Jump to content

Pieter Mostert

Member Since 09 Nov 2012
Offline Last Active May 24 2017 02:14 PM

#126753 Free And Open Source Glaze Software

Posted by Pieter Mostert on 16 May 2017 - 04:09 AM

I've been working on some software that helps you compose glaze recipes that have a specified oxide composition: given some set of ingredients, find which combinations of the ingredients give you the desired combination of oxides. Most of the recipes I come across use ingredients that aren't available here (I'm in South Africa), but I can reproduce them on the level of oxides. They don't always come out the same, but it's a good way of producing useable glazes.

Below is a screenshot of the user interface:


For each oxide and ingredient, the blue numbers on the left are lower and upper bounds that the user imposes, while the red numbers on the right are the calculated minima and maxima among all recipes that simultaneously satisfy each of the imposed bounds. The green polygon shows the possible values of the percentages of Potash Feldspar and Neph Sy (I still need to insert labels). In this example, I'm trying to reconstruct this recipe, but with Wollastonite instead of Whiting, and using my own potspar and Neph Sy instead of Custer, and a couple of locally available clays instead of OM4 Ball Clay. I've ignored MgO and TiO2, and slightly decreased CaO and KNaO, to allow for a bit of wiggle room.

If I increased the minimum K2O UMF value to 0.15, as in the recipe, the plot of Potspar vs Neph Sy diagram would change accordingly:



At the moment the program consists of a bunch of Python scripts, so most people will probably not want to instal and configure the extra bits and pieces necessary to run the program in its current state. I'd like to eventually turn this into a (free) stand-alone program you can run on your desktop computer, but since I'm not a professional programmer, I'll need some help to do this, and iron out some kinks along the way. If you're comfortable programming in Python, and have time to spare, I could use your help. You can find the repository on Github (published under the GPLv3), and details about the code structure and bugs in the README.pdf document. I've only tested it on my Windows 7 computer, and a couple of other people have had trouble getting it to work on their Macs, so be warned.

#126641 Ian Currie Test Tiles Forums?

Posted by Pieter Mostert on 13 May 2017 - 04:55 AM

A simpler way to work out how much base remains after removing 2ml from 48ml is to multiply by 46/48 = 0.95833...


So if you started with 30.64g per cup, you'd have 30.64g x (46/48) = 29.23g of the base left after removing 2ml.


At the next step you'd have 29.23g x (46/48) = 30.64g x (46/48) x (46/48) = 30.64g x (46/48)^2 = 28.14g of the base left.


If you continue in this way, you'll find the amounts differ slightly from the ones in the spreadsheet. The difference isn't large enough to make a practical difference, but for anyone worrying about where it comes from, it seems to be due to the fact that when you work out the amount of oxide removed, the spreadsheet takes the proportion of oxide in the mixture to be 0.02, whereas it's actually 0.02/1.02.

#113487 Nerd Playing With His New Toy

Posted by Pieter Mostert on 20 September 2016 - 03:25 PM

Nerd, I think you'll enjoy this blog post.

#112988 Seeking More Technical Understanding Of Firing To A Cone

Posted by Pieter Mostert on 13 September 2016 - 06:04 AM

Dick, I've always assumed that the reason a short hold at a higher temp equals a long hold at a lower temp is that the rate at which the chemical changes occur in the glaze or clay-body increases with temperature, above some minimum thresh-hold. I don't think thickness is the critical issue. If the rate at which the reactions occur is comparable to the rate at which cones soften, then a glaze or clay-body rated at, say, cone 6 should be mature regardless of the firing speed, provided the cone 6 cone is bent.


Of course, this is a big 'if', and assumes the temperature of the cone is the same as the temperature of the glaze / clay-body. Also, that's not to say the glaze will look the same, since things like crystalization and smoothing over of pinholes depend on the firing profile.

The Orton website does state that

The starting premise of cone development was that a blend of ceramic materials compounded to exacting proportions should behave similarly to the ceramic ware being fired in the kiln. A great deal of work later the starting premise was confirmed...

It would be nice to know what 'similarly' and 'ceramic ware' mean in this context.

By the way, according to this presentation by Dave Finkelnburg, fritted glazes melt from the outside in, while feldspar-based glazes melt outwards.

#110845 Mixed My A First Batch Of Glaze

Posted by Pieter Mostert on 01 August 2016 - 03:55 PM

Diesel Clay's right, of course. You don't need Brongniart's formula to work out how much water to add, and in fact this isn't why Brongniart came up with his formula. The usual way in which the formula is presented expresses the weight of the dry ingredients in terms of the weight and specific gravity of the glaze. This is useful if you have a batch of glaze that you've been using for a while, but want to modify the composition.

#110753 Mixed My A First Batch Of Glaze

Posted by Pieter Mostert on 30 July 2016 - 04:35 AM

I think the formula Celia referred to is Brongniart's formula, which effectively says that the specific gravity of a glaze is the weight of the dry ingredients plus the weight of the water (in grams), divided by the volume of the dry ingredients plus the volume of water (in ml). Really, it's just saying that specific gravity is the weight divided by the volume.

For most glazes, the volume of the dry ingredients is equal to the weight of the dry ingredients divided by about 2.5, so if D is the weight of your dry ingredients and W is the weight of water (which equals the volume of water, in our units), then

SG = (D + W) / (D/2.5 + W)

Using D = 1000 and W = 631 (the volume of 1 pint + 2/3 cups in ml, if I've converted it correctly), you get SG = 1.58. It's a little on the high side, so I'm a bit surprised you need to decant water, but perhaps the value of 2.5 isn't quite right for the glaze you're using.

If you want to find out how much water to add to a batch of dry ingredients to achieve a certain specific gravity, you can solve the formula above for W. You should find that

W = D * (1 - SG/2.5) / (SG - 1)

So if you wanted a SG of 1.5, for example, you'd need to add 800ml of water to your 1000g batch of dry ingredients.


EDIT: I've fudged over the fact that the volume of the dry ingredients isn't the actual volume you'd measure, it's the volume you'd get if you could remove all the air between the particles. But this isn't something you need to worry about.

#103726 Laguna 65 And Crazing

Posted by Pieter Mostert on 18 March 2016 - 12:44 AM

John Sankey has a good article on COEs, which includes a table of the different oxide expansion factors that various authors have calculated. Some of them vary quite a bit.


Another good article is on the digitalfire site. I think everyone who relies on calculated COEs should read this.

#98613 Fluorine In Glazes On Earthenware - A Technical Challenge

Posted by Pieter Mostert on 02 January 2016 - 04:26 PM

Tim, I think the corrosion may be due to the formation of silicon tetrafluoride. The following is a quote from Out of the Earth, Into the Fire by Mimi Obstler, in the section on fluorspar, and I'm guessing something similar happens with cryolite.

"At temperatures somewher between 1652oF and 2190oF (provided silica is present, the disassociating temperature of fluorspar in the absence of silica rises to 2300oF - 2500oF), fluorspar breaks apart and releases volatile fluorine gas. The liberated gas attracts silica from the glaze and/or claybody to form silicon tetrafluoride (SiF4), which escapes from the kiln, leaving behind the newly separated calcium. ...

"The consequences of using fluorspar can sometimes be injurious to the ceramic surface, the kiln interior, and even the potter. The formation of silicon tetrafluoride robs the glaze and/or claybody of needed silica. In addition, the forcible exit of silicon tetrafluoride is often marked by blisters and pinholes. ... Although there is a difference of opinion as to whether the liberation of this volatile gas injures kiln interiors (see Parmalee 1973), Ceramic Industry reports definitively that silicon tetraflouride corrodes furnace linings. Our experience confirmed the corrosive nature of silicon fluoride. A test series that added 20% - 40% of fluorspar to Sanders Celadon glaze not only corroded the kiln shelves, but also attacked the surface of adjoining pots. Even more important is the fact that tetraflouride gas is a known carcinogenic material"

#98419 Calcium Borate

Posted by Pieter Mostert on 31 December 2015 - 11:25 AM

Min, the figure of 5 microns given here is for the median particle size (half the particles are smaller than this). Mesh size relates to the maximum particle size. If the maximum particle size of Imsil A-25 were 5 microns, then yes, it would be 2500 mesh, but I suspect the rep you emailed just assumed the figure you gave was the max size without checking. I don't know what the actual maximum size is, but unless the particle size distribution is very narrow, the mesh size is likely to be significantly smaller than 2500. I haven't seen anything shows it to be 325 mesh, and even if it were, it would still pose more of a risk than Sil-Co-Sil 45, where the median particle size is about 10 microns.

On the other hand, if you look at the graph for Sil-Co-Sil 45, you'll see it has slightly less than 30% of its particles smaller than 5 microns, so the difference isn't massive. Maybe it's not worth the increased risk to use a silica where 50% of the particles are less than 5 microns, maybe it is.


Okay, just in case I made a mistake I emailed the manufacturer and got the info straight from the source. 


my email:




Quick question if I may, I am trying to determine the mesh size of your Imsil 1-25 silica.I see that it is 5 micron, does that correlate to 2500 mesh size?


Thanks in advance for clearing this up for me, 





Hi Madaleine,


At 5 microns, mesh size is too small to exist as actual screen size but for reference, yes, it would be 2500 mesh.


Many thanks,

Angela Le

Sales Representative

The Cary Company

C 630-215-9781 | O 630-629-6600



  • Min likes this

#98259 Calcium Borate

Posted by Pieter Mostert on 29 December 2015 - 07:08 AM

+1 for what Bruce said.

Just to put things in perspective, the silica you need to be worried about is that less than 10 micrometers in diameter (see http://digitalfire.c...xicity_251.html). From the spec sheet for Sil-Co-Sil 45 that Tyler linked to, that's about 50% for a 325-mesh silica. I couldn't find any particle size distributions for Imsil A-25, but given that its median particle size is 5 micrometers, compared about 10 micrometers for Sil-Co-Sil 45, I would guess that the percent of Imsil A-25 less than 10 micrometers is equal to the percent of Sil-Co-Sil 45 less than 20 micrometers, i.e. 80%.

Of course, this doesn't tell the whole story, since the smaller the diameter of the silica particles, the more hazardous they are, so you'd need to compare the tails of the particle size distributions if you really wanted to quantify how much more dangerous Imsil A-25 is than Sil-Co-Sil 45.

#97900 Calcium Borate

Posted by Pieter Mostert on 24 December 2015 - 08:48 AM

Are you sure about the figures for Turkish Colemanite? Pure Colemanite (2CaO.3B2O3.5H2O) is 50.8% B2O3, 27.3% CaO, but perhaps you're referring to another type of calcium borate.


I've heard Turkish Colemanite is quite variable (don't take my word for it though), so you might be better off using a frit like Ferro 3211, or Podmore and sons 2244. I have no idea how easy these are to get hold of, or even if they're still being produced. Apparently you can also get synthetic Colemanite, but I imagine it'll be quite pricey.

#90434 Home Made Kiln Controller

Posted by Pieter Mostert on 08 August 2015 - 06:04 AM

Besides the issue of support, the impression I got was that the Arduino is less finicky than the Raspberry Pi. But I don't have personal experience with either of them, so don't take my word for it. I think Bob's suggestion is good, but since you've already started using the Raspberry Pi you might want to check out this site for a controller that runs on a Raspberry Pi. It's for a gas kiln, but you should be able to use the code (available on GitHub) to get the controller to read the temperature.

#90270 Home Made Kiln Controller

Posted by Pieter Mostert on 05 August 2015 - 05:48 AM

High Bridge, I spent quite a bit of time researching DIY controllers earlier this year after my kiln's controller died. For a while I was heading towards using an Arduino Uno with the MAX31855K chip, but I ended up just getting a PID ramp/soak controller from Auber Instruments. I don't have your electronics background, so I didn't want to have deal with trouble-shooting any issues that might occur, and the Auber controller wasn't that much more expensive. But programming a controller would have been fun.


If you want to program a PID controller, this is an excellent webpage:
You'll have to translate the code to use with a Raspberry Pi though.

The guy who wrote the code, Brett Beauregard, is also involved in the production of an open source PID controller. When I was doing my research they were having some technical trouble, but they seem to be up and running again.


Lastly, I thought I'd throw this out for anyone else interested:


It's cheap, but too advanced for my skills and I couldn't find any reviews of how well it works.

Let us know how your controller turns out. I'm pretty happy with my Auber controller, but if I get a kiln that needs zone control one day, I'd like to try making my own.

#80987 When Does The Most % Of Total Shrinkage Of Caly Happen?

Posted by Pieter Mostert on 07 May 2015 - 04:33 AM

The dry-to-fired shrinkage of the clay bodies I use range from 1.1% for a white earthenware to 8.8% for a dark red earthenware, both fired to cone 4. I've only worked out the wet-to-dry shrinkage for a few clay bodies, and have shrinkage rates of between 4.5% and 7.2%, but obviously this depends on how wet the clay is to begin with.


You could easily work out the wet-to-dry shrinkage rate yourself by making a small bar of clay and comparing the length before and after it dries.

#80904 Community Challenge #2

Posted by Pieter Mostert on 06 May 2015 - 09:56 AM

Do those of us in the southern hemisphere get to make a plant pot inspired by autumn?