Last week I got an email notice of a new podcast that promised to “Expand Your Mind and Explode Outworn Beliefs!”

It was a podcast of an interview with Henry Halem in which the podcaster promised that Halem would reveal “…the truth about COE and its limitations in calculating compatibility!”

Of course we are delighted to learn of this brilliant new information since Dan’s been yammering about it for the last ten years. If you too have been living under a rock all this time, you can read the original under the title “TechNotes 3: Compatibility of Glasses on the Education page of this website.

But don’t let that stop you from buying Henry’s book – it’s got lots of other good stuff in it. Some of it is even his.

Seriously, we love the guy. Join us next summer when GAS gives him the Lifetime Achievement Award. And we buy yet another copy of Glass Notes.

What’s the hold up? It’s a color formulation problem. Ed just can’t become anything other than a red, yellow or orange glass – colors that are typically made with cadium/selenium oxides. Those of you who know a little glass chemistry will recognize them as the most ornery, irascible and unpredictable colors in any glassmaker’s palette.

Dignified? What’s dignified about sitting around in a tin can waiting for a glass chemist to get his act together?!

Cadmium/seleniums are so ornery, in fact, that I was pretty surprised when Uroboros’s Bill Ward recently appeared on the Warm Glass bulletin board to state that “…our 90 [COE glass] was designed to be compatible with Bullseye Tested Compatible and these have been comingled successful by many artists since 1989.”[sic]

Checking out the Uroboros website, I saw the same claim “Products at the 90 COE expansion point are tested compatible with Bullseye glass”

Over the years I’d heard about problems with mixing this so-called “90 COE” glass with ours, but I’d never seen any formal testing to investigate the issue. Bill’s remark got me into gear. (Or maybe it was Ed’s peevish spirit).

Either way, Dan drove on over to our local glass retailer and picked up samples of all the Uroboros red, yellow and orange glasses that he could find on the shelves.

TRIPLE FIRES: Measuring Irascible Glass

Because cad/sels have a tendency to change their internal composition on multiple firings – as might happen with a color-bar/full fuse/slump series of firings – we expanded our testing of these glasses years ago to include what we call Triple Fires.

In a Triple Fire test the same sample of glass is fired three consecutive times to a full fuse temperature. After each firing, the test sample is cooled to room temperature and viewed for stress at the interface between the chip and the standardized base glass. If, after any one of the three firings, the chip shows stress greater than our accepted deviation of 2 degrees low to 4 degrees high, it is considered incompatible with that standard and its related glass will not be labeled Bullseye Compatible.

After we instituted Triple Fire Testing both Uroboros and Spectrum glass factories followed suit with claims of “three meticulous test firings…measured for ….COE-shift”.

But – as I’ve been ranting all these weeks – unless a glassmaker explains exactly what their testing protocol is, I’d consider the claims to be just…well…marketing.

READING THE TRIPLE FIRE TEST

I hope you’re not totally bored after all these weeks of looking at fuse test bars. Because the ones coming up here are instructive.

Of the six Uroboros samples that we purchased randomly from our local glass store, four failed our triple fire testing.

This is what they looked like through the polarimeter.

First, the good news (for users):

UB’s 60-606.90 Grenadine Red Transparent passed the test. Against Bullseye’s base clear standard it showed to be within the acceptable range with 1/2 degree high on the first firing, 1 degree high on the second and 1/2 degree high again on the third.

The 60-357-90 Lemon Chiffon Yellow Transparent also passed with 1/2 degree high of measurable stress on the first firing, 1.5 degrees high on the second, and 1.5 degrees high again on the third.

Then our cadmium/selenium friends started to unravel.

The 60-608-90 Tangerine Transparent was just outside our standard range on the first firing at 5 degrees high. Then it went further out to 6 degrees high. On the third firing it went back just slightly lower at 4 degrees high.

Note that most glasses, when their chemistry alters due to more firings, will trend consistently in one direction – either going increasingly higher or increasingly lower. A glass that behaves erratically like this one usually indicates inhomogeneity within the piece of glass. But without access to the full run of glass from which this sample was taken, it’s really not possible to say for certain why it behaved this way.

The 60-356-90 Lemon Yellow Opal was compatible to our standards on the first firing at 1 degree high. On the second firing it went over the limit at 7 degrees high. Then it shot even further out to 11 high on the third firing.

The 60-351-90 Marigold Opal was a loss right out of the starting gate: 6 high, then 10 high, then 15 high.

The 60-604-90 Red Opal was the biggest problem in the bunch. It came closer to matching a “96” glass than a “90” with headlights flaring at 10 high, 15 high and finally 20 high.

Of course I’m very curious to know what Uroboros uses as their testing protocol, but since Bill wouldn’t answer the question at Warm Glass, we’ve had to try to figure it out for ourselves.

As I stated on Warm Glass, I’d certainly like a dialogue between factories that would help users really understand what compatibility in our field means. But I like my discussions in public. Until we can have them here, I guess I’ll just drone on alone.

Meanwhile, we’ve left Ed to muse on the prospect of his precarious afterlife as a cadmium/selenium jewel.

In the end, a stable cad/sel is a delicate work of art, not to be taken lightly or glossed over with vacant claims of “COE 90″ or “…the industry’s most rigorous testing”. Sometimes compatibility just isn’t so easy. But neither was Ed – and we loved him for it.

When will she quit??!?? Soon. Soon. Meanwhile, here’s a laboratory test glass with a measured COE of 89 (sorry, the 90 COE lab glass that we’d tested previously has been discontinued). According to popular understanding this glass should be within the range of expansion mismatch for a COE 90 glass (plus or minus 1 COE point). Not.

I’ve been ranting for weeks now about the misunderstanding of the Coefficient of Expansion within the studio glass community. Steve from Glasgow made the brilliant if obvious point in a comment to my April 11 posting that we’ll never get rid of this use of the COE as a shorthand equation for compatibility until it’s replaced with something else.

Steve’s right. Of course. We all know – and increasingly other manufacturer’s are even admitting – that matching COEs doesn’t insure compatibility between glasses. They further admit that they are testing for the stress at the interface of different glasses, NOT the COE. They are even confirming our assertion that viscosity is equally (we say more) important in the compatibility equation.

(Incidentally, I almost got an inter-factory discussion going over on Warm Glass, when Uroboros’s Bill Ward showed up briefly. I’d love to keep talking, but I guess my chit-chatting skills need honing…)

So Steve asks – correctly – if we shouldn’t be talking about the COE, then how do we talk about relative viscosities?

In short, matching viscosities is about matching annealing points. But that’s, again, taking us off-course. You can find incompatible glasses with the same annealing points, just as you can find incompatible glasses with the same COEs. It is a combination of those two qualities that will determine compatibility. They cannot and should not be separated. The combination is complex and is tied to the unique composition of the base Test Glass.

What is simple is that over 25 years ago, Bullseye developed a glass to use as a standard in testing its range of compatible glasses. It’s our base clear glass. It has a COE of approximately 91* and a measured annealing point of 990F. It will not match all other glasses with COEs in that range, nor will it match all other glasses with similar annealing points. To know whether a glass is “Bullseye Compatible”, you can either buy it from the factory where it’s been tested to our standards, or you can test it against our standard clear glass.

Once more. It is what it is. It says what it means. It’s not 90. It’s not anti-social. It’s merely accurate. Is this a disturbing concept?

Other factories can test against this standard too. In fact, when another Portland glass factory, Uroboros Glass, entered the fusing market, their owner stated that they’d tested their glass against our standard clear glass and that theirs fit ours. So why didn’t they just market it as “tested against Bullseye”? Guess.

Instead they chose to call it “COE 90.0” glass. Clearly, that designation let them come to the party without a hostess gift. Did that bother us? Of course, but not everyone is born with the finer social graces. And UB certainly wasn’t the first to try to genericize a brand. I’m not objecting to generics. I’m objecting to messing up technical information while creating them.

Today we have “System 96” which is a trademark registered to Spectrum Glass Company – who also happens to own the lion’s share of this product range. They are certainly aware that dilatometric tests do not yield consistent readings of 96 on samples of their glass, nor on samples of glasses said to match their glass.

Then we have the commonly used generic labeling of “COE 90” by dealers, manufacturers (not us!) and users that includes the historic standard (Bullseye’s 1101F clear glass), plus a lot of other stuff that may or may not have been tested against that standard.

And – to further confuse the scene – we have distributors who are marketing certain glasses as COE 90 when that claim contradicts the COE published by the manufacturer. And we have websites titled COE90 that – because they sell our glass – are mistakenly assumed to be owned by Bullseye.

And people think that WE are complicating this issue by arguing against the use of COE as a measure of compatibility?!

If not 90 and 96, then what?

“Bullseye-compatible” and “Spectrum-compatible”. Is that so difficult?

Any manufacturer who wants to claim compatibility with those brands should just say what they are doing: that they are testing their glass against Bullseye’s standard or testing it against Spectrum’s standard. Shorthand? How about TABS and TASS?

And if you think that’s too “commercial”, then just keep slaughtering glass science.

- Lani

* Because the measurement of the COE or Linear Expansion Coefficient is not precise, even in the best laboratories, we hesitate to claim better than an approximate measure for this quality.

PS. This is for Ed’s fans who didn’t want to remember him in a bag: Ed in a Tin. He seems to be enjoying the windowsill, but that CHINTZ would make him just….how do I say this….roll over in his….?

We did.

GETTING OUT FROM UNDER SOME SERIOUSLY OLD INFORMATION. Written almost 25 years ago by my partner Dan and his then-partner Boyce, Glass Fusing Book One was the first – and is still the most definitive – book ever written on the subject of kilnforming. Today, even Annie is too smart not to dig out from under that old story.

Yes, this entire mess started at Bullseye. We made a mistake. We (actually, it was Dan Schwoerer and Boyce Lundstrom. I wasn’t here at the time, Your Honor) believed that matching the LEC would insure that glasses “fit” when fired together.

They even went so far as to write “The range of fusing compatibility is plus or minus 1. Therefore, Bullseye #101F [now #1101F] with a coefficient of 90 would be fusing compatible with other glasses whose coefficients range from 89 to 91. Outside this range, undue stresses will develop in the glass.”

They soon learned that this was only true if the base formulas of the glasses being matched were the same. It was NOT true when mixing different glasses like opals and transparents. They learned, in fact, that sometimes the COE had to be different for the glasses to match.

How did they learn that lesson? The hard way, of course.

They’d been having a horrific run of a blue and purple streaky glass, a combination of an opal and a transparent. It was breaking up as it came out of the annealing lehr. The breaks were obvious incompatibility – along the interfaces of the two colors.

2105 – The glass that taught the Bullseye guys that matching COEs does NOT insure compatibility.

Samples of the blue opal and the purple transparent were sent off to a testing lab to check their expansions. The report showed that the LEC of the opal was slightly lower than that of the transparent. Problem solved: the formulas were adjusted in order to match their expansions.

Result? The breakage got WORSE. The COEs matched perfectly, but the two glasses were now even MORE incompatible.

Now what? When in a corner it’s sometimes a good idea to ram it into reverse. They reformulated the opal to make it even LOWER – to increase the differences in the COEs of the two glasses.

It worked. By making sure that the COEs did NOT match, they could insure that the two glasses fit.

But, before anyone gets the idea that the authors of Glass Fusing Book One were idiots, please read the original text. They knew that the COE wasn’t quite so simple. They cautioned that:

“…some glasses that have the same coefficient number, as determined by a laboratory, do not always fit each other when fused together.”

They went on to point out that how a glass behaved in the upper temperature ranges was also important.

But the damage was done. The EZ-users, the salesmen, the Fusing-for-Dummies set just grabbed the COE and made it into the astrological sign of our industry: “What do you use? 90 or 96?” It got worse:

“If you can fuse it to Bullseye and it doesn’t break, it must be 90!” Who cares if the manufacturer says it’s 94? Let’s just label it 90. It’s easier to sell that way.”

“Hey, this company says their glass is even closer to 90 than Bullseye. They say theirs is 90.0”

“Don’t you know that the standard in glassblowing is 96?” “Think of all the incredible combinations you can do by mixing all those color bars with a 96 glass!” “Wow, what FUN.”

A random sampling of chips from color bars by Reichenbach, Zimmerman and Gaffer fired against a base System 96 clear glass. Is this the same “compatibility” as the standard that Bullseye established for kilnforming so many years ago?

In the rush to make it all look so E-Z, users are being led astray from the basics of our field: the understanding of what compatibility means, especially as it relates to kilnforming. They are also being deprived of the opportunity to learn about a truly fascinating material. Glass, and the properties that make it so interesting to work with, can’t be reduced to marketing bytes without the user losing some of the rich potential that it holds for exploration, discovery and true creativity.

In the commercial carnival that has erupted in our field, essential, valuable, and necessary technical information is being sacrificed for sales dollars.

Does anyone care? You tell me.

Ed didn’t. He got pretty tired of living in a house full of obsessive glass fanatics and decided to bag it.

PS. Lest anyone think Ed ended it in a grocery bag, he didn’t. He was still alive – although a bit zoned in this shot – just enjoying one of his favorite activities on the day before his Big Check Out adventure.

I was about to unmask the fools who started the COE mess when I got a private email asking me why I was making such a fuss about compatibility standards when – by our own admission – Bullseye’s are likely tighter than they need to be.

First of all, that wasn’t quite the point of my rant, but I’ll take a momentary detour here to explain why this stuff matters…

When you’re starting out, the misinformation that you learn and the glass that you use may not create problems. Your projects and your equipment will likely be small. You can get away with breaking a lot of rules when you’re making earrings and even plates.

THE TORO. Designed and built at Bullseye: the heating controls, venting, and shelf configuration address the specific needs of large scale work.

When you scale up, what you learned as a beginner and the tolerances of the glass you work with can make the difference between success and costly failure.

This is a huge topic. Some of the major issues are:

• Understanding large kiln design
• The large shelf: materials, sources, maintenance
• Insuring even heating in the kiln chamber
• Firing schedules for larger work
• Problem solving off the charts

I couldn’t begin to cover this stuff in a blog. And lots of kilnformers don’t need it yet. But they DO need a strong foundation and understanding of the basics. I’m trying to wade through that stuff in my blog rant on COE.

For the Big Stuff, I strongly suggest that anyone who is working – or intending to work – on a large scale, come to Portland for BECon 2007 where we’ll be talking about this stuff, and more.

THINK ABOUT IT. Simplicity – when it’s big – isn’t quite so simple.

Now, back to the urban legend of the COE…