Ok, everybody is spilling their guts here now (or maybe just some of us as there need to be some business and trade secrets held back by those whose livelihoods depend on it), so I'll pile on too and try to relate a few things here to the other discussion about the Seger formula. I won't try to explain the Seger stuff, but just use what we have here to show its practical application. I don't use a hand calculator to derive the Seger unity formula, but rather a glaze calculation program on the computer (I use Glazemaster, but there is also Insight and others too) to just enter the recipe ingredients and amounts. The program does all the heavy lifting. Then I print them and lay them out on the table to look for patterns.

I have done this with about 60 crystalline recipes in all the standard books (Ilsley, Creber, Price, Shimba) plus others that have been acquired along the way. From this, I have found that despite the recipes' differences, their Seger unity formulae all seem to fit within relatively narrow limits. The sodium + potasium needs to be in the 0.2-0.3 range, with the total alkali fluxes (Na, K, and Li) summing within 0.25-0.4; further, the sum of the non-alkaline fluxes except zinc (i.e., Mg, Ca, Sr, and Ba) should be within the range of 0.1-0.15; and then zinc needs to be in the range of 0.5-0.6. Alumina and boron need to be very low at 0.03-0.05 each. Finally, silica needs to be between 1.3-1.8, yielding a Si:Al ratio of between 35-50. Translating those numbers back into general potter-speak, it's a glaze with a high proportion of it's flux coming from zinc, the rest of the flux preferably coming from sodium, plenty of silica, and not enough alumina for the glaze to stay on the pot when melted.

Kat and Bill have similar cone 6 glazes based on a mix of 3110 and 413 frits. (For reference, 3110 includes some K and some Ca with the Na while 413 includes some Mg with the Na; their proportions of Si and Al are a little different; and 3110 has some boron while 413 doesn't.) Compared to Bill's, Kat's has a bit less of the frits and zinc, more silica, and 5% strontium tossed in, while Bill's has less titanium and 2% lithium tossed in. But when you put them into Seger Unity, the fluxes are a near match except for the swap of Sr for Li; alumina and boron are a dead on match; and Kat has a tad more silica which is still well within the acceptable range for cone 10 glazes but just over the convention for published cone 6 glazes.

But now Bill pops in with his experiments with a frit 644-based glaze, which is where things get interesting. I have a similar glaze from the Ilsley book which I adapted to cone 6 (as a starting point for this trick, pull any kaolin and add 2-3% Li carb, and then test out from there until you get what you want...). My recipe is

frit 644-42%
Zn-24%
Si-27%
gerstley borate-1%
Ti-6%
Li-3%

Bill's experimental base is similar to mine except he chopped 6 off the frit and 5 off the silica so he could put in 10 of some other things to see what happens. When I put this into Glazemaster to get the Seger Unity, the results are predictable. The only flux in 644 is sodium, so that is lower, as is the silica, but everything else is close. Because the silica is low, the silica-alumina ratio is also too low - 21 vs. a minimum of 35 in the conventional recipes. A low Si:Al ratio means the glaze will be too stiff for the crystals to form easily. Also, there is no boron in the 644 frit, but more about that later. So now, on to the test additions.

Soda ash, is, well, sodium. Now the sodium content is back up where it belongs, but silica is still low. However, soda ash is soluable, so I suspect that some of that portion of the sodium technically in the glaze recipe actually dissolved and went into the ceramic of the pot rather than stay in the glaze.

Magnesium is in frit 413 of the other glazes, but 10% in this recipe is too much and takes the non-alkali flux total out of range. Also, the Zn at .47 is just under the .5 minimum.

Dolomite is about equal parts magnesium and calcium, but like the MgCo3 test above, 10% takes the non-alkali fluxes out of range and the Zn is still a shade low.

Frit 3134 had the biggies. Wow! This looks like another piece of puzzle! A few years ago on the old forum, Ivor Lewis from Down Under was discussing the esoteric chemistry of phase shift diagrams (still a mystery to me) and mentioned that a touch of boron in the lithium/silica phase shift brings the melting point down quickly, which is what led me to toss in that 1% ghastley borate with the lithium when I "adjusted" Peter's recipe. Among the frits, 3134 has a substantial amount of boron (23% by weight vs. 3% for 3110). I considered using 3134 for my boron source, but GB has more boron per unit so using just 1% of that gave me the trace I wanted and had less affect on the rest of unity formula. Keeping the boron to a trace amount is consistent with the commentary over in the "Seger for Koz" thread by Andreas Widhalm (copied over from the lost forum by John Tilton), in which he noted that some boron would help with crazing, but too much would kill the crystals. But now along comes Bill, throwing in 10% 3134 which drives the boron figure in the Seger Unity up by an order of magnitude to 0.066, over the 0.05 limit seen in most other cone 6 crystalline glaze recipes and well over the 0.04 limit seen in most cone 10 recipes. So, it appears to me that Bill has just proven both Ivor and Andreas to be correct. Cone 10 glazes have no lithium. Adding a slight amount of boron in a cone 10 recipe helps crazing but too boron much acts like too much alumina and kills the crystals, so sez Andreas. Cone 6 glazes use lithium to bring the melting temperature down, and boron working together with the lithium is a double whammy, so sez Ivor. Along comes Bill tossing in yet more boron with the lithium and he gets whoppers despite the silica to alumina ratio being a third lower than it should be. This is huge! (pun intended)

Strontium - interesting. Significantly increasing the non-alkali flux total with strontium gave satisfactory results while using Mg and Ca above did not. However, 10% Sr kept the non-alkali total within the conventional limits while similar percentage amounts of Mg and dolomite recipes exceeded the limits. Maybe try those again with only 5% MgCO3 and dolomite.

Whiting and talc - like the MgCO3 and dolomite tests above, just different proportions of Mg and Ca, and the Si:Al ratio is still low.

So, what to do now. The Seger Unity formulae suggest trying again with more Si and less Mg, dolomite, whiting, or talc.

Time for some more shake'n'bake!

cheers
dw