CraftWEB Hot Glass Talk

CraftWEB Hot Glass Talk (
-   General Hot Glass Discussion (
-   -   ring tests (

Shawn Watt 12-10-2016 10:10 AM

ring tests
When making ring tests and looking at all the variables. thickness of the wall ,how much color to clear ratio ,diameter of the ring and the gap itself. Looking at all this, how do you come to the final conclusion if the glass fits or not. I started using a micrometer and feeler gauge to make the measurements and am wondering where to draw the line for compatibility.
Example I am testing a black against clear
OD on the ring is 3.25 in
Thickness is .043 in
black to clear ration is about half with black on the outside
gap is .026 in

i am reading that gap as out of range. what are your thoughts?

Pete VanderLaan 12-10-2016 10:24 AM

It reads the glass on the inside has a gap and is slightly lower expansion than the outside glass. The proportions are very good. The very fact that the ring survived the saw is a solid indicator that the two glasses are reasonably matched. Ideally, you want the scored vertical part to simply sit at rest, neither opening or closing on itself. I would either add a small amount of soda ash to the black, or a small amount of silica to the outer glass. Small for me is about 25 grams in 30 lbs or so. Then melt and measure. One of the two glasses would of necessity have to be a batch glass to add the sodium or silica although it's possible to add soda or silica to SP87. Otherwise, it simply is what it is. At .026 I would try to close the gap.

Good Job.

Shawn Watt 12-10-2016 03:47 PM

Thanks Pete. So when you see a gap either way do you have a measurement that you say "once the gap goes behold this point it is not compatible to use with the clear?" I think i remember reading if a piece of paper fits the gap it tolerable. Is that right? According to the internet a standard piece of paper is .1 mm thick. Should i assume if the gap exceeds the .1 mm that it is out of range. Obviously all the other variables are going to effect that gap.Maybe im looking for a rule that doesn't exist.

Pete VanderLaan 12-10-2016 04:13 PM

my experience tells me that if the thing survives the sawing, It's likely to be compatible. My rings tend to be kind of thick and the ratio of each type of glass not so perfect. BUT if they survive the saw and then I can score it vertically and tap it open, I can see something I can judge. It really will break on the saw right away if it's out by much. A piece pf paper thickness would be a nice fit in my book. I do like to get it to just sit there.
My great advantage is that both types of glass are always batch glasses and that gives a real advantage when you're making the adjustments. I find that particularly true if you are putting a bunch of stuff in your batch that is not factored by E&T. The Third edition of Glassnotes had a huge error in it that listed factors for metallic oxides. That came from Paul Manners back around 1976 where he included a set of factors that were commissioned in the enamelling industry in I think 1925. They were based in a completely different temperature range than the ASTM standard dictates. In the fourth edition, I had pointed out that they should not be in there and Henry removed them.

Appen has some numbers for the colorant metals but you have to use the whole system and then you really want to do mole calculations, not percentages.Running a set of factors for the colorants is something that is still missing. The big glassmakers simply don't care since they never laminate production stuff.

Scott Benefield, who is a far more skilled glassworker than I am was able to do live tests on the blowing floor making thin blown rings. I never had the skills or the assistant it needed.Nice to get fast results though. When you pop them clean, the expansion is really exaggerated and takes some sober breaths to work with.

I do think that the ring coupled with the Hagy seal is a good combo. Learn to make seals. Pull tests can be very deceptive. I remember Ed Skeels doing cane pulls that looked terrible and then he annealed them and they came out absolutely straight.

The real trick is to not kid yourself on fit. When a piece blows up, we all have a weird way of rationalizing it. That's not good. Eventually the physics will have it's way. In this tortured round I've done recently with mating my unoxidized color batch replete with the tin and silver to the clear Tom mixes per my specs, I knew I had issues yet I had pieces that went through grinding and were set aside. Most broke after about six weeks. Some have yet to break. If you sell it, it will break soon. When I'm doing this junk, I still need to write on every piece the date and the batch info.

Trying to make everything perfect through the whole color system will make you crazy.Just remember the A to B to C logic I occasionally refer to. Whenever you try to put fifteen colors in one piece, Buddha's nose is getting quite out of Joint and he calls in Heidi Broderbund and her 15% rule. At that point you need really good muck boots for the deep shit you will find yourself in.

Shawn Watt 12-10-2016 04:33 PM

what do you mean by "live tests" ? Like self annealed cups ?

Pete VanderLaan 12-10-2016 04:40 PM

you're fast. I was simply editing my last post and here you are again. Do read my editing.

What Blenko could pull off which I'm sorry to say I don't have the skills for was taking a color in question- every morning and casing it in the clear or whatever of the day. They would blow that out ( a live test) really thin. Scott could take the thin things and shear out a ring of it hot. As soon as it cooled, it could be scored. We did it that way down at Josh Simpsons when I was working out the formulas for the Corona series. Make an ultra thin goblet cup, cool it a bit, scoreit and pop it open.

I like having the ring so I stick to what I normally do. It's reasonably constant. That ring goes in a bag, stapled to the other results and tied to the formula. I have something to refer to. In my first class they nailed the bags to the sheetrock wall in the hall. We had about fifty of them. It wrecked the wall but everyone could see what was happening. Teaching is quite an experience. I'd be into substance abuse in no time. when you get someone who's serious, it's important.

Scott Benefield 12-11-2016 04:48 AM

The ring tests that they do on Murano involve blowing a long bubble (with a wall thickness of no more than 1/16" and a diameter of around three inches) with the hot color to be tested gathered on top of the base clear glass. It is "cut" into a ring by wrapping a gather of hot glass around the middle of the bubble and popping it through thermal stress; then doing that again to make a ring of two or three inches. The ring is opened by laying another string of hot glass across one face. If done correctly, ring shouldn't have any annealing stress and should be thin and flexible enough to register any expansion mismatch between the glasses (i.e. opening up a gap or closing in on itself to make an overlap).

There are too many variables for that test to yield any quantifiable information--like measuring the gap to determine how much sand/soda to add/subtract--but it's immediately useful as a tool to visualize relative compatibility. Do you go forward with the day's production? Or did the batch person make a bad mix/melt? I'd say it's on the same order of usefulness as a pull test (thread test).

Pete VanderLaan 12-11-2016 08:03 AM

I find the more rigid ring to be more useful than what Scott describes in comparisons. BUT you have to anneal the ring material in my sample.Scott gets immediate results. I would be somewhat suspicious of an unannealed sample as I recall the Ed Skeels pull tests. I still make a Hagy seal if there's any doubt and the doubt usually comes with making new forays in color. The main point is that all the indicators should be pointing in the same direction.

Shawn Watt 12-11-2016 08:24 PM

thanks to both of you for the insight. the immediate results would be helpful. i will do some experiments this week with annealing a test cup and making a self annealed test. and then try what you say about skeels pull test, i will reanneal the ring and see what kind of results i get.This is exciting stuff

Pete VanderLaan 12-12-2016 07:01 AM

Ed's pull was between a fluorine glass and a transparent. I was floored by it but it makes sense if you consider the work done at bullseye. That remains an issue that the ring test and Hagy test actually deal with. If you can saw a cup that's part opaque and part clear. You either have a fit or you are very close. The gap is indeed the indicator.

In a Hagy seal, if you aren't pretty close ( I would say under 2 points) the deal just breaks up, another good indicator of trouble BUT when it breaks, you know nothing more.

Again, they are all indicators and should all point in the same direction. Anyone who simply says "It hasn't broken" is not only kidding themselves, they are unfair to the final client and they are unfair to the history of glass making. Historically, I am having trouble remembering finished pieces with tons of different colors in them as we see today and I think there's a reason for that. There is such a thing as being a conscientious craftsman. Durability is part of that .

Shawn Watt 12-19-2016 10:07 AM

Results are in. I tested the black glass with my clear with a variety of different ways. I did the "live test" where i made a thin cup and sat in down in some vermiculite just so it didnt get hit with a cold breeze or something. I made a cup and put it in the color kiln and just crashed it. and then i made cups and put them in the annealer at the normal cycle.
- Live test results- I made two cups, one with black on the inside and one has black on the outside of the clear. They both want to overlap. both test have similar thickness and ratio and outside diameter. so i figured that there was some sort of strain in the fast cooling. i cut some extra rings off the same cups to reanneal and see what happens.
Kiln crash test- that one was resting perfectly. No gap. I also cut an extra ring of the cup to reanneal.
- Annealed test- those cut and scored perfectly as well. no gap either way.
- The reannealed results- the live test with black on the outside, it originally had an overlap. After annealing the test it has a large gap of about .092 in and the ring with the black on the inside also has a gap but not as severe and is also out of alignment.Meaning they wont fit back together like a puzzle piece. The kiln crash test still was resting with no gap but was also out of alignment. Which made we wonder if i should not have scored them vertically before reannealing? maybe the heat and soak of the annealer warped them from heating unevenly or something?
i will try this all again on the next melt next week but the kiln crash test and the regular annealed cups so far are the only ones in common. Any thoughts or ideas on the process?

Josh Bernbaum 12-19-2016 10:54 AM

If it lets you get a clean cut on the saw then I think you're in a pretty good place, and the score results are just for additional fine-tuning of the fit.

Scott Benefield 12-19-2016 01:27 PM

If you have a simple polariscope, you can get some information from a 'chip test' (similar to the way that fusers test colored sheet glass against a base clear). Get a small 1" square of relatively flat clear glass and fully fuse (flat fuse) a 3/8" chip of the color to be tested on top of it. The polariscope--or just two pieces of polarized lens material set perpendicular to one another on either face of the test--should show any mismatch as colored haloes. But that's about all it tells you.

You don't have to make sheet glass to do this, you can just blow a small thin rondel and do it all in one step.

Shawn Watt 12-20-2016 09:06 AM

Thanks for the input josh and scott. Ill try the rondel thing out. I actually hot pop my test cups because they are so thin the saw just crushes them. im going to try making the cups a little thicker and sawing them and see what i get.

Pete VanderLaan 12-20-2016 09:42 AM

My minimal experience with ultra thin tests is that the gaps are far more exaggerated then they are on a thicker one. My tests are usually around 1/4- 3/8ths inch thick total. I put them in a zip lok with notes when I've done the scoring.

I don't care for the test they use at either bullseye or Uroboros. I would at that point prefer a Hagy seal. It tells you a lot more.

Art Freas 12-20-2016 12:36 PM

Maybe stating the obvious here but to do these tests you need color pots to gather the colors, correct? Also Scott, if you can video that process for the live test I would be interested in seeing it. Thanks much for all the information.

Shawn Watt 12-20-2016 01:21 PM

My cups are 1/16 inch at the most. Im sure they are quite exaggerated. I'm probably much closer than i think. Ill start making some thicker ones to compare.

Pete VanderLaan 12-20-2016 01:50 PM


Originally Posted by Art Freas (Post 133049)
Maybe stating the obvious here but to do these tests you need color pots to gather the colors, correct? Also Scott, if you can video that process for the live test I would be interested in seeing it. Thanks much for all the information.

You can do it with color rods. Just make the color the inner gather and then case in clear. Keep in mind, most color won't survive the saw from certain manufacturers. Gaffer should be fine with Spectrum, SP87 and Cristalica. Not so much with System 96. Kugler opaques won't do well.

The only reason you need a live test is if it is pot glass and you want to spend the day using it and not waiting for an annealer to cool.

Josh Bernbaum 10-20-2019 09:27 PM

Sorry to resurrect this thread, but I had a question about ring testing so I thought I'd stick it to this existing one.

The last color melt I did which I ring tested, like I do with all the colors I make, had only one crack run as I sawed it, but then after that I was able to saw a different section of the cylinder lower down and that ring sawed cleanly. That was color on the inside and clear on the outside. I don't feel comfortable with the result if it doesn't saw cleanly the first time, and if anything there was a tiny overlap in the score of the ring I was able to get. So two days later when I was at the furnace again I made another cup like that one and, for extra testing, one with clear on the inside and color on the outside. The same result with the redo of the first cup with color on the inside, and the one with color on the outside cracked more immediately and more violently when sawing. And I couldn't even get a ring out of that one. So question is, can I glean any more info off the even worse result with the color on the outside? Like would that tend to happen with higher or lower expansion of the color on the outside?

Pete VanderLaan 10-21-2019 09:36 AM

It's a solid indicator of mismatch to have it happen. I've tried on occasion and gotten similar rings from a cup where the initial sawing failed. They are often pieces where the vertical score just makes the ring break up. Either way, you're close but no cigar. The fact that reversing the casing sequence causes an immediate failure supports mismatch as well. If you pull a simple cane of the two glasses, what does that show you? Do that both annealed and not annealed.

I remember when Ed Skeels had a set of samples that pulled straight and then he annealed them and they bent over badly. We never pursued that but it remains in the back of my mind and particularly so with phosphate opals. The thing with any opals is the creation of, and subsequent collapse of material. It's worth noting in any of Peiser's phosphates, they are never cased in any other glass type. It's monolithic.

Josh Bernbaum 10-21-2019 01:53 PM

Thanks Pete. I haven't done a pull test, yet anyway, and in fact I've already scooped out the rest of that pot of phosphate just this morning. I've just seen such vastly different results from pull tests, which (I think) I'm doing correctly. For instance, on a few occasions when I've done them in the past, the pull test bend contradicted what both the spreadsheet and ring test results had been saying. I remember that being frustrating, like it wasn't really helping me any and just throwing more confusion into the mix. Maybe this is a limitation of the pull test since one is an opal phosphate and the other is a soda lime clear? I remember you saying pull test not reliable with vastly different glass bodies, like the fluorines for example. It's worth saying though that I had not annealed any of those previous pull tests.

I suppose I was just curious about the ring test where the outer glass was the phosphate and the inner was my clear and if the outer glass was showing that it was in compression or tension (?) if it broke more violently than the one with the clear on the outside. I was hoping that could be some kind of other indicator here. I have noticed in some pieces I do with cane, there seems to be less tolerance when the cane is on the surface of the piece, versus when the cane is cased, or even better, sandwiched between layers of clear. So just thinking out loud here, and if I need to do two test cups each time, one the reverse of the other, just for extra input on the fit. I think I just might have to now..

Pete VanderLaan 10-22-2019 07:09 AM

Surface tension is really important so when you put the mismatch glass out there, you see quick results.

On the pull test, the glass on the inside of the curve has a higher expansion
On the ring test the inner glass gapping open indicates the inside glass has a lower expansion.

The annealing part is something I threw in as curious and I can't say why. It is the case that substantially different formulas make the pull test less than reliable. Go read about it in Glassnotes IV

Josh Bernbaum 10-25-2019 02:38 PM

So on my latest phosphate white this week, both a pull test and what I could glean from a ring test where I barely got a ring to look at after cracks ran, indicated that my expansion of the white base is too high. Pull bends toward white and what I barely got of a ring section with white on the inside, clear outside (Spruce Pine), overlapped after scoring.

Debating whether to add sand or take away some of the soda and/or potash from the next mix. Sand is currently at 55.4%, soda at 14.2%, and potash at 6.2%.
Borax and strontium are both a hefty 5.1% each but they are supposedly there to help the phosphate do its thing I think. Total batch weight of the next mix will be 35.2 lbs., before I adjust any.

Question: Given that I seem to be a bit high expansion and I couldn't saw a clean ring, any recommendations on how much I should try and move the calculated LEC down on the spreadsheet? 1 point to start? Less? More?

Pete VanderLaan 10-25-2019 03:26 PM

Silica is actually a very powerful way to move expansion and if at a rather low 55%, it would be the material I would consider.

Is this still based in Dave's formula. I haven't looked at its percentages but these seem pretty strange

Try to recall that with the old bob held study, he established you could add up to 5% calcium mono phosphate and still be compatible with the basic clear. The problem was the calcium creating Apatite stones so going to Strontium and dumping the calcium out of the base was necessary. Since the strontium had a not dissimilar excpansion factor, this was not a tough substitution. It feels like you've strayed from the earlier materials.
So at this point, I'm assuming you are using Sodium tri phosphate for the phosphate source (Is it?). That should make for a fairly easy transfer from the soda ash as well.

I just never had this trouble with the Bross formula although I spent little time with it. Also, my clear was non commercial but matched SP87 on expansion and viscosity, which was the point.

Here's what Jordan did with this stuff:
In pounds
Sand 7.3125
Soda Ash 2.38875
Potassium Carbonate 0
Calcium Carbonate 0.4125
Strontium Carbonate 0.675
Zinc Oxide 0.225
Alumina Hydrate 0.3375
Lithium Carbonate 0.075
Antimony 0.03975
Potassium Nitrate 0.075
Sodium Tripolyphosp. 0.93

Total 12.471

Oxides %
SiO2 66.86
Na2O 16.4
K2O 0.327
LiO 0.278
CaO 2.13
MgO 0.006
SrO 4.35
ZnO 2.06
Al2O3 2.229
P2O5 4.95
Sb2O5 0.365
While I don't know your intent, you've wandered far from home. The silica alone is remarkable.

Josh Bernbaum 10-25-2019 03:39 PM

35.2 lbs overall batch weight. Yes, this is just about the same as Dave's, I just took away some sodium from his to put in some potassium from mine. And yes, STPP is my phosphate source.

What has me pulling my hair out is that after 4 melts of this that FIT, both with and without colorants, the last 2 melts don't fit anymore. I can't afford to lose much more hair with my already receding hair line from male pattern baldness..

All the variables I can think of are the same, I'm so confused, but the last two melts seem to be high in expansion now for some reason. I triple-checked my scale measurements this last time just to be sure. What's the biggest difference I can think of? New pallet of SP batch I'm melting now, but given their good track record on consistency, that's gotta be unlikely that the expansion of their mix lowered any, right?

So if we are looking at sand to move first, what would you think first jump should be on a 35.2 lb. total mix?

Pete VanderLaan 10-25-2019 04:16 PM

what did the simple pull test show you?

I do not like that silica content.

Go back and re read this entire thread. It my help.

Josh Bernbaum 10-25-2019 05:08 PM

The simple pull test showed that the phosphate white was higher in expansion than the SP clear. It bent toward the white a decent amount I'd say.

Here's a formula posted here by Dave in 2016:

Sand 97.5
Soda 36
Lime Hyd. 5.5
Strontium 9
Zinc 3
Borax 9
Alumina Hyd. 4.5
Lithium 1
K Nitrate 1
STPP 6.5

Mine is very similar to that, just put in some K carbonate instead of some of the sodium carb. I calculated his percentage of sand at 56%, by dividing 97.5 by the total weight of 173. But maybe I'm going about that percentage calculation totally wrong, because I see Dave listed the SiO2 content of the above formula at 65.32%. I'm guessing that's more in the range of what you'd like to see for sand content. If Dave is correct on that, then that will be very similar to where my sand content is at as well.

Josh Bernbaum 10-25-2019 05:16 PM

In fact I just checked on your spreadsheet and the silicon % of my mix is 64.94%. I don't know what I did wrong with my division and what I thought was more like a 55.4% sand content. Maybe I'm just calculating weight percentage and not thinking hard enough about oxide percentages which the spreadsheet seems to be accounting for?

Pete VanderLaan 10-25-2019 06:43 PM

I'd say, get back to basics.

64% silica is low and it would benefit from better durability even though one can make the argument that Phosphorus is a network former.

If you can work at getting rid of calcium, and boosting other stabilizers at that moment, then look to Strontium and zinc.

I continue to not like borax. If I was trying to lower melt conditions, I would head for nitrates. Lithium is really expensive. Sodium is certainly cheaper than other alkalines such as potash but I love the luster from potassium as well as its ability to promote nice color. The down side is it makes a fatter glass. Potassium nitrate is a great balance.

If this was easy ( it's not) when you get what you want, no one will be able to copy you. That's been an advantage for me now for over thirty years. I'm not a very good glassblower.

Pete VanderLaan 10-25-2019 06:52 PM


Originally Posted by Josh Bernbaum (Post 145705)
In fact I just checked on your spreadsheet and the silicon % of my mix is 64.94%. I don't know what I did wrong with my division and what I thought was more like a 55.4% sand content. Maybe I'm just calculating weight percentage and not thinking hard enough about oxide percentages which the spreadsheet seems to be accounting for?

you have to look at the oxides supplied, not the compounds used. Sodium carb, NaCO3 supplies NaO. Now do the atomic weights for the carbon and oxygen lost. . It's seriously different. SP I think supplies around 81 lbs of actual oxides per 100 lbs batch.
Go back to GlassnotesIV. Henry has a list of materials to oxides supplied. My spreadsheet does it as well but you took a bad side road here.

Josh Bernbaum 10-25-2019 07:05 PM

Everything else being equal (which is the case as best I can determine), how unusual is it that I'd have a base mix which fit SP clear four times in a row and then all of a sudden it didn't? I pay a lot of attention, in fact I obsess over details and minutiae most of the time (which is a affliction), and I am having trouble finding many if any differences between the first four that fit the clear and the last two which did not.

Pete VanderLaan 10-26-2019 08:16 AM

Sometimes someone has to stand outside and look at the aassertions. One of the things about phosphates and fluorines, or most anything that promotes either phase changes or crystals is how easily they can change in the pot. I can't speak to your details but it's hard to beleive that it worked four times and suddenly didn't. Something changed. You need to figure out what that was.
Rtae of charge, temp of charge, new materials, scales, simple error, humidity. moon was in feces...

When you moved the phosphate to the outside, I assume that was gathered. How thick was that casing. I get criticized periodically for using too thick a layer in the ring test. It's John's frequent response as to why he prefers the hagy seal. I still conduct all tests, dilatometer, ring Hagy and pull. You might consider doing a small increase in the silica, put the phosphate back on the inside, see what occurs as long as you can't do a hagy seal. If you like make the hagy seal, anneal it and send it over to me.

All times are GMT -5. The time now is 02:30 AM.

Powered by vBulletin® Version 3.7.2
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd. Opportunity Network. 2008. All Rights Reserved.