Book of Errant Pages

What is this picture and why is it representative of failure?

This picture shows my CNC mill in progress. Attached to the table are the supports and linear rails for the x axis. The part being held by the engine hoist (ominously foreshadowing my error) is the gantry. This component would ride the x axis’ rails to move back and forth along the table. The y axis linear rails can be seen attached to the gantry. The z axis and tool holder are not installed in this image.

The heavy steel is very rigid and does not rack even when being driven from only one side with a moderate load at the other. So why is this a failure? The key lies in the aforementioned ‘heavy’ steel. I had intended for the gantry to be installable by a single person. It turns out I am in fact capable of building a gantry so heavy I cannot lift it.

The gantry assembly will likely be installed and removed many times over the course of the mill’s construction. Even using the lift it is very difficult to line up all the parts which need to be attached to each other and damaging the surface of the rails is a very real danger.

So this is a failure of design. The designer (me) failed to take into account the weight of the component which would need to be installed compared to the available resources needed to install it (unfortunately also just me).

I already have the aluminum with which I intend to build the new gantry. Hopefully the new design will work out better though I am not thrilled at the prospect of drilling and tapping all of those holes over again.  

An epoxy composite is simply an epoxy resin which has other stuff mixed or laid into it before it hardens. The most familiar composite material made with epoxy is carbon fiber. Though we call the end product a ‘carbon fiber’ item it is more technically a ‘carbon fiber reinforced polymer’ item with the polymer in this case referring to the epoxy resin. Parts made this way are very light for the strength they offer but are expensive.

Reinforcements for polymers are typically cut fiber strands or whole fabric sheets. These relatively large reinforcements are typically ‘laid up’ into a mold and then saturated with resin. I wanted to use other less conventional (and less expensive) materials with epoxy resin as a filler to see what I could create with a specific focus on castable materials. A castable resin mix would need a viscosity low enough to run into and fill the parts of the mold. In addition the filler material would need to address some of the thermal problems encountered when creating thicker parts in epoxy.

As epoxy cures it releases heat. The hotter the epoxy is the faster it cures. These two aspects of thermosetting polymers (including epoxy) limit the thickness which can be cast. If the part is too thick it will not be able to dissipate the heat generated during the curing process resulting in a thermal runaway. In a thermal runaway the heat from curing speeds the curing generating more heat causing a rapid spike in the temperature of the curing resin. The addition of a filler will increase the volume of the epoxy and function as a heat sink to regulate the speed of the reaction.

Samples created for this experiment were of two sizes. The smaller is approximately 40mL in volume while the larger (such as the above pictured) are approximately 90mL in volume. This first sample is about 50% aluminum powder by volume. When working with fillers it is important to consider how much space the filler will consume as compared to the resin. Since it is troublesome to measure these materials by volume their densities are consulted and the mass to be mixed in is determined which allows the appropriate amount to be scaled out.

Reducing resin volume to less than 60% when using aluminum as a filler significantly increases the viscosity of the mixture making it both more difficult to mix as well as pour.  

Because of the iron present this sample has ferromagnetic properties though the resin’s presence greatly diminishes them.  

This 90mL sample is 20% aluminum, 30% iron. The addition of the aluminum seemed to decrease the viscosity of the resultant mix compared to the previous iron sample. Otherwise it is exactly what one would expect compared to the previous iron or aluminum only samples.

This 40mL sample is the first of the marble powder samples. Powdered marble makes up 40% of the volume of all the marble samples. All individuals presented with these samples determine them to be some type of stone with a minority guessing them to be marble. None supposed them to be a composite.

After this experiment I discovered cultured marble used in bathroom vanities and many other solid surface countertops are created using a similar process. These samples are different in their use of an epoxy resin as compared to the more conventional polyester resins used in commercial products.  

This sample had a black ink added after the mixing of the resin with the marble powder. The contrast is more pronounced in other samples. It is key the colorant not be added until after the resin has been completely mixed. Excessive mixing of the ink into the resin would create a uniform distribution of the colorant removing the striated effect seen.

Blue and black colorants were used in this sample. Various effects are possible via the manner in which the colorant is added and mixed into the resin as well as how it is poured into its mould.

This sample was filled with both marble powder and whole pieces of marble in ratios similar to those used for course and fine aggregates in concrete. Both items (the epoxy sample and concrete) are similar as they are both composites of a large aggregate (to take up most of the volume), a fine aggregate (to fill the smaller spaces), and a binder to hold everything together (cement in the case of concrete and epoxy in the case of the sample). The red spots on the sample are the uncleaned remnants of the polishing compound used to polish the marble chips. The marble powder and resin matrix does not seem to respond well to polishing compound though it can be sanded to a relatively fine grit.

For a comparison of how a ‘clear’ resin (one without reinforcements or modifiers) cures compared to a resin with filler consider the following. The 90mL samples discussed above are just over an inch in depth and did not exceed 80F during their curing process. A 90mL block of clear resin cast into the same mould as the above samples exceeded 240F during its curing process. The high temperature experienced during curing caused the piece to deform in its mould. This might be correctable with a rigid mould (the mould used for these samples was a flexible silicon cookie bar mould) but this would likely lead to problematic stresses in the sample.

In conclusion these fillers seem to be capable of allowing epoxy to be used to cast larger parts than normally possible with clear resin. General appearance seems to be highly mutable based on added colorants and fillers. Further tests will be necessary to determine the machinability and general strength of these samples.

For working in soft materials such as paper, softwoods, or plastic you may want to consider a scalpel instead of an X-ACTO knife.

The classic #11 X-ACTO blade has a thickness of 0.02” whereas the comparable #11 scalpel has a thickness of 0.015” making it easier to move though the material it has cut. This narrower blade also makes the scalpel a bit more flexible than the X-ACTO. My experience so far has shown the scalpel to be easier to use when making tight and detailed cuts. They also seem to last as long as my X-ACTOs in similar applications.

The attachment method for an X-ACTO normally involves using a screw to apply compression to the blade to hold it in place. If the user gets an X-ACTO stuck in a piece of material and pulls back with enough force to overcome this compression the blade will leave the handle.

 A scalpel attaches differently. It slides down a grove in the handle and snaps over the end ensuring it is not possible for the blade to come off of the handle without lifting the tab at the back.

Somewhat surprisingly scalpels can be had for cheaper than X-ACTOs. From Amazon a pack of #11 X-ACTOs with 5 blades can be bought for $4.05. The equivalent scalpel, also a #11, can be had in a 100 pack for $17. That works out to an X-ACTO costing $0.81/blade compared to a scalpel at $0.17/blade.

X-ACTO knifes have a numbering scheme similar to but not totally comparable to that used for scalpels. In both systems the blade number indicates the shape of the blade and not its size. Some blades in both systems refer to a similar profile.

The #10, #11, and #22 blades of both systems are comparably shaped. Some blade numbers do not match such as the #15 (which in X-ACTO is actually a saw) so be careful when acquiring direct replacements for your X-ACTOs.

Scalpels will come individually packaged in a sealed sterilized film. The sealed film alleviates the need for an oil film to protect the blade in storage (which is common with X-ACTOs) so there is nothing to clean off of a new scalpel before use.

If you want to give a scalpel a try go out and buy a #3 handle and some #10 (round point) or #11 (angled point) blades. There are many different profiles but most of the ones you will likely use in a crafting scenario will fit the #3 handle.  

There are many stories whose plot involves an object compelling its owner to use it. If the object were something like a sword you would likely have either a horror story or murder mystery. I often think this is very true for tools. Once you have a hammer everything starts to look like a nail and nails exist to be hammered. As I own a variety of tools I occasionally find novel uses for them. This weekend it was the micrometer I heard calling. As this is a measuring instrument I needed something to measure. I decided this would be paper.

There is much paper at the Barnes and Noble so there I headed early on a Saturday morning to measure their great stocks of it. The looks from both patrons and clerks was interesting. I enjoyed trying to imaging what they thought I was doing especially as they likely had never seen a micrometer before.

I had thought there would be a great variety in the thicknesses of the paper between volumes but was surprised to discover there was not. Conventional paperback novels are basically low resolution ink holders made of natural fiber paper and came in at 4 thou (four thousandths of an inch or .004”) with exceptional examples as low as 3.5 thou or as high as 5 thou. Black and white graphic novels are thicker with an average thickness of 6 thou. Color content seems to be mostly printed on glossier paper using synthetic fibers to achieve a higher resolution with better color fidelity and averaged around 3 thou. Black and white content printed on synthetic media also tended to average around 3 thou.

Larger format items such as color how to guides or coffee table books were more varied. Almost all are full color and ranged from 3 thou to 5.8 thou in thickness.

In hindsight I suppose this is not surprising to find the thickness of the paper tied to its application. Keeping standards for paper thickness also allows manufacturers of printing equipment to make reasonable assumptions about what their systems must accept.  

I do not cook, or bake, or execute any other from of food preparation beyond pouring cereal into a bowl milk. For this reason I am deficient in the realm of cooking implements. Thus when my sister arrived at my house with the intention of baking cookies in my oven I scrambled to find substitutes for the needed tools. The impending adventure led to some alternatives which I believed demonstrated greater usefulness than their traditional counterparts.

In the two years I had owned my house I had never used the gas oven. Little trust was placed in the oven’s metered dial indicating temperature and as there was no display of current temperature we would be unable to judge when preheating was complete.

My IR temperature gauge filled the role of thermometer admirably. In addition to the simple ability to register a temperature the meter may be aimed at different segments of the oven’s interior. This allows the user to find hot spots in the oven without laborious experimentation.

After the cookies were placed within the oven, whose preheating had been empirically confirmed, it occurred to us I was not in possession of any form of oven mitt. Various inferior and unimaginative alternatives were offered before I derived the ultimate solution.

Welding gloves being designed to protect their user from molten metal proved most effective. So great is the insulation of these gloves the user may maneuver hot surfaces at a leisurely pace. There is no haste necessary to drop a hot item before the heat becomes unbearable. As quality gloves may be had for $25 I cannot imagine a case in which a conventional oven mitt would excel them.  

Among the various loots I transported home after visiting my family for Christmas was a cake of most delicious construction. On arriving home I decided to cut the cake and store the individual pieces in the freezer. At this point I realized I had no implement which would neatly cut the cake. This was not a surprise for my expansive collection of power tools is inversely mirrored by my nonexistent supply of cooking utensils.

After some thought I decided to use dental floss. If wire cutting works for cheese, which is nominally more dense than cake, floss should work for cake.

This worked very well. The floss cut the cake easily and as its length is easily varied it is possible to cut the entire length of the cake at one time.

My original intent was to press the floss through the cake and then thread it out of one side at the bottom, thus cutting in one pass. This produced a cut edge so fine the icing seemed to reconnect after the floss had passed and so I drew the floss back up through the cake thereby cutting with the floss twice. The image shows the result of this double cut.

 

Having multiple monitors makes working on almost any task easier. The additional displays allow a great number of windows to be visible concurrently thus obviating the need to waste time looking through the taskbar for them. There are a few disadvantages though the chief being the consumption of vast amounts of desk real estate. To address this issue I constructed a monitor stand to hold all of my displays above the desk. This allows the utilization of space under the monitor for something more useful than a monitor stand.

  This stand is constructed of 32 feet of steel U channel, 16 feet of angle iron, 4 triple jointed monitor mounts, and 73 bolts. The U channel is SuperStrut which may be found at Home Depot. The frame utilizes two 80 inch segments of SuperStrut as its primary upright members. These vertical members reside on the outside of the desk and connect via ½ inch bolts through the desk to parallel elements on the inside of the desk. Tightening these bolts compresses the back plane of the desk between these parallel vertical members. This lets the desk take the load transferred from the stand and distribute it over a larger area than would be possible with bolts and washers alone.

Three 60 inch horizontal members connect the outer vertical supports together via right angle brackets. This forms two rectangles above the level of the desk. Spaces within these areas are candidates for monitor mount locations. When the horizontal location of a mount is determined two strips of angle iron are attached via angle brackets to the upper and lower horizontal members which define the area the monitor is being mounted in. These strips run vertically and parallel to each other with a ¾ inch gap between them.

   The arms I choose to use for this stand each have 4 holes which would conventionally provide a means of mounting the arm to a wall. In my case they are mounted to the vertical angle iron strips. A rectangular piece of 1/8 inch steel is drilled to match the hole pattern in the base of the mount. The steel plate is placed behind the strips of angle iron while the mount is attached to it via bolts from the front. These bolts pass through the gap between the strips. When these bolts are tightened the mount and steel plate compress the sides of the angle iron securing the mount.

 In addition to the monitor mounts this stand has also been fitted with a pair of backlights. Four monitors put out a considerable quantity of light and the presence of a dark background behind them can be uncomfortable over time. These lights are simple 18 inch under cabinet lights normally used in kitchens. Plywood was cut, glued, and painted in the form of a J hook which rests on the center

horizontal member of the frame.  

I am fond of writing with my Mont Blanc Meisterstuck rollerball however my fear of it coming to harm prevents it from leaving my desk at home. So the idea occurred to me to attempt to replace my Pilot G2 Limited’s internal refill with a Mont Blanc refill.

The process of changing out the refill was disappointingly simple. In fact further study revealed many fine pen refills would fit my G2 Limited body with little modification(Mont Blanc rollerballs and ball points) or no modification(Pelikan rollerballs). In the case of the Mont Blanc rollerball it is only necessary to remove a small portion of the plastic nub at the far end of the refill. The shoulder of the pen will fit into the pilot with no changes.

I started writing with my Mont Blanc modified pilot at the office and after some time I began to notice difficulty in discerning between my modified pen and my conventional G2. Curiosity struck and I decided to conduct an ad hoc survey to see if the occupants of my office could tell a difference between the modified and conventional G2.

Fifteen people were presented with a silver G2 Limited which contained a Mont Blanc rollerball refill as well as an unmodified coal gray G2 Limited. Each was simply asked which they would prefer to write with and why. Eight preferred the conventional G2 with the balance more favorable to the modified pen. Of especial interest was the language used to describe individual preferences. Regardless of which pen was preferred it was described as smoother than the other.

Line preference was evenly distributed between the G2’s thicker line and the Mont Banc’s thinner line. To be technically accurate here I must comment that these pens both produce lines of a around 16 thousandths of an inch. The G2’s line appears thicker because it is darker due to its greater ink deposition rate.

Of additional interest were comments made of each pen’s weight. Several users immediately noted the greater weight of the modified pen doing so instantly upon lifting the second pen without going back to the first for comparison. The normal G2 has a mass of 24.6 grams while the modified pen comes in at 27.4 grams. While this makes the modified pen 10% heavier than its counterpart I was still impressed with the ability of an individual to detect a 2.8 gram difference in mass with such speed.

Presently I find myself alternating between them based on task. When taking swift notes I make use of the G2 while reserving the Mont Blanc for slower tasks.