to select the forward or reverse control point for the current point. These are useful when the points are too close to make out without an extreme zoom in. Since two points can coincide (and a control point can coincide with the point it controls) it may be that no amount of zooming would help.

Although as mentioned above the Print dialog is unable to print, it can still export an image which I can then print using other programs. In this dialog you supply a text string along with a point size, a choice of rendering engine, and anti-aliasing choice (all of which can differ within the text). You also pick a value for resolution in pixels per inch, global to the entire string, and the rendered characters are shown on the screen. The bitmap can also be exported to a file.

One problem is that the saved image does not have the resolution specified in it, so opening the image in another program will show the letters the wrong size.

Another problem is that I want to use this feature to print oversize versions of the letters to make templates for the pantograph. The obvious way (at least, to me) to do this would be to select a resolution multiplied by the enlargement I want. So for instance if I wanted to see my 18 point @ sign at 10× magnification printed at 300dpi, I would ask the dialog to show me an 18-point @ sign at 3000 (10×300) dpi. That doesn’t work, though. Any time I change the resolution to any value above 300dpi the change is ignored. Instead I have to ask for a 180-point @ sign at 300dpi. In this case it amounts to the same result because I don’t currently have alternative outlines for different sizes. I do, however, plan to add the other sizes to the same font project, and each will have slightly different outlines. Then my only choice would be to generate and install the font as a Windows font, and use some other program (perhaps a custom one, or a PowerShell script) to draw the letter I want at the resolution I want.

This window also has problems if the generated bitmap is much larger than the window.

Finally, this form seems to scale my outlines so the interline spacing is the size I have selected, and despite all my attempts at removing extra vertical spacing, this form seems to insist on adding 200 design units (my design is drawn 1000 units tall) so I actually have to ask for a point size 20% larger than what I want. All this is ignoring the small error introduced by the fact that the point they are using is probably not 0.013833…″

Despite this rough start, I have an @ sign that looks good enough to work with, and I have the template outline printed off. I still have to work out the details of turning this into a template proper, selecting a cutting bit and tracer to match, and accounting for ink gain. Not to mention making a matrix blank to cut!

Making a Display Matrix, but Which One?

Posted on by kpmartin Posted in Kevin, Matrix Making, Matrix Markings — No Comments ↓

It seems natural for me, as a first project using my new pantograph engraver, to try making a matrix. Rather than designing a whole new family of typefaces (or even a single typeface), I decided to make a single character. What people seem to be looking for the most lately is the at-sign @, for use in e-mail addresses.

In the Monotype matrix numbering system, the @ is considered a “Commercial Symbol” and so has markings between 20P and 29P on the “face” section of the marking (located on the front of a cellular matrix). The body size is marked as usual in the “size” section of the marking (located on the side of a cellular matrix), and being a symbol, there is a third designation of 2 because @ is Commercial Symbol number 2 (number 1 is the percent sign) marked on the back of a cellular matrix. These markings are generally augmented with alphabetic codes indicating set-width and unit-width.

For instance my specimen book lists 3 different 10-point @ symbols, coded “2-10Vb 20Po”, “2-10U 20Po”, and “2-10U 25Po”. The ‘20P’ or ‘25P’ indicates a Commercial Symbol, the ‘2’ indicates an @ symbol specifically, and ‘10’ is the body size in points. ‘Vb’ indicates a set-width of 9½ points, while ‘U’ indicates 10 points. Finally the ‘o’ indicates the correct width is 18 units wide (i.e. 18 18ths of the set-width). In general with symbols there is no correlation between numbering and appearance; in this case the first one is a slightly narrower version of the second one, while the third one has slightly less variation between the thinnest and thickest strokes.

The number of styles available varies with body size. The smallest is “2-4½Y 20Po”, the largest in cellular matrices are “2-12T 20Po”, “2-12S 20Po”, and “2-12S 25Po”. Although other symbols are available as cellular matrices up to 18 points, the @ symbol is not among them. In contrast to the spareness of variety in the @ sign, my specimen book lists 27 different percent signs in 12-point. Many of these have “Symbol” designations like “1-12S 26Po” but some have face-specific designations like “1-12Tb 609Jo” which is a 11½-set (‘Tb’) 18-unit (‘o’) percent sign made to match face 609J (20th Century Ultrabold). Such markings do not specify that it is a Commercial Symbol, so a pilcrow ¶ (Reference Symbol number 1) specific to the same face would have the same markings if one existed and was the same width.

The @ symbol is also available as display matrices, the smallest being the 14-point “14-21 No.2” in the specimen book. The ‘14’ is the body size, the ‘No.2’ refers to Commercial Symbol number 2, and the ‘21’ seems to refer to the 20P-29P range on cellular mats for Commercial Symbols. Other examples, though, disprove that last statement, such as the dagger “18-25 No.3”; the dagger is Reference Symbol number 3, and so would have a second number in the 40’s (Reference Symbols) rather than the 20’s (Commercial Symbols).

Some other symbol display mats, though, lack the ‘No.n’ designation, and in such cases the second number is the number of the matching text face, so for instance there are percent and pound-sterling signs both marked “24-42” which have an outline style clearly intended to go with text face 42 (DeVinne Outline). In these cases, though, there is nothing identifying the mat as a Commercial Symbol nor naming the specific symbol; you have to recognize it from looking at the actual mat cavity shape, just as you have to distinguish an ‘A’ from a ‘p’.

The largest display @ sign is “36-21 No.2” and there is exactly one of each standard display size. There are no alternates nor any face-specific ones.

So I have set myself a project to make a matrix for casting an @ symbol specifically designed to match a face. The font I have chosen is 18-point number 217, Swing Bold. There are two reasons for this choice: One is that I have recently cast some Swing Bold and have done some close-up study of the letter forms because my mats did not match the sample letters I given to match. The other is that this face is a script style (somewhat, at least) and so the join of the ‘a’ and its enclosing circle flows naturally. The origins of the symbol are not known for certain and there are many theories, but they pretty much all seem to come from informal handwriting rather than calligraphy and thus the flow from the ‘a’ to the circle is natural with script or italic faces. It is much less clear how the ‘a’ and circle should be structured and styled if one wanted to match a roman face or, worse yet, blackletter.

Such a mat would appear to be rightly marked “18-217”.

If this all works out I’ll do it for other sizes, and maybe expand to include ‘/’ and ‘#’ which are also becoming popular. Making them for other faces is a more long-term thing.

From drought to downpour

Posted on by kpmartin Posted in Gardening, Kevin, Weather — No Comments ↓

We’ve had several weeks here with essentially no rain, and the garden has been looking rather sad. I planted tomatoes, cucumbers, melons, and basil around May 24th, with promise of rain (finally!) in the forecast but it did not materialize, so I’ve has to hand-water the transplants.

Yesterday the rain came with a vengeance! It rained so hard that there was a big washout in the garden.

The main washout. The land slopes towards the camera. Soil washed out from the upper part of the garden deposited in a delta in the foreground, almost smothering some eggplants.

The main washout. The land slopes towards the camera. Soil washed out from the upper part of the garden along the schoolyard (about 12m away) deposited in a delta in the foreground, almost smothering some eggplants. The darker soil is composted manure which I had spread on the soil and has now been redistributed by the water.

The finer particles that didn't deposit in the delta continued to the first trench of potatoes, filling the trench with mud.

The finer particles that didn’t deposit in the delta continued to the first trench of potatoes, filling the trench with mud. I was going to fill the trench as the plants got bigger, but the weather has done it for me. I might actually have to rescue some of the plants that have been covered.

Water flowing from the schoolyard dug away at this 30cm slope, especially where I had dug up a large weed recently. The cucumbers plants (against the trellis), melon (center right) and onions (far right) were not really affected, now were the weeds (left).

Water flowing from the schoolyard dug away at this 30cm slope, especially where I had dug up a large weed recently. The cucumbers plants (against the green trellis), melon (center right) and onions (far right) were not really affected, nor were the weeds (left). However, just beyond the lower right of the photo, the flowing water dug away about 5cm of soil, probably taking out some beans I had planted.

Quite a bit of soil also washed onto the lawn at several places along the edge of the garden.

Quite a bit of soil also washed onto the lawn at several places along the edge of the garden.

This erosion probably only damaged some planted beans, because it mostly flowed where I had nothing planted. However there are other areas of lesser erosion that might have done in some carrots, beets, and parsnips that haven’t yet sprouted, and I think I lost at least one Brussels Sprout plant. Things are just too muddy to take a closer look yet.

Another 10 to 15mm of rain are expected today, but as a steady rainfall, not all at once like yesterday.

 

Challenge Guillotine (still) for sale

Posted on by kpmartin Posted in Sold — 1 Comment ↓

We still have our Challenge 265 power guillotine for sale.

I’ve removed the VFD and converted the wiring back to plain 208V 3-phase. The problems with the VFD tripping turned out to be due to a defective current sensor in the VFD itself, and the motor was not drawing the current levels mentioned in the previous posting.

We have also reduced the asking price from $825 to $600.

More information on Matrix pneumatic valves

Posted on by kpmartin Posted in Caster Computer Control, Equipment Acquisition, Repair, and Maintenance, Kevin — No Comments ↓

This week I was having another look at the pneumatic solenoid valves I want to use to run my Monotype from a computer. I was somewhat disappointed that the valves I had were the Normally Open (N/O) type and designed for higher pressures than I would be using.

While doing some tidying up I noticed that the valves did not all have the same part number on them!

The original valves I tested were BX758DE2A324, rated for 0-8 bar (0-120PSI) and normally-open operation, but the ones I noticed lately were BX758DE1C324, rated for 0-4 bar (0-60PSI) and normally-closed operation. A quick examination of all the valves revealed I had six of each of these, and six BX758XX2A3JJ (0-8 bar N/O but not sure about these other than JJ means special driver electronics are required to control the solenoid power).

I tested one of the newly-found valves and it did indeed seal properly in N/C mode, which is great because this will substantially reduce power consumption and heating of the valves. Unfortunately, despite the lower pressure rating, their flow at 20PSI when open varied between almost none and barely adequate, as it did with the first valve I tested.

I took apart the valve and found, as I expected, that there was more pressure from the spring push rods holding the valves closed. Opening the back of the valve to remove the springs revealed the difference:

Matrix Valve SpringsThese springs are tiny, about 2.7mm (0.106″) diameter. The lower spring is the weak one from the N/O valve. It has many coils and is thus a soft spring, and when installed it is hardly compressed at all, thus generating little force holding the valve closed. The upper spring is the strong one from the N/C valve. It has fewer coils making a stiffer spring, and its longer free length means it is compressed on installation, generating a much stronger force holding the valve shut. The spring push rods are identical in the two valves.

Both springs are made from the same wire. As an experiment I stretched one of the soft springs from the N/O valve and trimmed off the excess turns to make a spring resembling the one from the N/C valve, installed it in the valve and tested it. It gave a good seal when closed.

Thus is appears I have the N/O vs. N/C problem solved. I have six valves manufactured for N/C operation, and six more that I can convert to N/C by altering the springs. Twelve valves are enough to make 3 control units to run 3 casters.

As for the airflow in these valves, I charted the shutter thickness against subjective air flow as I had done with the first valve, and found a similar relationship between the values. One difference is that in the N/C valve, the shutters appear to be about 0.05mm thicker for a given airflow than those in the N/O valve previously tested. This much thickness difference would amount to about two units of airflow difference so on the assumption that the second valve has more space in it allowing good flow with thicker shutters, I’m going to be trying swapping shutters between valves to see how it affects air flow.

So for my attempts to shrink the rubber seals on the shutters, keeping them in boiling water for several hours (the mildest way I could think of to force out absorbed oils) had no effect.

Top