Slump Depth - Kiln Forming Myths 30

You can slump only 5cm or so per firing. If so, why don't drop rings fail?

This is really about a comparison of deep slumps and free drops.

Multi-stage Slump

·         The stages of a deep slump give more control of the design

·         The shape of the result is determined by the shape of the mould

·         The wall thickness will be much more consistent

·         For its size the deep slump is lighter than the drop. 

·         The multiple stage deep slump is designed to have the minimum of cold work.

Single Stage Slump

·         If you attempt to do a deep slump all at once you run the risk of ruffle around the upper edge. 

·         There are likely to be a large number of stretch marks on the outside.

·         The design will be distorted to varying degrees along the surface.

·         The wall thickness will vary greatly.

·         You lose the advantages of the multi-stage slump without gaining the advantages of a drop vessel.

Drop Ring Vessels

·         Drop rings require higher temperatures. 
·         The glass thins as it stretches through the ring, so you need to start with a much thicker blank than slumping. 
·         There will be a thick base in relation to the sides.
·         The design will stretch, and if properly designed will be very pleasing.
·         The walls of a drop vessel will vary from thick at the bottom to thin at the top.
·         The collar needs to be cut off the vessel and cold worked to smooth.
·         The process of falling through the ring needs to be monitored to avoid an excessive drop causing distortion or an insufficient drop causing the need for grinding a flat base for the vessel.

Reversibility of Boron Nitride

After using Zyp/MR97, can I sand it off and use kiln wash?

Some people are applying boron nitride to ceramic moulds for the "smoother" surface.  Boron nitride is an excellent separator for metal moulds and casting moulds whether metal or ceramic. But it has limitations, including the price and requirement for a repeated application at each firing.  Some are beginning to wonder if they can go back to kiln wash after having used the boron nitride. Some say you cannot unless you sand off the separator.

The general experience has been that you can't apply kiln wash on top of the boron nitride. It just beads up and flows off, because the boron nitride creates a non-wetting surface that survives relatively high temperatures.  The water in the kiln wash mixture merely beads up or washes away. This means the kiln wash in suspension has no opportunity to adhere to the mould.

The most accepted way to get rid of the boron nitride is by sandblasting. Then apply kiln wash as normal. The sandblasted ceramic mould previously coated accepts kiln wash with no difficulty. In the absence of a sandblaster, you can use a sanding pad. You do need to be cautious about taking the surface of the mould when using abrasive removal methods, as the ceramic is relatively soft in relation to the abrasive materials.

The difficulty of removal of the boron nitride means that you have to think carefully about which moulds you put it onto.  If the mould has delicate or fine detail, removing the boron nitride risks the removal of some of the detail.  This indicates that this kind of mould, once coated, should not be taken back to the bare mould to change the kind of separator.

Another use of boron nitride is to spray a very small amount on a fiber strip to be used for damming. This will give you fewer needles as it provides a non-wetting surface at relatively high temperatures. This allows the glass to slide down the fibre paper without hanging up and creating the needles.

One advantage of kiln wash over boron nitride is that you do not have to reapply every firing as with boron nitride. With the boron nitride you need to apply before every firing.  It is best to use a paint brush to dispose of any lose material before giving a light re-coating. Not a whole lot is required on subsequent coatings.

If you are using boron nitride to get a smoother surface to the object, consider using a lower slumping or draping temperature, as this will also minimise mould marks.  

Super Glue - Kiln Forming Myths 29

The use of super glue in the kiln causes cyanide gas

This is not true.  But because it is such a persistent belief, a lot of detail is given below.  In short the precautions are: 

• use the minimum amount, 
• use an organic gas face mask, 
• do not wear natural fibres or gloves, 
• let the glue cure before placing it in the kiln, 
• have the solvents at hand while using the glue.

Super Glue Safety

Super glue is frequently used as a temporary fixative in assembly of kiln forming projects. There is some concern about safety, as it is known that super glue is made from cyanoacrylate, which it is feared will break down in the kiln into cyanide gas.

Greg Rawls, a certified industrial hygienist says

"I looked at the MSDSs for several forms of super glue. The main component is Ethyl 2-cyanoacrylate, which has a TLV of 0.2 ppm which is relatively toxic. [However,] the thermal decomposition products are carbon monoxide and carbon dioxide. I did not see a reference to cyanide gas. However, as I recall cyanide gas dissociates into elemental carbon and nitrogen at about 800 F. Since you use it in such small quantities, I would not worry about it. In my opinion the worst thing that could happen is you glue your fingers to the glass."

Safety issues

To treat the safety issues seriously and determine if you feel Greg Rawls' view is justified, you need to look at the issues of toxicity, reactions, adhesion of tissue, ventilation, first aid and decomposition products in the whole context.

Toxicity

The fumes from cyanoacrylate are a vaporized form of the cyanoacrylate monomer that irritate sensitive membranes in the eyes, nose, and throat. They are immediately polymerized by the moisture in the membranes and become inert. These risks can be minimized by using cyanoacrylate in well ventilated areas. About 5% of the population can become sensitized to cyanoacrylate fumes after repeated exposure, resulting in flu-like symptoms. It may also act as a skin irritant and may cause an allergic skin reaction. On rare occasions, inhalation may trigger asthma. There is no single measurement of toxicity for all cyanoacrylate adhesives as there is a wide variety of adhesives that contain various cyanoacrylate formulations.

The United States National Toxicology Program and the United Kingdom Health and Safety Executive have concluded that the use of ethyl cyanoacrylate is safe and that additional study is unnecessary. 2-octyl cyanoacrylate degrades much more slowly due to its longer organic backbone that slows the degradation of the adhesive enough to remain below the threshold of tissue toxicity, so the use of 2-octyl cyanoacrylate for sutures is preferred.

Reaction with cotton

Applying cyanoacrylate to some materials made of cotton or wool results in a powerful, rapid exothermic reaction. The heat released may cause serious burns, ignite the cotton product, or release irritating white smoke. Users should not to wear cotton or wool clothing, especially cotton gloves, when applying or handling cyanoacrylates.

Adhesion of the Skin

Various solvents and de-bonders can be used. These include:

·         Acetone, commonly found in nail polish remover, is a widely available solvent capable of softening cured cyanoacrylate

·         Nitromethane

·         Dimethyl sulfoxide

·         Methylene chloride

Commercial de-bonders are also available.

Warnings include:

·         It is a mild irritant to the skin.

·         It is an eye irritant.

·         It bonds skin in seconds.

·         Any skin or eye contact should be copiously flushed with water and medical attention be sought immediately.

·         Do not attempt to separate eye tissues – the bond will separate naturally within a few days.

Precautions

·         Use goggles.

·         Do not wear cotton or wool clothing while using super glue

·         Ventilate the area well. Since cyanoacrylate vapours are heavier than air, place exhaust intake below work area. Activated charcoal filters using an acidic charcoal have been found effective in removing vapours from effluent air so the bench top air filters are suitable for use while using super glue.

·         Avoid use of excess adhesive. Excess adhesive outside of bond area will increase level of vapours.

·         Assemble parts as quickly as possible. Long open times will increase level of vapours.

Evaporation Effects

·         The effects of heating cyanoacrylate are not completely known. The flash point is known to be greater than 85ºC. As a precaution do not remain in the area of the kiln after that temperature has been reached.

·         The decomposition products are carbon monoxide and carbon dioxide. There is no reference in the literature to cyanide gas. It is highly unlikely that heat will cause the release of cyanide gas at any time during the heating. To be certain, you should make sure the evaporation of the glue is complete before firing the kiln.

See this tip for the use of super glue in kiln forming. http://glasstips.blogspot.co.uk/2013/12/super-glue.html

Relieving Stress at Corners

The most frequent locations of high stress in a piece is at corners or points.  The stress seems to be concentrated there and thus they become the most vulnerable parts of the piece.

Although the above image is of a plastic drawing triangle, it illustrates the point. The stresses are concentrated at the points and right angles whether inside or at the edge. The rainbow effect of some of the stress points show that those are the location of extreme stress.  If you see any of that in your glass, you need to check for compatibility and certainly anneal it again more slowly if it is compatible.  Remember though: slow annealing of incompatible glass will not enable incompatible glasses to fit together and become compatible.

Of course, the main thing that we do is to ensure the anneal is adequate to reduce the stress at these points.  It is important in a piece that has points, right angles and other abrupt changes in angle that you are more conservative in your annealing soak and cool. 

Further, if you are tack fusing, the stresses will be greater than on a full fuse. This is because the pieces of glass are not fully incorporated and tend to expand and contract independently of each other and of the main piece.  Also, the lower glass is shaded from the heat by the upper pieces on heat up. On cool down, the lower glass looses heat more slowly.  These two main effects, although there are others, require that the annealing is done much more slowly - two to four times more slowly than a piece of the same thickness.

One simple means of reducing stress before the start of the fusing process is to nip the corners off.  And slightly round the internal angles.  This requires only a very small piece to be taken from the corner or point to reduce the stress in the final piece. This is particularly important in tack fusing projects.

This nipping of the corners also removes the frequentl sharp points that some right and more acute angles develop during the cool down.  Glass, even of 6mm and more expands with the heat of the fusing.  As it cools toward the annealing temperature, it contracts.  The glass at the corners has to contract further than the edges, and so leaves a sharp point where it was unable to fully round. Removing only a small piece of glass from the corner removes enough mass to counteract this effect of contraction.

Hot Short Firings - Kiln Forming Myths 28

The hottest temperature for the least time always gives you best results.

It is difficult to imagine where or how this instruction arose.  Just as “low and slow” is not always the answer, so this also has its application, but not as a general practice.

In general, I try to get my fusing work done in 10 minutes at the working temperature.  Any less time there and I feel I am trying to go too fast. 

Advancing very fast normally requires a higher temperature than a slow advance, to get the same result.  Also with a higher temperature you do not need to have as long a soak as at a lower temperature.

It is more difficult to get repeatable results with fast firings.  A more controlled rate of advance will allow the controller to cope with any variations (e.g., power, or mass of material being fired) present. 

But you need to know why you are doing the AFAP for as short a time as possible.  It can be useful for small and jewellery scale items.  It certainly is not applicable to larger or thicker items. 

For slumping, it may be that the reverse of the headline suggestion could be the appropriate response.  Slow advances allow the glass to gently conform to the mould without excessive stretching.  This is also helped by using a low temperature and a long soak. 

These observations show that the injunction may be appropriate for some work, but most kiln work is better done with a slower, lower, longer approach.  This means slower rates of advance, lower target temperatures, longer soaks.

Spacing of Pieces on the Shelf

It is natural that we should want to put as much onto the shelf as we can to maximise the number of pieces from each firing.  But, when you are placing the pieces remember that glass expands as it heats up. When the glass is at its maximum expansion, it will be much less viscous than at lower temperatures and so will stick very easily to any neighbouring piece it touches.

Although the final size of a two-layer piece is the same at the end as the beginning, they do expand to a larger size during the fusing process.  My experience shows me that a 6mm piece can expand as much as 5mm, depending both on temperature and size.  This means that I treat 10mm as the absolute minimum space between pieces. But, because of the size of my fingers, my normal minimum placing is 20 mm apart as that is a comfortable space between my fingers and the other glass.

Thicker pieces expand to become larger after fusing than they were at the start. These pieces spread more during the firing than the 6mm piece.  A 9mm piece may expand by about 3mm at the finish – depending on size and temperature.  But during the firing, it may expand as much as 9mm. This means that 20mm is an absolute minimum between pieces that are 9mm thick at the edges, even though they may be only 6mm over most of the area.

The tip is to avoid over-filling your kiln shelf.  By trying to get too much production in one firing you may find a number of pieces stuck together at the end, thus eliminating any savings on glass or space. 

Didymium Lamp Working Glasses - Kiln Forming Myths 27

Lamp working glasses can be used to look into the kiln at high temperatures.

Definitely not! 

Didymium glasses are used by lamp workers to protect eyes against sodium flare – the yellow glow coming off the glass in the torch flame.

In kiln forming, the radiation that our eyes need protection against is infrared.  Welders’ goggles do this, but didymium glasses do not.  Welders goggles and helmets are much cheaper too.

Edge Working Options for Glass

There are a number of standard options for the worked shape of edges.  The simplest is to have a seamed edge, where just enough sanding is done to take the sharpness from the edge.

The next is to have an arris where more glass is removed, usually as a chamfer, but sometimes in a rounded, bullnose effect.  These are commonly used for glass that is to be toughened.

Flat chamfered and often polished edges are quite common also.

Bevelled glass is very common on mirrors as this reduces the reflection of the inside of the frame holding the glass.

As you can see from the attached illustration, there are a number of standard edge treatments, although some of them are uncommon.

The seamed, arrised and flat polished edges are easiest to create by hand grinding.  The other more fancy edges require machines.

Slump Point Test

Revised 7.7.21

At a time when we are all going to be trying a variety of glass of unknown compositions to reduce costs of kiln working, the knowledge of how to determine the slump point temperature (normally called the softening point in the glass manufacturing circles) and the approximate annealing temperature becomes more important.  This is called the slump point test.

This test can be used to determine both the slumping point and the annealing soak temperature. This used to be required when the manufacturers did not publish the information. It continues to be useful for untested glasses.

The method requires the suspension at a defined height of a strip of glass, the inclusion of an annealing test, and the interruption of the schedule to enter the calculated annealing soak temperature.

A strip of 3 mm transparent glass is required. This does not mean that it has to be clear, but remember that dark glass absorbs heat differently from clear or lightly tinted glass. The strip should be 305 mm x 25 mm.  If you are testing bottles, you may find it more difficult to get such a long strip.  My suggestion is that you cut a bottle on a tile saw to give you a 25 mm strip through the length of the bottle.  Do not worry about the curves, extra thickness, etc.  Put the strip in the kiln and take it to about 740C to flatten it. Reduce the temperature to about 520C to soak there for 20 minutes.  Then turn the kiln off.  

Suspend the strip 25 mm above the shelf, leaving a span of 275 mm. This can be done with kiln brick cut to size, kiln furniture, or a stack of fibre paper.   Make sure you coat any kiln furniture with kiln wash to keep the glass from sticking.

The 305mm strip suspended 25mm above the shelf with kiln furniture.

Place some kiln furniture on top of the glass where it is suspended to keep the strip from sliding off the support at each end. Place a piece of wire under the centre of this span to make observation of the point that the glass touches down to the shelf easier.

The strip held down by placing kiln furniture on top of the glass, anchoring it in place while the glass slumps.

Also add a two layer stack of the transparent glass near the suspended strip of glass to act as a check on whether the annealing soak temperature is correct. This stack should be of two pieces about 100 mm square. If you are testing bottles, a flattened side will provide about the same thickness.  This process provides a check on the annealing temperature you choose to use.  If the calculated temperature is correct there should be little if any stress showing in the fired piece.

The completed test set up with an annealing test and wire set at the midpoint of the suspended glass to help with determining when the glass touches down.

The schedule will need to be a bit of guess work.  The reasons for the suggested temperatures are given after this sample initial schedule which will need to be modified during the firing.
Ramp 1: 200C per hour to 500C, no soak
Ramp 2: 50C per hour to 720C, no soak
Ramp 3: 300C per hour to 815C or 835C, 10 minute soak
Ramp 4: 9999 to 520C, 30 minute soak
Ramp 5: 80C per hour to 370C, no soak
Ramp 6: off.

Fire at a moderate rate initially – 200C/hr to 500C - and then at 50C/hr until the strip touches down. This is to be able to accurately record the touch down temperature.  If you fire quickly, the glass temperature will be much less than the air temperature that the pyrometer measures.  Firing slowly allows the glass to be nearly the same temperature as the air.  

Observe the progress of the firing frequently from 500C onward, unless it is float glass you are testing. Then you can start observing from about 580C. Record the temperature in Celsius when the middle of the glass strip touches the shelf. The wire at the centre of the span will help you determine when the glass touches down.  This touch down temperature is the slump point of your glass.  You now know the temperature to use for gentle slumps with a half hour soak.  More angular slumps will require a higher temperature or much more time.

Once you have recorded the slump point temperature, you can skip to the next ramp (the fast ramp 3).  This is to proceed to a full fuse for soda lime glasses. Going beyond tack fusing temperatures is advisable, as tack fuses are much more difficult to anneal and so may give an inaccurate assessment of the annealing. Most glasses, except float, bottles and borosillicate will be fully fused by 815C. If it is float, bottles or borosilicate that you are testing, try 835C. If it is a lead bearing glass, lower temperatures than the soda lime glass should be used. In all these cases observation at the top temperature will tell you if you have reached the full fuse temperature. If not add more time or more heat to get the degree of fuse desired.

While the kiln is heating toward the top temperature you can do the arithmetic to determine the annealing point.  To do this, subtract 40C from the recorded touch down temperature to obtain an approximate upper annealing point.  The annealing point will be 33C below the upper point.  This is approximate as the touch down temperature is by the nature of the observation also approximate.  

The next operation is to set this as the annealing soak temperature in the controller. This will be the point at which it usually possible to interrupt the schedule and change the temperature for the annealing soak that you guessed at previously. Sometimes though, you need to turn the controller off and reset the new program.  Most times the numbers from the last firing are retained, so that all you need to do is to change the annealing soak temperature.

The annealing soak should be for 60 minutes to ensure an adequate anneal. This may be excessive for 3 mm glass, but as the anneal test is for 6 mm, the longer soak is advisable. The annealing cool should be 83C/hr down to 370C. This is a moderate rate which will help to ensure the annealing is done properly. The kiln can be turned off at that temperature, as the cooling of the kiln will be slow enough to avoid any thermal shock to the annealing test piece.

When cooled, check the stack for stress. This is done by using two polarised light filters. See here for the method. 

Squares of glass showing different levels of stress from virtually none to severe
 (no light emanating for no stress to strong light from the corners indicating a high degree of stress.)

If the anneal test piece is stressed there is a problem. There could be a number of reasons for the inadequate annealing. It could be that the glass has devitrified so much that it is not possible to fuse this glass at all. If you also test the suspended strip for stresses and there is very little or none, it is evidence that you can kiln form single layers of this glass. You now know the slumping temperature and a suitable annealing temperature and soak for it, even though fusing this glass is not going to be successful.

Other reasons for stress due to inadequate annealing could be that the observations or calculations were incorrect.  

• Of course, before doing any other work, you should check your arithmetic to ensure the calculations have been done correctly. I'm sure you did, but it is necessary to check.  If they are not accurate, all the following work to discover the difficulties will be fruitless.

• The observation of the touch down of the suspended strip can vary by quite a bit - maybe up to 15C. To check this, you can put other annealing test pieces in the kiln.  This will require multiple firings using temperatures in a range from 10C above to 10C below your calculated annealing soak temperature to find an appropriate annealing soak temperature.

• If stress is still showing in the test pieces after all these tests, you can conduct a slump point test on a strip of glass for which there are known properties. This will show you the look of the glass that has just reached touch down point as you know it will happen at 73C above the published annealing point.  You can then apply this experience to a new observation of the test glass. 

Fast Ramps - Kiln Forming Myths 26

Firing AFAP harms your kiln.

This may be a hangover from the time when ceramic kilns were being used commonly.  There certainly is a tradition of this kind in ceramics practice.  However, nowadays we are firing in kilns with light weight bricks or fibre, or a combination of the two, making this less relevant.

The light weight bricks are much less subject to temperature shock than the dense ones.  Fibre is completely unaffected by rapid changes in temperature.

Firing as fast as possible is much more likely to damage the glass you have in the kiln than the kiln itself.  It is also likely to have over runs in temperature.  The controllers compare the actual increase in temperature with that requested by the schedule.  It takes time for the controller to “learn” the rate of advance being achieved within the kiln.  On fast rises in temperature, it does not have the capacity to stop the input of energy early enough to prevent the kiln temperature rising beyond that which is programmed.  This can lead to unexpected and unexplained results (unless you think about the effects of an AFAP rate on the controller's computer).

Dog boning in Slumping

Often even in shallow rectangular moulds the sides pull in during the slump.  To know what things to try to correct this effect, you need to understand why this effect is occurring.  These two pieces show the effect in different ways.

ebay 0916_slump_01

 This slump shows that even with thick glass the sides curve inwards even on shallow slumps.

theglassundergroundnj.org

This slump shows the interesting effect that the further up the piece you look, the greater the curvature. This relates to the greater amount of movement required by the glass to conform to the mould at the outer edges.

Why

During the slump of a rectangle or square the whole shape of the glass sheet is changing.  It is slightly stretching to form into the “hollow” of the mould, but it cannot stretch evenly all over, especially at the corners.  If you think of the analogy of Draping a piece of cloth into a rectangular depression, you will find it wrinkles up at the corners if you smooth it at the sides. This indicates the material is attempting to overlap there as it does not have a dart to take up the excess cloth.

This similar to what is happening to the glass sheet.  It is relatively thicker at the corners than along the sides.  Therefore, it does not slide down the mould at the corners as on the sides. It is simply thicker and is compressed by the movement of the glass at the sides.

Prevention

The question is how to use that knowledge to avoid or minimise the dog boning during the slump.  There are probably lots of methods, but three have occurred to me and others.

Add more material along the sides.  This involves fusing a piece with shallow arcs rather than straight sides.  This gives more material to counteract the dog boning effect when slumping a rectangle.  The difficulty is getting the proportions of the arc correct in relation to the length of the sides. You also need to ensure the arcs on the sides are not so much larger than the mould that they slump over the edge.  This means the whole piece will need to be cut smaller than the mould.

Remove material at the corners.  This takes the opposite approach.  To avoid the increased amount of glass at the corners, you remove some of it.  That is, you round the corners of the pieces to be fused. How much you will need to round the corners is a matter of experience, but is a shorter learning curve than cutting the edges in an arc.

Reduce the temp and increase soak time.  This approach requires less skill in cutting a shape.  It relies on giving the glass time to relax into mould with a minimum of stretch.  You need to find the lowest practical temperature at which to slump.  This will be the temperature at which you can first see the deformation of the glass in the mould.  Hold the temperature there for as long as it takes – possibly one or two hours. It is likely that you will still need some rounding of the corners of the glass, but only your experience will determine that, and if so how much.

Cold work the edges until straight.  This can be done by hand or by machine.

Further information is available in the ebook: Low Temperature Kiln Forming.

You Can Re-fire 3 Times Only - Kiln Forming Myths 25

Bullseye claims that you should only fire a piece 3 times

       

No. They only say the glasses are tested three times and that you are on you own after that.

There is not a general answer that can be given for the number of times you can fire a piece.  In general, Bullseye glass (and probably others, although they do not state what their limits of confidence are) can be fired three times with confidence.  Beyond that you need to do your own testing.

Bullseye states: 

At Bullseye, glasses known to be fairly stable are tested by firing to a top temperature of 1500°F (815°C) and soaking for 15 minutes before annealing. Once cooled, these tests are viewed for stress through polarized light and graded accordingly. We fire glasses known to be less stable three times to make sure they'll perform well under multiple firing conditions, such as those used to fuse and slump a plate.

If you have plans for multiple re-firings, tests are needed. The tests should replicate the temperatures, colours and thickness of the proposed project.  You probably do not need to reproduce the size of the project in these tests though.

Results from each firing should be tested for stress and these tests should include a test for annealing each time. 

You may wish to note that I have fired up to 7 times on several two layer with powder pieces.  Many people fire more times successfully.  It is my belief, but I have no proof, that multiple firings of a piece to slightly lower than full fuse will be more successful than each of them being to the full fuse.  My practice is to go to a rounded tack each of the firings subsequent to the first full fuse, but the final firing will be to a full fuse if I wish a gloss finish.  If I do not, my final firing will be about 10C - 15C below full fuse.

Dog Boning Causes

I fired a one-layer piece of glass and it shrank. What did I do wrong?

Cause

This result relates to the thickness that glass, under kiln forming circumstances achieves.  The combination of gravity and viscosity lead to this effect.  As the glass becomes less viscous (more runny), the surface tension is greater than gravity and so it becomes thicker at the edges.  This additional glass is supplied from the edges and to some extent from the interior. The glass in the middle becomes thinner, allowing in certain circumstances bubbles or holes to appear.

This illustration from Fusedglass.org shows the effects of gravity, which is related to mass, and viscosity.  The lack of mass means the surface tension allows the glass to draw up to be come thicker, forming the classic dog boning appearance.

Prevention

Knowing why this occurs allows you to take come precautions, when firing single layer pieces, to help prevent the shrinkage, often known as dog boning.

Fire larger

You can cut the glass larger than the final piece will be.  After firing, you cut it down to the size you want.  You may have to do a bit of cold working to get a rounded edge to the glass before any further processing.

Fire lower  

You can fire at a lower temperature for a longer time.  You will need to observe to determine when the glass begins to shrink. Either stop the temperature rise and soak there for a time, or reduce the temperature a little and soak for as long as needed to get the surface texture wanted.

Fire oval or circular pieces.  

With these shapes the shrinking is not so obvious, as it occurs all the way around.  With rectangular pieces, as the glass shrinks, the corners become thick more quickly and so do not shrink as much, giving that dog bone appearance.  Rounded pieces become thicker all the way around more evenly and the shrinkage is not so obvious.  However, you still get thinning in the interior which can lead to holes or bubbles, so observation is still necessary to prevent excessive thinning and bubble formation.

Fire thicker

The real prevention is to fire two layer pieces as that is the thickness at which viscosity, surface tension and gravity are in balance.  So the glass does not change size at kiln forming temperatures.

Cold work

Alternatively, you can cold work the edges back to straight parallel edges.  This can be done by hand grinding or by machine.

Further information is available in the e-book: Low Temperature Kilnforming.