Wednesday, 14 March 2018

Cooling the Kiln

“Why does my kiln take so long from boiling point to room temperature?”

The rate at which a kiln cools is dependent several factors:
  • ·        The mass of the kiln. Some kilns have dense insulation bricks.  These are very good at holding heat, and release it slowly.
  • ·        Its insulation characteristics. Other kilns have light weight bricks or fibre insulation. Both these materials have less mass and can release heat quickly at high temperatures, but much less slowly at lower temperatures.
  • ·        The environment. The temperature of the surroundings has a big effect at lower temperatures.  The amount of air movement around the kiln also influences the cooling rate at these lower temperatures.

The physics of heat transfer determine the cooling rate. if all other factors are the same, the rate of temperature fall is faster when there is a greater temperature differential.  And it is slower where the temperatures are closer together.  You can see this by comparing the rates of fall at 800⁰C and at 300⁰C.  It is much faster at the higher temperature and slower at 300⁰C.  You will also notice that the kiln cools more slowly at the lower regions when the outside temperature is high than when low.

Rather than waiting hours or days for the kiln to get 
to room temperature, there are some things you can 

·        Open any vents or peep holes your kiln has. Not only are peep holes good for observing the progress of the kiln work, they are important in cooling.  Their relatively small size insures that there is not such a great air exchange that could cause thermal shock.  The temperature at which you do this is relative to the thickness or variation in thickness of the pieces in the kiln, of course.

·        Open the kiln lid/door a little. As the temperature fall rate reduces, you can crack the kiln open a little.  Many times, you need to put a prop under the lid to keep it open only a little.  Again, this should only be done at a low enough temperature to avoid any thermal shock to the glass.

·        Create greater air movement around the kiln.  You can of course create greater air circulation around the kiln by opening doors and windows, or by a fan.  If you use a fan, it is best to avoid direct air current from the fan onto the kiln. This is because when the vents or lid are open, dust can be spread over the glass and throughout the studio.  If using a fan, it is best to have the kiln closed.  Some kilns have powered ventilation to speed cooling, but these are usually industrial.

How do I tell if I am cooling too fast?

The risk of opening your kiln after the end of the second part of the annealing cool (generally around 370⁰C) is thermal shock from the relatively cool air contacting the glass and cooling one part too much, causing a break or fracture.

You can select how fast a cool rate is safe for your piece and programme that into the controller down to room temperature.  Doing this does not use any more electricity than simply turning the kiln off.  The controller will only put more energy into the kiln if it is cooling more quickly than the rate you set. 

And this is the point of programming to room temperature.

When you vent your kiln, and have the controller set for a cooling rate, it will only add more heat if you have opened the kiln too much.  If you hear the controller switch on the elements, you know to reduce the size of the opening, because it is cooling faster than you set the rate to be.  This makes for a safe, but more rapid cooling than just letting the kiln cool with no ventilation.

"My controller shuts off when I open the kiln."

If your kiln does not allow any opening of the lid/door without the controller switching off, you need an alternative.  In this case, you will need to take note of the temperature drop over set periods to learn if the temperature is falling too fast or too slowly.  Usually 15-minute intervals are all that is required.  Record the temperature at the switch off and before venting the kiln. Vent the kiln. Fifteen minutes later record the temperature. Multiply the difference by four to get the hourly rate.  If that rate is above the one you intended, close the venting a little.  If it is less, open the venting a little more. Then record the temperature after another quarter of an hour. You continue to do this until you are satisfied you have settled on the rate of cooling you intended.

You must exercise patience with the cooling.  

The larger, thicker, more important the piece is, the more caution is required. 

Wednesday, 7 March 2018

Kanthal vs. Nichrome

Both Kanthal and Nichrome are high temperature wires.

Kanthal is the trademark (owned by Sandvik) for a range of iron-chromium-aluminium (FeCrAl) alloys used in resistance and high-temperature applications. The first Kanthal alloy was developed by Hans von Kantzow in Sweden.

“Kanthal alloys consist of mainly iron, chromium (20–30%) and aluminium (4–7.5 %). The alloys are known for their ability to withstand high temperatures and having intermediate electric resistance.”  So, it is often used in kiln elements.

“Kanthal forms a protective layer of aluminium oxide (alumina) when fired.”  This layer resists further oxidisation of the elements when firing.  Aluminium oxide is an electrical insulator with a relatively high thermal conductivity.  Ordinary Kanthal has a melting point of 1,500°C.

“Kanthal is used in heating elements due to its flexibility, durability and tensile strength.” Its uses are widespread, with it being used in home appliances and industrial applications as well as glass and ceramic kilns.  As an aside, it is being used in electronic cigarettes as a heating coil as it can withstand the temperatures needed in this application.
Based on Wikipedia and other sources.

Nichrome is an alloy of various amount of nickel, chromium, and often iron.  The most common usage is as resistance wire.  It was patented in 1905.

“A common Nichrome alloy is 80% nickel and 20% chromium, by mass, but there are many other combinations of metals for various applications.”  Nichrome is silvery-grey, corrosion-resistant, and has a high melting point of about 1,400°C.

It has a low manufacturing cost, it is strong, has good ductility, resists oxidation and is stable at high temperatures.  Typically, nichrome is wound in coils to a certain electrical resistance, and when current is passed through it, the resistance produces heat.  This is probably the most common material used for kiln elements.

When heated to red hot temperatures, the nichrome wire develops an outer layer of chromium oxide, which is stable in air, being mostly impervious to oxygen.  This protects the heating element from further oxidation.  However, once heated the nichrome wire becomes brittle and must be heated to red hot before bending.

Based on Wikipedia and other sources.

Sunday, 4 March 2018

CoE Varies with Temperature

Information from Bullseye shows that the Coeficient of Linear Expansion changes rapidly around the annealing range.

The following is from results of a laboratory test of Bullseye clear (1101F)
Temperature range.......................COE
20C-300C (68F -­ 572F).................90.6
300C-400C (572F - ­752F).............102.9
400C-450C (752F - 842F).............107.5
570C-580C (1058F-1076F)............502.0

Bullseye glass is probably typical of soda lime glasses designed for fusing.

The change of CoE by temperature is further illustrated by Kugler (a blowing glass) who state their CoE by temperature range. Remember CoE is an average expansion over a stated range of temperatures)
CoE 93 for the range 0C-300C
CoE 96 for the range 20C - 300C
CoE 100 for the range 20C - 400C

The extension of the range by 100C has a distinct effect on the average expansion over the (larger) range. 

This shows why it is not helpful to refer to CoE without also mentioning the range of temperature.

In addition, here is an illustration of the effect. 

(If the owner of this illustration comes across this, please let me know, as I have lost the source)

Wednesday, 28 February 2018

Cordierite/Mullite vs. pizza stones or tiles

Description of the materials

Cordierite refractory shelves are generally combined with mullite to achieve low expansion rates.  These are most often manufactured as solid slabs, although there is an extruded version with hollow channels along the length, given the trade name corelite.

Cordierite is magnesium, iron and aluminium in a cyclosilicate form (or rings of tetrahedra).  It is named after its discoverer, Louis Cordier, who identified it in 1813.

cordierite/mullite shelves

Mullite is combined with cordierite in small amounts to increase strength and reduce the amount of expansion. It does this through the formation of needle shapes that interlock and resist thermal shock. It also provides mechanical strength.

Mullite was first described in 1924 and named for an occurrence on the Isle of MullScotland, although it occurs elsewhere, usually in conjunction with volcanic deposits.   

Pizza Stones and Tiles
Pizza stones are a variant of baking stones where the food is placed on (sometimes heated) stones.  Baking stones are a variation on hot stone cooking, one of the oldest cooking techniques. The stones are normally unglazed tiles of varying thicknesses.  What is said of pizza stones also applies to tiles.


Pizza stones  

Ceramic tiles and pizza stones are essentially the same things.  Some tiles may be thinner, especially if they are not large. In both cases, the ceramic is a poor heat conductor and the thermal mass means care needs to be taken in rapid heating and cooling of tiles and of baking stones. These are dry pressed which give a coarser surface texture than cast shelves.  All these ceramics are generally fired at about 1100C, so they can withstand kiln forming temperatures.  They are adequate as small shelves, but will deform over larger areas over time.

Cordierite-Mullite kiln shelves and furniture.

This formulation of materials has an extremely low coefficient of thermal expansion that explains the outstanding thermal shock resistance of these kiln furniture materials. They are also strong although heavy. Cordierite/mullite shelves are sintered, to allow the mullite needles to form, and fired at 1400C+, higher than tiles (which are most often fired at about 1100C).

This material can be cast, dry pressed or extruded. 

Cast shelves are the cheapest of the methods and provides a smooth surface.  These are used for kilnforming glass, and low temperature ceramic firing. 

Dry pressed shelves have a higher temperature resistance than cast. For this reason, these are often marketed as ceramic shelves, even though the cast shelves are fine for smaller areas.  These are more expensive than the cast shelves.

Corelite, a brand name for extruded shelves with hollow channels, is often used where larger shelves are required, as the weight is less than the solid cordierite. Extruded shelves are ground smooth after forming.

pizza stones


Pizza Stones and Tiles
Due to the thermal mass of pizza stones and the material's property as a poor heat conductor, care must be taken when firing.  Firing quickly can break the stone or tile.  The stone or tile should be fired slowly to just under the boiling point and soaked for a couple of hours to eliminate any dampness in the material.  This probably should be done each time kiln wash is applied.  Because it is porous, a baking stone or tile will absorb any liquid applied, including detergent. They should be cleaned with a dry brush and then plain water if further cleaning is necessary.

Pizza stones and tiles should be checked for having straight and level surfaces. It is not a priority for these to have flat surfaces as for glass and ceramics shelves.  If by placing a straight edge on the surface you can see slivers of light, the shelf needs to be smoothed.  You can do this by grinding two of the proposed shelves together with a bit of coarse grit between.  This best done wet to avoid the dust getting into the air.

Cordierite/mullite shelves do not need this level of preparation, unless they have been stored outside.  It is possible to kiln wash and air dry for a few hours before placing glass on the shelf and firing.  This difference is the low rate of expansion (CoLE 19, if you are interested).

corelite shelves

The extruded corelite shelves are made with cordierite/mullite, but are more delicate due to the hollow channels along their length.  They should be fired slowly to just under the boiling point of water to eliminate the moisture.  It should be fired to 540C with a pause before going to the top temperature.  The shelf should be supported at 30cm intervals under the shelf to minimise breakage.  The whole surface of the shelf should be filled rather than having just one heavy piece; again this is to minimise breakage.

Wednesday, 21 February 2018

Flat shelves

Can I use a pizza stone or a tile for the shelf?

Yes. but, you need to be consider how flat the stones are.

Choose the flattest, smoothest stones you can find.  Take a ruler or other straight edge with you to select the flattest.  Hold the straight edge vertically, and look for light coming from between the edge and the surface of the stone.  Choose the ones with the least light showing.

Determining how flat the stones are
You can make the stones very flat and smooth when you get them to your studio.  Put the surfaces together face to face and move one against the other in a circular motion.  After minute or so of grinding, lift and take note of the areas which are showing the effects of the grinding. Where the stone has not been affected, are the low spots.  The number and depth of the low spots will determine whether you wish to continue to even out the variations in the surfaces.

You can speed the grinding by putting a slurry of grit between the two surfaces.  You can use a coarse grit of 100 or less in the grinding. Place a small pile of the grit and make a depression in which to put the water.  Mix into a runny paste.  And place the other stone on top and begin to move the upper stone in multiple directions.

Keeping the grinding surfaces damp will prevent any dust from the grinding getting into the air. You will hear a difference in sound when the slurry begins to dry out.  This is the time to add a spritz of water to the grinding materials.  As you check from time to time, you will see the areas that already are ground and those that are not yet evened.  The grit will remain in the depressions and be clear from the higher areas.  Push the grit onto those clear areas to continue the smoothing and flattening process.  Continue until the surfaces of both stones are smooth and flat.  This probably will not take much more than a quarter of an hour.

It is advisable when smoothing ceramic or glass materials to wear a dust mask. The dust from both are irritants, although not carcinogenetic.

When the stones are smooth, they need to be carefully dried.  If you have the time, you can leave them to air dry for a few days.  Even then you need to fire them to just below the boiling point of water and soak there for several hours.  Keep the vents open, or the door/lid propped open slightly.


It will continue to be important to fire up slowly to keep the stone from breaking from thermal shock.  The most rapid expansion of the ceramic is in the 200⁰C to 250⁰C range. This means that the rate of advance of firings should be slow until 250⁰C has been passed, no matter what the glass might survive.

Wednesday, 14 February 2018

Drapes over cylinders

Draping glass over cylinders or similar shapes presents some ordinary problems in a problematic combination.
  • ·        In general, the glass is a long rectangle
  • ·        The glass is supported on a long thin part of the mould
  • ·        The glass is usually high in the kiln
  • ·        The mould is heated unevenly
  • ·        The material of the mould influences the way the glass is heated
  • ·        The characteristics of the glass interacting with the mould material

Narrow glass
Especially in smaller kilns, a long rectangle will receive uneven heat.  The short edges of the glass are nearer the sides of the kiln than the long edges are.  This means that the ends nearest the sides are in relatively cooler parts of the kiln in a top fired kiln.  It is the opposite in a side fired one.

Long thin support
A drape on a cylindrical mould means the glass is supported on only a long thin part of its substance.  This further increases the temperature differential in the glass.   The unsupported glass receives both radiant heat and heat transmitted through the air, allowing the unsupported glass to heat faster than where the glass is in contact with the mould.

Elevated glass
Glass high in the kiln – the effect of placing glass on top of a cylindrical mould – heats more unevenly than on the shelf. 

Uneven mould heating
The mould directly under the glass will be shaded from radiant heat, but will continue to be heated by convection of along the lower sides.

Mould material

The two common mould materials are steel and ceramic.  These gain heat at different rates.  The steel generally heats more quickly. The ceramic is usually thicker, so with a greater mass, and the heat transfers more slowly through the ceramic than an equivalent mass of steel.

Glass characteristics
Glass is a good insulator of heat.  This means that heat transfers to the mould supporting the glass more slowly than through the air.

The question becomes how to overcome or at least alleviate these limitations.

Relatively narrow glass sheets that extend near side elements will heat those narrow edges more quickly than on the long sides.  Top fired kilns often have the opposite problem, as the short sides may be in the cooler part of the kiln. The usual solution is to reduce the rate of advance, or to baffle the hot parts.  Either of these should work well in this circumstance.

The long thin support of the glass creates the problem of a heating differential.  The glass may be in contact with half a centimetre of the mould all along its length. The glass and mould heat at different rates.  The normal solution to this is to slow the rate of advance.  The slower rate of advance can be combined with periodic soaks 100⁰C intervals.

Elevated glass
Glass high in the kiln needs special care, as the heat is more uneven there than most parts of the kiln on the heat up.  A general rule of thumb is that the radiant surface temperature given by the elements evens out at a distance from the elements.  This distance is determined by the distance between the elements.  The radiant temperature evens at a distance that is one half the distance between the elements.  If your elements are 100mm apart, the radiant temperature will only be even 50mm below the element.  Any glass closer than this will require slow schedules to overcome this uneven heating.

Uneven mould heating
As described earlier the mould will be heated by convection current of the hot air, rather than directly the radiant heat from the elements.  To reduce this difference, the rate of advance needs to be slow.

Mould materials
Although there are other materials, steel and ceramic are the most common materials from which moulds are made. Steel gains heat much more quickly than ceramic.  In the forms used for glass draping, ceramic has much more mass to heat than steel.  Steel also transmits the heat more quickly.  This means that a steel mould can give a hot line under the glass, and ceramic a cool line.  Reduction in the rate of advance will assist in overcoming this differential heating.

Experience has shown that a very slow rate of advance to a soak of 20 minutes at 100⁰C will allow the temperature to equalise between the glass and mould.  However, too fast a rise after that will cause thermal shock possibilities.  So, increase the rate of heating by 50% to another 20 minute soak at 300⁰C.  Follow this by a rate twice the initial rate to 500⁰C for another 20 minutes as a precaution.  Then proceed to fire at a normal rate.

These precautions are not necessary on the annealing cool as the glass will be in contact with the mould.

Glass characteristics
Glass is a good insulator, so the heat passing to the mould will be less than through the air.  With steel, this will give a hot line and with ceramics a cool line.  Slowing the rate of advance will help reduce this differential.  Experience has shown that placing a sheet of 1mm fibre paper over the mould will also help to reduce the effect of the temperature differences.  You can place a sheet of Thinfire or Papyrus over the fibre paper to retain as smooth a surface as possible.

The best defence to the thermal shock of glass on a cylindrical mould is to reduce the rate of advance with periodic soaks to equalise the temperature.  The addition of fibre paper to the cylinder is an added protection against uneven heating from a hot or cold spot on the mould.

But why does the glass break at right angles to the length of the mould?

I have talked of the long thin contact line between the mould and the glass. “Why does the glass not break along the length of the glass?” I hear you ask.

In thermal shock, the break will occur on the line of least resistance.  In these cases that is on the short sides.

Sunday, 11 February 2018

Glass Cutting Surfaces

There are several considerations about your surface for cutting glass.

Make sure you are putting the glass on a flat surface. If the surface is uneven, it will give difficulties in scoring and breaking.  This means that large sheet timber is an excellent surface.  These boards need to be securely screwed down to the bench structure to avoid any warping.

There is some advantage to having a slightly cushioned cutting surface. This will help accommodate glass with a lot of texture and those sheets that have slight curves in them. 

In this example the user has placed corrugated cardboard under the glass for cushioning, but with a hard surface underneath

Consider ease of cleaning.  As you score and break glass, small shards will be left on the cutting surface.  The tell-tale squeaks as you move the glass indicate there is other glass under the sheet. These shards and any other small almost invisible things under your glass can promote unwanted breaks. Also, if there is glass or other grit on the surface, it may scratch the glass. So make sure you brush the cutting surface clean frequently.

An example of a ready made cutting bench.  It has the advantage of being easy to clean and compact when not in use. 

Think about the size of sheets you will be cutting.  Large sheets often have minor imperfections in texture, or some bowing.  These benefit from a slightly cushioned surface. It also allows the sheets to be put down onto the surface with more confidence that it will not break in contact with the bench top.  But if you are cutting mostly smaller sheets, they benefit from a smooth hard surface to support the whole of the sheet especially when cutting long thin or curved pieces.

An example of a large cutting bench with composition board top surface

Some of the materials used are sheet boards (such as marine plywood, MDF, and other composition boards), short pile carpets,  thin rubber or foam sheets, dining table protectors and pin boards. 

All these are useful for cutting each with advantages and disadvantages.
  • Carpets and foam can trap shards of glass, so have to be cleaned very carefully to avoid retaining sharp glass within the pile or foam.
  • Smooth, wipe-able surfaces avoid trapping glass, but can be slippery. Choose one with a non-slip surface.
  • A slightly cushioned surface is good for large sheets
  • Smaller sheets of glass are best cut on smooth hard surfaces, providing support for all of the glass sheet.
You can also consider, as in the example above, the use of different cutting surfaces on top of the larger smooth and hard surface.  This allows adaptation to the needs of your glass without duplicating surfaces.

Before scoring, clean the glass on both sides, to ensure any sounds you hear when moving the glass relates to glass shrds on the bench rather than grit on the glass.  At the very least, clean along the cut line, as this makes the action of the cutter smoother. The grit on the glass actually interrupts the action of the wheel, so you get a staccato effect in the score line.

Saturday, 10 February 2018

Soldering Fumes

Exhausting Soldering Fumes

Health and Safety
The health and safety of working with lead and solder are a great concern of many people.  Greg Rawls, the acknowledged expert in glass working health and safety, puts soldering and lead work in perspective.

Soldering lead came for stained glass does not usually present an inhalation hazard if the area is well ventilated and you are using an iron and not a torch. With normal soldering, you are melting the lead at temperatures that are NOT hot enough to create a fume.
Lead fume is the inhalation exposure issue. Fumes are very small respirable particulates that are made with heat. Liquid chemicals give off vapours.
Avoid exposure by ventilating the area when soldering, especially if using a torch instead of an iron. Open a window and turn on a fan!  Wash your hands thoroughly when finished working with lead. There are specific products for this purpose.
Use a P95 or P100 respirator when concerned about lead exposure.

There are commercially made fume traps which often have an activated charcoal filter and can be effective.  A simple desk top fan blowing away from you can be effective in well ventilated areas, if you are working on your own. (otherwise it blows the smoke toward others.)

An example of a fan drawing fumes away from the person soldering

Making a fan
Exhausting fumes while soldering is a safety issue. If you happen to have an outdoor screened-in studio a simple fix can be had with a computer fan! You can scavenge such a fan from an older used computer ready for disposal. Simply cut four timbers 50mm square or 25mm x 100mm to fit around it as a box. Attach a long electrical cord to it with an approved plug. Attach a screen to both sides. Plug in. An additional feature is to attach an activated charcoal filter (as used for cooker hoods) to the front of the fan. This removes particles and some fumes.

Always set a fan to draw fumes away you, generally pointing it so that it is blowing the fumes in the same direction as the larger air flow in the studio. A very large fan doesn't always do the job alone, since the fumes seem to rise and find your nose. However, with an additional small fan sitting right next to where you are currently soldering, the fumes just move away.

Wearing an appropriate dust mask as illustrated by the Bohem Stained Glass Studio is the best solution.

Wednesday, 7 February 2018

Needling in Bottle Moulds

Sometimes people experience sharp, needle-like points on the bottle after it is slumped.


As the bottle expands and softens, it conforms to the surface of the mould.  When the cooling begins, some parts of the glass are trapped in the tiny pits of unevenness that always exist in the mould or in the separator.  As the glass retreats, the glass is stretched until it breaks off, leaving the sharp needles.


Remedies relate to separators and temperatures.  This of course, assumes you already have a good coating of kiln wash or similar separator on the mould.


These additional separators can be fibre paper or powders.  Thinfire laid on the bottom of a bottle mould can provide additional separation between the bottle and the glass.  This works, because with a slow rate of advance, the Thinfire will have turned to powder before the bottle begins to slump. This powder will not interfere with any designs on the mould.  Papyros will work on smooth moulds, but not so well with textured bottle moulds, because of its more fibrous nature.

This use of powered paper indicates that you could use a cheaper solution.  Just dust a fine film of kiln wash on the mould.  I do this by placing the powdered kiln wash in a sock and shake the sock above the mould.  This will allow an almost invisible layer of fine powder to fall onto the mould.  This is enough to provide an additional layer of separation between the glass and the mould.


It is quite common for people to slump bottles at tack fusing temperatures to do both the flattening and the slumping at one firing. This is quite hard on the mould and softens the glass enough to promote the needling. 

It may be better to use two firings – one to flatten using tack fusing temperatures, and one to form the bottle at slumping temperature.  This lower temperature will avoid the needling, as the bottle will not soften enough to form the needles during the slumping. The reason many people avoid this is because the bottles tend to devitrify on second firings.  If you do this two-stage slumping, you will need to apply a devitrification solution to the upper surface of the flat bottle to try to prevent it.

You can take a different solution to the two-stage firing.  As lower temperatures reduce the possibility of needling, you can simply soak for a longer time at the slumping temperature than a normal one stage tack and slump.  You will need to peek in at intervals to determine when the slump is finished, of course.  After a few firings though, you will get a good idea of the amount of time required to complete the slump. An additional advantage is that at the lower temperatures, devitrification is less likely.

Tuesday, 6 February 2018

Foiling Tight Curves

Foiling tight curves often leads to splits in the foil.  This note gives information on how to generally avoid the splits and a repair method when they do occur.

Make sure the piece of glass is free of dust and oil and that there are no sharp ledges. Find the flattest part of the glass piece to begin the foiling.

To get the foil to stick to the curve and fold over onto the glass with the minimum of tears, use your fingers first to burnish the sides of the foil onto the piece of glass beginning on the outside curve.

As you move to the inside curved portion gently and progressively ease the edge of the foil with your finger toward the surface of the glass. Easing and burnishing one side of the curve at a time will give you better results toward getting the foil to stretch onto the glass without tearing the foil.

Finish by burnishing the foil down using a fid. This helps keep the foil firmly adhered to glass through the soldering.

Dealing with splits
Tears in the foil line happen. Clean up the broken foil lines with a craft knife and the solder line will look nicer when project is finished. You can also patch the tears by placing a small section of foil over the broken foil.  Place the scrap on one side, burnish it, and fold it onto the edge.  Ease the foil onto the second surface and burnish all the surfaces again.  Then trim the scraps that extend beyond the rest of the foil with a craft knife.

Monday, 5 February 2018

Foiling and Soldering Small Pieces

There are several approaches to dealing with small pieces in copperfoiling:

No-foil approach
One approach is to have some of the pieces held in place by over-beaded solder without foil on the tiny piece, but it is patchy at best and likely to lose pieces in the long term.

Bevel approach
A very good and strong approach is to partially 'bevel' the edges of each piece on both faces. Grind at 45 degrees until the very edge is only 1 mm thick. Then use foil that is 4 mm wide for 3mm thick glass. For 4 mm glass, you will use 5.4 mm foil. Make sure that the foil covers only the bevelled edges and does not extend outside them.

Solder into the 'V' formed by the bevelled edges. Don't over-fill the joints as you don't want solder outside the 'V'. It also is best if the panel is supported underneath the area being soldered by a wet sponge to more quickly cool the solder.

With the solder contained by the 'V', the solder lines will be of constant width throughout the piece. Best to practice this technique on some scraps before you start the main job.

This approach will minimise the amount of light blocked by the foil - important with tiny pieces - while still providing the strength of fully foiled pieces.

Triming approach
If you have to have really small pieces, just foil them as you would any other piece, and burnish it as normal. Then take a very sharp craft knife (Exacto or similar) and trim the foil so that just a little tiny bit of foil is on the front and back of the piece.

No glass approach
Tiny pieces are really tedious to work with. So if the piece is going to be black or really dark, for example a small hummingbird's beak, or a bird’s eye, don't bother with glass but just fill the space with foil and solder.

Lead free Solder

There are some problems to overcome when using lead free solders. 

One is that all, except for expensive compositions, lead-free solders have a higher melting temperature than tin/lead compositions.  The table in this link shows the melting temperatures.

Most lead-free solders have a wide pasty range, so careful attention needs to be paid when selecting the composition, if you want a eutectic, or nearly so, solder.

Some eutectic solders are:

65% tin, 25% silver with a eutectic temperature of 233C.  It is known as “Alloy J” and patented by Motorolla.

99.3% tin, 0.7% copper has a eutectic temperature of 227C. It is expensive.

96.5% tin, 3.5% silver has a eutectic temperature of 221C.  This is slightly lower than the tin/copper composition but more expensive.  It is also likely to rob copper from the soldering bit, although it is easier to solder with as it has excellent wetting properties.

Lower eutectic temperature solders are available:

91% tin, 9% zinc has a eutectic temperature of 199C.  It corrodes easily and has a high level of dross.  This makes it a poor choice for copper foil work.

42% tin, 58% bismuth has a low eutectic temperature of 138C.  It is a well-established solder, but it is expensive.

48% tin, 52% indium has the lowest eutectic temperature of 118C, but it is very expensive.

Copper bearing solders

Another problem is that a solder without lead, robs copper from the soldering bit/tip, and even more so at the higher temperatures lead-free solders normally require.  One means of avoiding the rapid deterioration of the soldering bit is to use solder with a small amount of copper included in the composition. As little as 0.5% can be useful.  Normally, nothing greater than 1% is required to extend the life of the soldering bit.

Eutectic copper bearing solder
However, only one of the commonly available solders is eutectic. This is 99.3% tin and 0.7% copper with a melting temperature of 227C.

Copper bearing solders and pasty ranges
Other copper bearing solders are available. Most of them have high temperatures and wide pasty ranges making them less useful for copper foil work.
Near eutectic solders
97.25% tin, 2% Silver, 0.75% copper has a small pasty range of 217C – 219C, making it a nearly eutectic solder and suitable for copper foil, except for its high melting temperature.

91.8% tin, 3.2% Silver, 0.5% copper has a pasty range of 217 – 218C, also making it a near eutectic solder suitable for copper foil; again, except for its high melting temperature.  With its high silver content, the solder is expensive.

95.5% tin, 3.8% silver, 0.7% copper has a pasty range of 217-220C.  This also has a small pasty range, but may be similar in cost to the 91.8% tin composition.

95.5% tin, 4% silver, 0.5% copper has a pasty range of 217 – 225C.

95.5% tin, 4% silver, 1% copper has a smaller pasty range of 217 – 220C, but may be more expensive.

Other copper bearing solders 
94.6% tin, 4.7% silver, 1.7% copper has a wide pasty range of 217 – 244C.

96.2% tin, 2.5% silver, 0.8% copper, 0.5% antimony has a
smaller pasty range of 217 – 225C and may be slightly cheaper because of the reduced silver content.
95.5% tin, 4% Copper, 0.5% Silver has a pasty range of 217 – 350C and is the usual lead-free plumbing solder.  The high melting temperature of 350C makes it unsuitable for most copper foil applications.

97% tin, 0.2% silver, 2% copper, 0.8% antimony has a high melting temperature and wide pasty range of 287 – 318C., which makes it unsuitable for copper foil.  It is known as “Aquabond”. 

95.5% tin, 4% silver, 0.5% copper has a pasty range of 217 – 225C.

95.5% tin, 4% silver, 1% copper has a smaller pasty range of 217 – 220C, but may be more expensive.

94.6% tin, 4.7% silver, 1.7% copper has a wide pasty range of 217 – 244C.

96.2% tin, 2.5% silver, 0.8% copper, 0.5% antimony has a
smaller pasty range of 217 – 225C and may be slightly cheaper because of the reduced silver content.

Lower temperature copper bearing solders
94.25% tin, 2% silver, 3% bismuth, 0.75% copper has a pasty range of 205 – 217 which is smaller than many of the other copper bearing solders.

90.7% tin, 3.5% silver, 5% bismuth, 0.7% copper, with a pasty range of 198 – 213C, has a lower melting point than many other copper bearing solders.

93.4% tin, 2% silver, 4% bismuth, 0.5% copper, 0.1% germanium has a relatively small pasty range of 202 – 217C, but because of the incorporation of rare earth metals may be expensive.