Origin of COE as compatibility measure in studio glass

[Tom Littleton]

Henry,
That is not correct. Where did you get your info?

I have to say we got lucky with a short learning curve but we did have Nick's help. The first stuff that got melted at Penland was made to the orginal formula and which had no lithium. As I remember, the stuff was so hard to melt it got cooked at high temp for 24-36 hours perhaps the source of you stories. We were trying a relatively large grain local sand. We tried screening that sand to get a smaller grain but that didn't work. I remember only two people got some of those first runs. We tried changing the source of the sand and added the lithium at the same time and by the way it was the maximum amount suggested. We should have tried changing one thing at a time and used less lithium and then increasing it in a series of tests but that was not done. The result was the 92. We only took out a little soda to make the 87.

[Pete VanderLaan]

Originally posted by Lani McGregor
OK, Pete, now we are into magic: Bullseye wasn't founded until 1974.

well, you are right. I actually included the items from a few other issues that popped up just for effect. The price on the glass was right though. I do have the ad announcing the grand opening of bullseye.
40th anniversary indeed!

[Dave Bross]

Wow! spectacular thread! Do continue!

If anyone has that book that HKL did in the late sixties there is a lot of discussion of what went on with Erwin at the factory as regards melts. I read the book a few years back in the Seattle library so I don't have a copy to reference now.

Tom,

Are there any other reasons besides expense of material that you wished you had used less lithium?

Just spoke with my friend the chemist at the Leerdam plant.
In his memory Kugler colours where used in the Crystal production all the time. Kugler was asked to make a melt to fit but if it did not the clear was adjusted.

It was tested by production doing the ring test.
The lab used the "method Dr. Patmos". Dr. Patmos worked for Philips and developed a test where a standard clear was welded to the glass to test. After annealing the weld was cut and polished and viewed through a micro polariscope to measure the tension.
Necessary for the developments of all the different light systems Philips developed.
This is all pre 1950.
His remark on the COE, LEC and viscosity was very practical, to my friend its all very theoretical if you know the LEC you know the COE and for him viscosity has to do with machine speed. Than again he is all about container glass. Density of 2.5 at room temperature but only 2.25 at 1400 C (2552F)
Batch calculations with expansion is only practical empirical info, measure and you will know.

[Pete VanderLaan]

Originally posted by Robert Mickelsen
glass. For instance, there is an opal green that is notorious for crackling when incased. But the measured LEC is dead-on that of clear boro glass. Why doesn't it fit?
Fascinating discussion folks. Please continue...

When you say it is dead on, who measured it and how. Is it a pull test? I simply do not trust pull tests in opal/clear matches as even remotely reliable. They only work in formulas of similar composition. It would require dilatometry or a ring or trident seal test to get accuracy here. Chrome glasses are refractory in nature and the problem is compounded by opacity. If it is a boro supplier, I have so far not been overwhelmed by the current group of suppliers being terribly thorough. It's a comparatively young industry when it comes to color and actual batching of colors.

Originally posted by Pete VanderLaan
I really think that L.E.C. works so well for glass blowers that the viscosity issues have just about never concerned them.

Pete, IMO what works well for glass blowers is glass blowing. Not LEC. I happen to know three glassblowers who started a glass factory and found out – while trying to fit different colored glasses together in streaky sheets – that after sending in numerous tests to laboratories (late ‘70s) they could only make the glasses match by INCREASING the difference in their COEs (0-300C). Everything they thought they knew about expansion and fitting colored glass had only taken them in the wrong direction.

Where I think the problem lies, is in the gross generalization that such things as "96" guarantee you a clear blue sky and great sailing.

Exactly. COE is half of it. Viscosity is the other half. One without the other is certainly a place to start, but is misleading. Matching COEs only works if the base compositions are the same or very similar. Who can make a broad color palette with only one base composition? Bullseye has at least five different base compositions for its line of colored glass. If we made the expansions the same on each of these they’d be grossly incompatible.

At the time BE developed its Tested Compatible line Boyce (ever the salesman) liked to call it “90 COE” and, IMO, intended to take over the kiln world at that “expansion”. Dan was never completely comfortable with the corruption of glass science entailed in that kind of marketing. Of course they could have called the glass “990 AP” instead of “90 COE” and been equally right – and equally wrong. For accuracy, it would have been more accurate to call it a 90/990 glass.

I’m just curious as to where the COE as a synonym for compatibility first entered the marketplace and hot shop. When did color bars or batch start to be listed/distributed in those terms?

And I’m talking measured (i.e. at 0-300C) not theoretical or calculated COE. Lots of people were doing calculations. But to our knowledge the practical testing done in factories on a daily basis has always measured the “fit” between glasses, i.e, the combination of expansion and viscosity – NOT the COE.

When did SPB first list a measured COE? I’m pretty sure that what Tom and Henry are talking about in terms of Labino’s formula is calculation.

[Robert Mickelsen]

Originally posted by Pete VanderLaan
When you say it is dead on, who measured it and how. Is it a pull test? I simply do not trust pull tests in opal/clear matches as even remotely reliable. They only work in formulas of similar composition. It would require dilatometry or a ring or trident seal test to get accuracy here. Chrome glasses are refractory in nature and the problem is compounded by opacity. If it is a boro supplier, I have so far not been overwhelmed by the current group of suppliers being terribly thorough. It's a comparatively young industry when it comes to color and actual batching of colors.

I know the green was tested on a dilatometer and was on the money. The color manufacturers have tried for years to solve the problem to no avail. You are absolutely correct in what you say about the present group of suppliers.

Originally posted by Pete VanderLaan
When you say it is dead on, who measured it and how. Is it a pull test? I simply do not trust pull tests in opal/clear matches as even remotely reliable. They only work in formulas of similar composition. It would require dilatometry or a ring or trident seal test to get accuracy here.

Just so we're all straight on what's being measured:

a DILATOMETER is used to measure the expansion, e.g. COE(0-300C) or “LEC”. This is the measurement of an isolated glass.

The following three tests measure the combined effect of expansion AND viscosity on two different glasses:

RING TEST

TRIDENT SEAL TEST

BULLSEYE’S CHIP/BAR TEST

The PULL TEST is also a rough measure of the combined effect of expansion and viscosity on two different glasses, BUT in our factory experience it is very frequently unreliable due to on the skill level (or lack thereof) of the person performing the test.

[Pete VanderLaan]

Originally posted by Lani McGregor
terms?

And I’m talking measured (i.e. at 0-300C) not theoretical or calculated COE. Lots of people were doing calculations. But to our knowledge the practical testing done in factories on a daily basis has always measured the “fit” between glasses, i.e, the combination of expansion and viscosity – NOT the COE.
*************************************
When did SPB first list a measured COE? I’m pretty sure that what Tom and Henry are talking about in terms of Labino’s formula is calculation.

Actually you have hit on the first point of confusion. How about 17C-300C instead. When I did the work at Los Alamos labs with this, we started at 0C just like the ASTM standard would have suggested. Nothing made real world sense. Once I shifted to 17C as a start point, things looked right. 17C is of course the temperature at which lab workers are comfortable while wearing white coats. I have to be very careful when I'm measuring four inch rods for expansion that the start temperature is consistent. In summer, it easily goes up to 23C so I need a cooling bath for the canes and dilatometer.

The labino stuff is calculation and it is consistent with the E&T numbers if run up to 250C ( I think). E&T did not originally go to 300C and that's why those numbers have never made real world sense. It seems to me it was either 250C or 225C. I cannot recall right now and I don't want to look it up.

When Croucher and I sat down about eight years back to talk about expansion we both felt that a 93-93.5 would have been a better standard but SP87( which is really a 96) was the elephant in the room. If we wanted to sell color that was not 45 percent lead we were going to target SP87. While Kugler just got around the fit issue by dumping lead into the glasses, we really wanted lower lead contents wherever possible and no lead if it was not critical to the glass. Kugler and company all use GLASMA 70 or 71, I forget which but it is their proprietary batch. It has a predetermined L.E.C which they add colorants to. It explains why Kugler is all over the map. John and I both made the color first, and then adjusted the glass to a 96.

I agree about the multiple formulas. I have six base glasses. I think John had seven the last time we talked. The trouble with the viscosity issue is that I can't explain it. Nick couldn't explain it. Why did lead seemingly defeat the compatibility problem up until you pierced the surface tension of the piece with a saw or a grinder? Would a high Bismuth glass do the same thing ( moot since bismuth is too expensive).

I don't think anyone listed an L.E.C. on their stick until Croucher. I never did. I advertised its expansion but never put it on the rod. I liked the parrots too much.

Maybe glassblowers are more like the tarot card of the fool. Walking towards the cliff and always the presumption that they will turn at the last minute without ever knowing the cliff was there.

[Pete VanderLaan]

Originally posted by Robert Mickelsen
I know the green was tested on a dilatometer and was on the money. The color manufacturers have tried for years to solve the problem to no avail. You are absolutely correct in what you say about the present group of suppliers.

***************************
Is this that green that was a square stick and came from the phillips plant in NY along with the opaque white stick about three years back? Secondly , who did the dilatometry tests? Do you really know firsthand that it was tested? That's what I mean about the suppliers.

Originally posted by Pete VanderLaan
John and I both made the color first, and then adjusted the glass to a 96.

Pete, I’d still argue that you weren’t adjusting to a 96 COE, you were adjusting to fit SP87.

Do all glasses with a 96 COE fit SP87?

And you think viscosity is confusing?

BTW what’s the AP of SP87?

Originally posted by Pete VanderLaan
Why did lead seemingly defeat the compatibility problem up until you pierced the surface tension of the piece with a saw or a grinder?

I don’t understand how the compatibility problem is even “seemingly” defeated if it’s only compatible until you cut into it.

[Steven O'Day]

In the mid 80's Therese LaHai was at CR Loo and wanted to try and figure out the in-compatability problem. She had the base glasses and colors tested and published the COE numbers in the catalog as a rough guide. I don't remember any talk of viscosity. I think at the supplier level before this it was caveat emptor, the numbers used were the ones from the manufacturers or the E & T theoretical.

[Pete VanderLaan]

Originally posted by Lani McGregor
Pete, I’d still argue that you weren’t adjusting to a 96 COE, you were adjusting to fit SP87.

Do all glasses with a 96 COE fit SP87?

And you think viscosity is confusing?

BTW what’s the AP of SP87?



I don’t understand how the compatibility problem is even “seemingly” defeated if it’s only compatible until you cut into it.

*********************
That's why I said that the viscosity issue has never gained any traction. Glassblowers seem to be content with the notion that they can't cut and grind their work made with a lot of german color rods. It's bizarre but people seem very accepting of it. I don't disagree with you. It would seem as long as it doesn't actually explode, it's OK. How many people have you seen who still try to use bright canary yellow opal from Kugler. It doesn't fit anything on the planet. It still sells and if you don't grind it, it holds together for some time if you use cullets that are nearer to a 90.

I have been a vocal minority for a loooong time that the glass made in the 70's and 80's well may be the cullet of this century in due time. It's like pushing back the tide. I am simply stating the way I perceive things to be.

As to all glasses that have 96 C.O.E. fitting Sp87, no I don't think that but I think it based on the thickness of the two glasses that are present. I make opals at 96 that saw and grind with SP87 just fine if they are under one inch thick. Go over that and they don't fit. Why? my hunch is annealing range, not point.

You live in a casting world where thickness is a day to day issue. Things crack. They don't crack in blownware and the schools have absolutely no technical interest in this crap at all anymore. They have all adopted Dale's out of sight out of mind approach to broken glass.

As to the argument about what we were aiming at, I measure my glasses as a 96 on a dilatometer. I do that because I can usually measure SP87 as a 96 or very close. I make the number my target these days. I don't actually have any SP87 in the building right now and haven't for about a year. If SP varies, and it does occasionally, I simply can't try to follow the day to day vaguries of the soap opera. I need to be within 1.5 ten thousandths., that's all.

The annealing RANGE of SP87 is about 890 to 945 depending on thickness and how fast you want to anneal it. I put away large grinding blanks that are thick at 995 for about an hour before dropping to 940. If I don't, I have really cracking issues on grinding them or I have to anneal them at 940 for way to long a time.

When Croucher and I made our presentations at Corning back in 2001, Frank Wooley, senior melt engineer at Corning was there and got into the Q &A very actively. I made the assertion that a substantial number of "Checks" in pieces were really being mistaken for incompatibility when they were actually annealing problems. Frank became very animated in agreement. We talked about it for some time.

[Tom Littleton]

Dave,
Expense is the biggest reason for not wanting to use a relatively large proportion of lithium but it also has a bad rep as a very aggrssive flux. My feeling is though (without actual data to back me up) that by the time you replace the fluxing action of the lithium with something else whatever you come up with will be just as corrosive.

Lani,
I can't remember when I started putting a measured COE in the sales material. I think fairly recently and likely when everyone else started talking about COEs and stuff and started to look at the 87 as a standard. No more than 8-10 years at the most but I'd have to check and I didn't always keep good records of when I made changes. It seems like it became more of an issue when the Dichroics came out and people wanted to use them in their furnace worked glass and the 90 standard stuff would not work with the SP 87. And likely after Pete learned how to build and operate a a dilatometer.

Originally posted by Tom Littleton

Lani,
I can't remember when I started putting a measured COE in the sales material. I think fairly recently and likely when everyone else started talking about COEs and stuff and started to look at the 87 as a standard. No more than 8-10 years at the most but I'd have to check and I didn't always keep good records of when I made changes. It seems like it became more of an issue when the Dichroics came out and people wanted to use them in their furnace worked glass and the 90 standard stuff would not work with the SP 87. And likely after Pete learned how to build and operate a a dilatometer.

Tom, thanks. This is the input I was looking for. COE as a measure of compatibility seems to have showed up in the vocabulary of studio glass very recently (certainly quite a while after BE published Glass Fusing Book One in 1983 and our side of the market – i.e. kilnworkers - grabbed the “shorthand” tag of “90 fits 90” with all its half-accuracies and commercial agenda).

I'm not saying that the batch or color manufacturers are wrong to provide COE info, just that users think this equates to compatibility when any colored glass maker knows it doesn't.

Tom, do you provide an Annealing Point for SP 87? (Not the annealing range – the AP as determined by the ASTM test)? A softening point?

"Kugler and company all use GLASMA 70 or 71, I forget which but it is their proprietary batch. It has a predetermined L.E.C, which they add colorants to. It explains why Kugler is all over the map. John and I both made the colour first, and then adjusted the glass to a 96."
That's Bullshit
Excuse me but Kugler was there decades before Glasma ever was thought of. Kugler started to line up with Glasma just to provide a base glass that would hopefully fit there range of colour glass to all of those studio melters, if people where able to melt it in the first place.
The basics of "Kugler crystal" mixed by Glasma has absolutely nothing to do with the different base glass that Kugler Colour or rather frau Fridrich develops her colour ranges around, other than the two fit most of the time hopefully.

I just wanted to say I've spent the past hour and a half of my work day reading this, its more fascinating that the history channel and the military channel put together (and if you know my love of weaponry and old stuff that says alot). So thanks guys, and yes I'd be interested in researching or helping to organize info if anyone wants to try and put together some kind of documentry archive of it all.

Justin

[Robert Mickelsen]

Originally posted by Pete VanderLaan
***************************
Is this that green that was a square stick and came from the phillips plant in NY along with the opaque white stick about three years back? Secondly , who did the dilatometry tests? Do you really know firsthand that it was tested? That's what I mean about the suppliers.

No, it is an opal green (NS jade) manufactured by Northstar. Paul Trautman did the dilitometer tests in an attempt to discover the reason for the cracking problem. I believe he did them and related the true results to me because there was no reason for him not to. I might be suspicious if I were not such a close friend of his. He divulged the results to me while were discussing what the problem could be if not the LEC. That is when we first discussed the possibility of viscosity curves... which is why bells went off when Lani mentioned it.

I might be wrong... the jade problem may be entirely different from what is being discussed here... but I am still reading this thread egerly every day (instead of working like I should be) anyway.

[Dave Bross]

Thanks Tom.

I'll bet your assumption about whatever flux you use being equally corrosive is probably true. I suspect the amount of alumina in the glass is important here too.

One other thing Pete has (repeatedly) had to remind me about is that our old friend viscosity is THE major player in how fast the glass will eat the pot.

So there's one thing "for sure" about viscosity.

Lani,

While we're in history channel mode, who solved your expansion/viscosity fit issues in the early days and how?

[Pete VanderLaan]

Frank Wooley had said to me that no chemical was particularly more corrosive in a glass than any other but it was the viscosity of the glass at a fixed temperature that determined how quickly a pot was attacked. In that capacity I think you could assume that Lithium is a very powerful flux for lowering viscosity. A little goes a long way.
As to the GLASMA issues, what Durk says goes contrary to what I have been told over the years. I was told that GLASMA became involved when German environmental law blocked the mixing of the base formulas in Germany. At that time the mix moved to GLASMA and much was standardized. If this were not the case, then it would be inexplicable to me that the company would make so damn many different expansion coefficients. It is not hard to standardize them if one can simply adjust the formula but if it's premixed then you have problems. When Klaus sold the company, my impression was the sale of a formula box and little else. The product suffered quality issues for years and produced a sufficient amount of defective product to cause US suppliers to stop dealing with them. As far as I could see at the time under Frau Fredrich, there was an apparant lack of control over process. They have gone on to build a large factory and try a variety of processes such as cast rod which have had genuine problems.

[Pete VanderLaan]

Originally posted by Robert Mickelsen
No, it is an opal green (NS jade) manufactured by Northstar. Paul Trautman did the dilitometer tests in an attempt to discover the reason for the cracking problem. I believe he did them and related the true results to me because there was no reason for him not to. I might be suspicious if I were not such a close friend of his. He divulged the results to me while were discussing what the problem could be if not the LEC. That is when we first discussed the possibility of viscosity curves... which is why bells went off when Lani mentioned it.

I might be wrong... the jade problem may be entirely different from what is being discussed here... but I am still reading this thread egerly every day (instead of working like I should be) anyway.

*************************
Draw me several feet of cane of it, about 4mm in diameter which is around the thickness of a lead in a pencil. Vary it up and down a bit in thickness to be sure, I will test it.

[Steve Stadelman]

Paul has not had Northstar for a while now. Is it an old batch?

[Rollin Karg]

Originally posted by Robert Mickelsen
For instance, there is an opal green that is notorious for crackling when incased.

Robert, if that green is a fluorine opal this may help.
Years ago when we first started making colors a guy told me that the fluorine opals are made of billions of little fluorine inclusions.When you reheat you start to lose these little inclusions to the atmosphere and the more you heat the more you lose. Also the hotter the flame the more inclusions you lose and when you lose these little guys you change the LEC. The same thing happens in the pot when we make them. The longer they stay in the pot the more the LEC changes and most likely the viscosity changes too.
I remember how we were struggling to make the colors work at the time and that bit of news was just one more thing to torture us.
Rollin

Robert,

The jade green was never tested for COE by Paul. It was tested by 2 other people who wanted to know about the glass. Both found the glass to be in the 35 range.

The glass does not contain fluorine and the cracking in the early days (prior to mid 98) was poor formulation. After that date the cracking was closely related to reduction during working or phase seperation issues during extended "garaging" at high temperatures.

And yes, the viscoity curve and rate of heat loss during cooling of the borosilicate glass is very important to fit, maybe more so since the metal oxides tend to much denser in the boro glasses than in the soda-lime.

I was told, by whom I have no idea, that the fluxing power of the alkalis has to do with ion diameter and the ability of the ions to enter and spread the tetrahedral bonds of silica. My copy of Handbook of Chemistry and Physics doesn't have data on ion sizes, but it is almost as old as I am. Do recent editions have such info? Which is larger among sodium, potassium and lithium? Does it make sense with expansion and what we know about fluxing power?