sifting

Below is a photo of my third attempt at home milling and sifting, which resulted in a flour very similar to my favorite "high extraction flour", Heartland Mills Golden Buffalo flour. The processes used on my second and third tries are explained further below. Additional photos of the process have been posted.

First Try

The first bread from my home milling and sifting project, blogged earlier, looked like a 100% whole wheat bread. Unfortunately, I still hadn't figured out a way to do home ash content testing, but from the results, a guess at the ash content of the flour that went into my first try might have been something like 1.4%. So, it had some of the darker material sifted from it and therefore had a lighter crumb than a 100% whole wheat flour might have produced, but the color and flavor was closer to 100% whole wheat.

Second Try

My second try was a little lighter but still closer to whole wheat in character. I allowed the sifting process to go on longer and used a couple of passes. After one pass through the Retsel mill at a fairly fine grind and then sifting through a stack of sieves (#25,35,45,60,70,80) on my sieve shaker, the breakout was as follows. A video of the equipment in operation is posted for fun.

I then took the 219g caught by sieves 35,45, and 60, and re-milled them at about the same settings as the initial milling. The output of this second milling was then fed through the #60 sieve. The output was 53g of coarse material caught in the #60 sieve and 156g of somewhat creamy, grayish flour that went through the #60 sieve.

I then created a flour that is about 82% extraction by combining the all the flour that fell through the #60 sieve on the first pass with enough of the flour that fell through the #60 sieve on the second pass to constitute 82% of the total output. The resulting flour was lighter than on my first try, but the bread that resulted still had a color more like a whole wheat bread, although slightly lighter in color. The flavor was noticeably different, though. The second try had a flavor with far less of the grassy flavor of a whole wheat bread. Again, this flour was made before I had a way to test for the ash content, but I imagine from the color of it, that it was probably about 1.25% ash content. It was slightly darker than Heartland Mill Golden Buffalo flour. My second bread also had 5% whole rye and 10% whole spelt in it, as did the first one, so part of the whole wheat character of these loaves is caused by the addition of 15% whole grain flour.

Third Try

I received my Meadows 8 Inch Stone Mill and decided to have another go at milling and sifting. Of course, the new mill works differently than the Retsel. The stones are much larger and turn much faster. I can't seem to get the grind anywhere near as fine as the Retsel will produce with just one pass. However, the Meadows mill is far faster, especially when re-milling flour. The Retsel takes forever to re-mill flour, and seems to heat up too much on a second milling. The Meadows Mill takes less than a minute to grind a few cups of grain, and re-milling the output takes only slightly longer.

I was happy to discover that for the amounts I would normally do - not more than a 5 pounds at a time, the flour was very cool coming out of the mill. In fact, it was noticeably cooler in temperature than the flour coming out of the first pass with the Retsel mill. I imagine that equation would reverse for much larger amounts, as the Meadows would heat up over time to a higher temperature, given the large stones turning at much higher revolutions per minute.

This time I went for about a 70% yield. I realize in retrospect that my first pass was probably too coarse, which resulted in only about 600 grams going through the #60 sieve and 323 grams caught in the #25 sieve, out of a total output of 1815g. I then re-milled the middlings from that sifting, and the output was 350g through the #60 sieve. One more pass resulted in an output of another 244g through the #60 sieve. The flour coming through the #60 sieve from this pass was lighter than previous attempts.

I did another sample of about 300g which was milled at the finest settings a couple of times. The result was a finely milled whole wheat, more like what would be done on a very fine first pass with the Retsel. The result was sifted through a #25 and 50 sieves to get 240g of flour, with only 15g of "bran" caught in the #25 sieve and middlings of only 40g. This was probably too fine. I'm slowly beginning to understand what setting of coarseness of the mill will result in a good distribution of particle sizes for more efficient sifting to get the flour desired.

The resulting flour was actually 68% of the total flour made during this session trying a couple of different strategies. This time, I was able to measure the ash content, at least approximately, using the home ash content measurement mentioned in a previous blog entry. The ash content is around 1.05%, maybe a little lower than Heartland Mills Golden Buffalo flour, which their site says is around 1.13% ash content and I calculated to be around 1.2 with my test, such as it is.

Resulting Bread

A similar bread to previous attempts was made with this flour. However, I omitted the 10% spelt and raised the hydration to about 81% to compensate. I may have gone a little too far with the hydration, as I had some trouble getting the loaf to hold its shape well. Due to some unanticipated distractions, the loaf was about 20 minutes late getting into the oven, so it was also slightly overproofed. The result was therfore flatter than I would have liked. However, the crumb, crust, and flavor were all very good. I believe this loaf is very similar in most ways to country miches made with Heartland Mill Golden Buffalo flour in the past. The color is a little darker, but I believe that has more to do with the fact the flour is not aged, as the ash content clearly indicated that my flour was lower in conductivity than the Golden Buffalo flour and should therefore be a little closer to white flour than the Golden Buffalo flour. The texture of the dough and the general behavior of the flour while handling it seemed very similar to what I have experienced with the Golden Buffalo flour. By the way, the wheat berries used for this flour was Heartland Mills M2, which may be similar to the wheat berry product they are using to create the Golden Buffalo flour. Overall, I'm extremely happy with this result. The flavor and freshness of the home milled flour is a delight, and the prospect of being able to freshly mill a desired grade of flour on demand is pleasing.

Future Attempts

Now that I have a better feel for the right mill settings, my plan is to do a multiple pass approach, this time hopefully more systematically and with better mill settings. The outputs of the various passes will be saved and ash content measurements performed on each one. Hopefully, I can then make the process much more efficient and flexible. With ash content measurements available, blends can be created based on ash content of the final flour desired, and hopefully better yields will result for the same ash content, with better coarseness settings on the mill on the first and subsequent passes.

My home sifting project resulted in "middlings", a term I may be using incorrectly. What I mean by middlings is the stuff I sifted out that is finer than bran but was coarser and darker than I wanted for the flour being produced.

This output of my milling and sifting process had a coarseness similar to semolina or maybe a little more coarse. It was a fairly dark brown. I refrigerated it, thinking it might be useful for dusting a couche or some other purpose eventually. To some extent, I was hoping to discover some good food use for this part of my output, which should contain a fairly large nutritional content, since it has much of the darker, vitamin-rich outer layers of the wheat berry in it. My more whole grain oriented friends might be less disapproving of my use of less than 100% whole wheat flour in some of my breads, if I could show that the other parts of the whole grain are still being used. Also, my wife is more interested in whole grain nutrition, so she asked me to save it, probably also imagining some good use she might discover for very freshly ground outer layers of the wheat berry.

The nice thing is that I can see this output will be consumed nearly as quickly if not more quickly than the bread that was made from this sifting session. My whole wheat loving friends would be happy, since we would be eating 100% fresh ground whole wheat by eating the bread and having the cream of wheat middlings and bran for breakfast.

This morning it occurred to me that the "middlings" were a lot like cream of wheat in consistency, just browner. I decided to try making "cream of wheat middlings". I forgot to measure, but roughly speaking the recipe was 1.5 cups water, 1.5 cups skim milk, 0.5 tsp salt, 1.5 cups of "wheat middlings", and about 0.5 cups of "wheat bran", the coarsest output of my sifting process. I then brought it to a strong boil, dropped the heat to low, and let it simmer, stirring periodically, for about 15 minutes.

The resulting gruel was served with some milk poured on it, and some brown sugar sprinkled over it. My 13 year old son wolfed this concoction down with great delight, saying it was very good. I thought it was a great breakfast, more flavorful than cream of wheat and probably nutritionally much superior, and it would have significantly more bran fiber, for those who might like that aspect of it. I tried adding raisins to some of it, which I thought made it even better but my son thought detracted from it.

Recently, I've been attempting to grind and sift my own flour. The grinding is straightforward with a Retsel Mil-Rite, an excellent home stone buhr mill or my new Meadows 8-inch stone mill. However, the mysteries of sifting the flour have been less straightforward. A subsequent blog entry will deal with my progress on grinding and sifting my own flour. The sifting project motivates the need for measuring the ash content of my flour.

Ash Content

Ash content in general is the percentage of inorganic matter in a sample of some material. It is used in many different ways to analyze agricultural products, at least, based on some cursory sampling of articles on the internet.

About.com says defines ash content as:

The nonvolatile inorganic matter of a compound which remains after subjecting it to a high decomposition temperature.

A traditional method for determining ash content is to place a sample of known weight in a furnace at high temperature (600F or higher) for a number of hours (12 hours, for example) such that all the water, volatile compounds, and organic matter either evaporate or burn. After that, the remaining material is weighed. Ash content is the weight of remaining "ash" expressed as a percentage of the original weight of the sample. The remaning mass will be the inorganic non-volatile compounds that were in the original sample.

Flour ash content in Europe is measured using a dessicated (dried out) sample  of flour, so the original weight of the sample doesn't contain any water. In the US, a moisture content of 14% is assumed (typical for white flour before it is dried out), so US numbers for ash content differ from the same European measure by the amount of water in the original sample.

An Important Characterizing Measure of Wheat Flour

Ash content is widely used in Europe to classify flours. When you see "type 55", for example, the 55 refers to the ash content, which would be 0.55% of dry matter in this flour. In the US, it is often available by searching a manufacturer's or supplier's web site for flour specifications (often hidden somewhere hard to find), or more often, by calling someone in their testing department.

Why Ash Content

The inorganic matter in a wheat berry is heavily concentrated in the outer layers, such as the bran, various seed coatings, and the germ. As you traverse from the outer coatings to the outer endosperm and then to the inner endosperm, the concentration of inorganic matter steadily drops.

During milling, the flour is ground, then sifted, then ground again, and sifted again repeatedly. When the milling process is complete, a large number of bins of product will result from very coarse to very fine, and from very dark to very light flours. The whitest flours will have less ash content, and the darker flours will have more ash content. At this point, various grades of flour may be created by blending the flour from the bins.

Ash content then summarizes how much of the outer layers made it in to the final flour, regardless of how it may have been milled, sifted, and blended.

The importance of measuring ash content was immediately obvious to me as I tried to mill and sift at home on my own. An infinite number of possible permutations of grinding and milling could be imagined. For example, I tried grinding very coarsely, then sifting, then grinding the coarser results of the sifting again, then sifting again. Another version was grinding very finely and sifting into more and finer sizes. I also tried grinding coarsely, then regrinding, then sifting. Of course, the possibilities are endless. In each of these cases, flour resulted that made good bread, seemed light in color, and fine in texture. The difference to the eye and the feel in the hand was not great between one and the other, at least not to me, a first-time home miller.

Measuring ash content of my results would make it possible to know at least approximately how much of the outer layers had made it into each type of flour resulting from the various grinding and sifting processes tried. Also, once a given process is adopted and used consistently, calculating the right blend of the various outputs of the milling process to achieve a desired ash content, depending on the type of flour needed, should also be fairly easy.

The Theory

Distilled water doesn't conduct electricity. However, if some salt is dissolved in distilled water, it will conduct electricity. The ions contributed by the salt are charged particles that will travel through the water in the field created by the voltage difference on the electrodes of the conductivity meter to create a flow of electric current. The higher the concentration of salt, the higher the conductivity of the water and salt solution will be. The diverse mineral content in the inorganic matter that makes up the "ash content" of the flour ionizes the water in the same way described above for salt. If the flour has a larger amount of "ash content" it will also contribute a larger quantity of ionizing compounds to water, increasing the conductivity. 

The Equipment

To measure conductivity you need a conductivity meter. In the field of water quality measurement, "Total Dissolved Solids" is a standard measurement, but it is essentially a measure of the conductivity of the water being tested. So, you can use either a "conductivity meter" or a "TDS Meter". In my case, I had obtained a Hanna 9813 pH meter a number of years ago, and it turns out it also had a conductivity meter function. However, it was easy to discover conductivity meters on the internet, by searching on terms like "Conductivity Meter", "TDS", "Total Dissolved Solids", "Water Quality Meter", and so on. One place I found was http://www.technika.com. Also searching on "Hannah Meter" might work, since that's the brand of meter I have that has both pH and conductivity meters, both useful functions for flour measurement.

You might wonder why a standard digital multi-meter wouldn't work. I tried to use one unsuccessfully. First of all, you would have to carefully mount the probes to maintain the same distance apart and total surface area exposed to the water. However, it gets worse. The DC current used by a digital multi-meter to measure resistance causes the ions to build up on the electrodes, so the measurement just goes higher and higher the longer you leave the electrodes in the water. Conductivity meters made for measuring water impurities use AC current to measure the conductivity so the above problem with an ohm-meter doesn't occur, have probes made of less reactive conductors, and are designed to maintain proper spacing of the electrodes.

The Method

I found a couple of papers on the internet describing methods of measuring ash content with conductivity. One was especially useful for home measurements and was titled, "Electrical Conductivity of Flour Suspensions and Extracts in Relation to Flour Ash." published in 1977 in the Journal of Cereal Chemistry. The method described below was derived from the discussion in this paper.

The method is very simple. Mix 100 grams of distilled water (should be distilled water to get good results) and add 5 grams of the flour to be tested in a container. Stir thoroughly to completely hydrate the flour. Periodically stir for about 12 hours. After the flour has settled to the bottom of the jar, measure the conductivity of the water. For the best measurement, allow the flour to settle on the bottom so there is clear water to measure. The clear water will have a higher conductivity than recently stirred and cloudy water. At first the conductivity rises, as the various compounds that contribute to the conductivity of the water dissolve, but at some point the conductivity will stabilize. In my case it took a long time, maybe 12 hours or so, for the conductivity to stop changing. The conductivity measured can then be calibrated by measuring flours with known ash content and fitting a curve of conductivity to the known ash content. In practice it looked very linear, so even a simple proportional relationship would give reasonable results, based on my admittedly minimal sampling.

The table above shows measured conductivity in ppm, as the meter represents it for TDS or "Total Dissolved Solids" in parts per million salts for a hydroponic solution and also shows conductivity in the more standard measure of milli-Siemens per cm. I don't know the ash content, but based on some flour specification information from Heartland Mill, I filled in rough numbers and then used them to approximate the ash content of my "71% yield, fairly white bread flour" sifted from a couple of passes with my new Meadows 8 inch mill and a couple of siftings with a number 60 sieve in my new SS-100 Econo-Shaker sieve shaker.

The method in the paper heated the samples to boil them for a short period, then cooled and centrifuged the samples to create a clear liquid with the dissolved minerals in it. I didn't want to deal with boiling or somehow obtaining a centrifuge. OK, maybe you could put your jars in bags, tie them to some rope and spin them like Argentine "bolas", but I recommend patience. It was unclear what the effects of boiling were from this paper, but it seemed to affect the measurement in some unexpected way. So, my approach is to keep it simple and just wait for the conductivity and the flour to settle, even if it takes a while.

Summary

You can obtain a reasonable estimate of ash content by mixing 5 grams of flour with 100 grams of distilled water, stirring periodically for a few hours and then measuring the stabilized conductivity and comparing to the same measurement for some reference flours of known ash content. I proceeded to make one of my favorite miche recipes and found this flour to give very comparable results to Heartland Mill Golden Buffalo flour, which is of similar ash content. The difference is I can mill my own version of the Golden Buffalo flour and obtain it absolutely fresh when called for. In addition, measuring and recording the ash content of the output from the various passes of grinding and sifting should allow me to blend the outputs in the right proportions to obtain a desired ash content for recipes that may call for more refined or less refined flour.