homemade biltong dryer AKA the meat wardrobe

Biltong is a cured, air-dried meat. Some people dry it outdoors for a little extra tree and insect flavor, some hang it in a closet for a hint of cedar and lint, and some use a low oven and live without baked potatoes for a whole week. An artificial heat source is optional, although temperature and humidity obviously have a large effect on the drying time.

For my first attempt last year I wanted a somewhat controlled and sanitary environment for drying the biltong, as well as one that wouldn't conflict with other household activities. I ruled out the garage, oven, and my sock drawer and figured I'd build something.

My thoughts quickly turned to the wardrobe moving boxes we purchased several years ago. They were so expensive that I just couldn't bear to throw them out. They have over 10 cubic feet of volume, a shirt and meat hanging rod, several cutouts for carrying and venting, and a convenient access panel. I also had a small thermostatically controlled forced air heater with a safety shutoff. Perfect!

I cut a piece of scrap plexiglas, probably left over from the construction of some reptile enclosure, for a viewing window.

This spare oven thermometer goes down to 100 degrees F, which seems fine for the low end. The heater, pretty close to its lowest power and thermostat settings, holds the temperature just fine at 100 to 120 degrees.

I lined the bottom of the box with aluminum foil for sanitation. At least initially the meat does drip a bit. Here's my five pound first batch hanging.

And here it is all done.

One reasonable modification would be to raise the height of the bar. There's plenty of room to go up with it, and very long cuts of meat might hang low enough to be undesirably close to the heater.

Overall it ends up being a convenient and multitasking tool. The box breaks down for flat storage, and the heater keeps the garage smelling meaty fresh when I'm out there in cold weather.

testing the $3 popover pan

While dropping of some stuff at a thrift store I made my usual book and electronics scavenging run and found a nearly mint condition "professional quality" black steel popover pan in the original box for $3! (I was looking at more than books and bits, obviously.) The heavy black steel is supposed to lower the baking temperature by 25 degrees and decrease the baking time by 10% or more.

For starters I tried the recipe on the front of the box. At the time I didn't have any all-purpose flour so I figured I'd experiment with what I had: whole wheat pastry flour and white whole wheat flour. I also added the "speck of soda" penned in by the previous owner. I wasn't surprised when this didn't produce the ideal result. Probably any significant proportion of whole wheat with its jagged shards of bran is going to prevent the batter from capturing the steam that makes the popover pop. They did puff to a degree, but the interior was fairly solid. Still, they were light and tasty, like a whole wheat German pancake.

After restocking the cupboard with unbleached AP flour I tried again, omitting the soda speck. Now that I think about it, the author of that modification surely meant baking powder, not soda. Perfect results this time.

Alton Brown's popover recipe in I'm Just Here For More Food is a little different, adding some butter, using some water and less milk, and using one less egg. I'll have to try that next.

first coffee roast in modified Popcorn Pumper

I cleaned off my work bench the other day, organizing and storing the bazillion fasteners, tools, scraps, instruction manuals, and spare household pets. It was a beautiful blank slate for the first time since we moved in, what, five years ago? And it was literally crying out for projects!

While cleaning up the rest of the garage I found the Popcorn Pumper, which I haven't used for roasting coffee since finding the Poppery II. Seemed like a good time to toy with a few modifications. I started small, simply bypassing the thermal switch (and thereby creating a magnificent fire hazard). This is just a matter of soldering a wire between the terminals on either end of the thermostat and fuse, more or less according to this Engadget how-to. This keeps the heating coil on constantly, delivering significantly greater heat. (This great article has more explanation of the circuitry in these things.)

I also made an aluminum chimney to replace the bulky plastic hood, which really does not care for these temperatures. This was just some leftover aluminum roof flashing, rolled and fastened with J-B Weld.

The first trial run was excellent! I took it outside, preheated for a minute, and added enough beans for there to be some agitation but no swirling. I stirred occasionally with a wooden spoon, which is now possible since I can get straight into the chamber. The difference in heat was obvious. Roasting time was greatly reduced, and it's clear that any desired roast can be achieved without baking the beans. The chaff was all blasted directly out the chimney.

I didn't take any measurements but the results are clear. This is an enormous improvement over the original machine and the Poppery II as well. This was Sweet Maria's Classic Italian espresso blend and I took it to a very dark roast with a speed and consistency I've never seen.

2010-09-13 update: Tried again, this time shooting for a lighter roast. 4.65 ounces for five minutes produced this:

It was just entering second crack. That may be a little faster than is desirable. The roaster could have handled more beans. Adding beans generally speeds things up but I'm not sure that would be true in this case where I'm doing some manual agitation until the beans are light enough to fluidize. I guess the best approach will be to vary the supply voltage.

stovetop versus hot air roasting test 1

I'm finally getting around to doing some controlled tests between hot air coffee roasting and stovetop popper roasting. For this test I used Sweet Maria's Moka Kadir blend.

Stovetop roasting was outdoors at about 50 degrees, high humidity, no wind. Hot air using the Poppery II was indoors at about 68 degrees. Beans started at room temperature in both cases.

I roasted 10 ounces in the stovtop roaster, ending up with 8.43 ounces. Temperature was 550+ at bean drop. I ran it hotter than usual, staying above 350 for the duration and climbing to around 450 at the end. It came out at 4:20 with heavy smoke and some beans well into second crack. This was a fast and uneven roast.

I roasted 5 ounces in the hot air popper, ending up with 4.2 ounces. I preheated the popper for a few minutes and agitated the beans for the first few minutes as well. By three minute it was fluidizing unassisted, with a few already very dark and others barely colored. It came out at 8:00 with moderate smoke, when I judged the color to be comparable. It was even, as usual for hot air.

The closeness of the final (proportionate) weights may be a good indicator of the average darkness being nearly identical.

Toss-cooled both outdoors. There's a nonlinear relationship between mass and cooling time; the smaller batch takes perhaps one fourth as long. I don't have a way to measure the internal bean temperature so don't know whether that's also a factor.

Stovetop results:

Hot air results:

Tasted after resting for a day, pulling single shots. The grind was a little fine (30+ seconds). The hot air beans were a touch slower.

The stovetop-roasted coffee's initial moderate acidity mellows into sweetness, some fruit. Can detect both the very dark and bright tones, probably due to variability in bean doneness. It has both a slightly underdone and a slightly overdone quality, but it's quite good. I think it will be better in a day but it's totally drinkable.

The air-roasted coffee produced much more crema and the espresso had a much stronger coffee aroma. The acidity stays on the tongue considerably longer, with the sweetness blending into it. There is less complexity in the flavors and it's just not as lively. More body. Also good, but I think I slightly prefer the stovetop variant.

For the next test I should slow down the stovetop roast and try to reduce the variability in darkness. This can be done by starting at a lower initial temperature, running at a lower gas setting, or using more beans. Weather is a factor, too. While my opinion is that some variation is actually desirable, being responsible for the more interesting flavors, a better comparison can be made if it is limited.

Overall, a pretty successful test. I'll happily drink both of these roasts straight, with milk, and as Americano.

auriferous espresso in the Poppery II

I tried five ounces of Sweet Maria's Auriferous Espresso blend in the Poppery II. This was an amount that I thought would work well, based on the first attempt at four and the second attempt at six ounces.

I let the popper preheat for a few minutes, and gave it a shake now and then during the early going since it was not fluidizing. The beans in this blend are generally larger than the beans in Moka Kadir; perhaps that's a factor. At ten minutes I pulled the plug, with it very audibly in second crack. Color is even. Final weight was 4.13 ounces.

There was enough smoke to set off the alarm, even though I was running the fan on the range hood. That did not happen last time; I was surprised.

After a couple of days rest, I'm pretty pleased. It's a touch dark but I don't detect any other flaws. I think I'm ready (on a day with slightly less revolting weather) to do some head-to-head comparisons between the Poppery II and the stovetop popper.

second use of Poppery II

The Poppery II does work better with more beans. The first attempt was with four ounces of Moka Kadir; the second was with six. In eight and a half minutes it produced a glossy dark roast with plenty of smoke and cracking. Final weight was 4.87 ounces, or 86% of original weight. That's an interesting result. I usually see something closer to 80%.

The main problem is that while the beans are heaviest they aren't getting much agitation. It takes a few minutes before any degree of fluidization is evident. Some of the beans at the bottom were done very quickly. Giving it a shake now and then seems to help.

Tipping it back doesn't appear to be a good idea. It seems to interfere with the fluidization. A few beans do pop out when it's level, but not as many as I'd feared.

This roast ended up slightly uneven, but I think it will be better next time. Things to try:

• no tilting
• occasional manual agitation
• five ounces

The immediately-pulled shots were an improvement. Lots of crema, though not the foaming torrent from the last roast. The under-roasted flavor is gone and the solvent notes are faint. It's still pretty bad, as experience suggests will be the case before resting for a couple of days, but I think it may turn out all right.

After resting for a day there was some improvement. It's darker than I usually take Moka Kadir, with a bit of char. Aroma is nice, with chocolate dominant.

Update: It's not improving. It's not producing good espresso at all, in fact. I wonder whether I just need to clean the espresso machine.

Update: Indeed. That thing was disgusting. It's now decalcified and every user-accessible component is spotless. Whisked out the grinder, too. Pulled shots pretty close to the ideal time/volume. The result is a much cleaner-tasting cup. My main objection is probably the darkness of the roast, which I think has destroyed some of the fruitiness that I've tasted in Moka Kadir in previous roasts.

first attempt at roasting with a Poppery II

I discovered another hot air popcorn popper at the thrift store and felt it worthy of the $3 investment. It's a Poppery II and is compact and quiet compared to the Popcorn Pumper.

I immediately roasted four ounces of Moka Kadir. Interestingly, my results were almost identical to the first time I roasted in the Popcorn Pumper. After 15 minutes I shut it down, having produced almost no smoke, no sustained cracking, and at a much lighter roast than I would normally use.

The result for the first shots I pulled were about as originally described. Under-roasted, solvent aftertaste, very little fruit or sweetness, and low aroma. After a day of rest the only change was that the sweetness may have developed a bit and the aroma was improved.

An interesting effect was that the espresso was produced very quickly and with an absolute torrent of foaming crema. This behavior lessened slightly on the second day. I didn't experiment to see how far I'd have to reduce the grind in order to mitigate it, since four ounces (3.2 after roasting) is not much to play with. There may be some underextraction that would account for a bit of the flavor, but I don't think that's the main factor.

I will try again with six ounces to see whether I can get results more like these. However, that quantity is going to be pretty close to jumping out of the top of the roasting chamber. I may need to tip it back slightly. This is a nicer piece of equipment to use, but at this point I don't know whether it can get the job done.

Update: After a few days, no real change. It is not going to improve with age. This is just a bad roast.

sourdough bread in an oval casserole

This was the no-knead 1-2-3 technique, with all-purpose flour plus three tablespoons of wheat gluten. The main idea being tested here was the shaping of the loaf by using a different baking vessel, a glazed oval stoneware casserole, rather than the huge cast iron dutch oven. I've been consistently disappointed by the tendency of these wet no-knead doughs to produce a very flat loaf.

I had the bright idea of turning the dough right onto a sheet of parchment paper, where I thought I'd easily do the tucking and shaping. Moron. Maybe that would have worked if I'd oiled or floured the paper first. This sticky dough attached itself rather enthusiastically to the paper so rather than doing anything to worsen the situation I simply picked it up and dropped it into the casserole.

The oven and casserole were at 450. I baked covered for 20 minutes then removed the loaf and set it on a cookie sheet for another 20 to finish.

I'd have used my baking stone if it hadn't shattered beneath the last pizza I made. I did not find any suitable unglazed tiles on my last trip to Home Despot.

I cannot wait for my new camera to arrive.

The experiment was a success in terms of shaping the loaf. The proportions are just about what I've wanted. There are some odd crinkles because of the parchment paper, but I think this technique would work fine without the paper. There's heavy cracking on the top. Not sure what causes that. Failure to fold the dough into the loaf shape, perhaps? I did give it a couple of shallow slashes, so it can at least be said that they did not prevent this tragic crustal event.

It lost a bit of volume as it cooled. It's all good news from there, though: excellent moist, chewy, stretchy crumb and a delicious crust with good tensile strength but still easy to eat.  It has the usual sourdough flavor from this starter: a little tangy but not overly assertive. I'm pretty pleased with this result.

Next time I'll try without the parchment paper, and will perform the four folds to see how that affects the surface. Could also try leaving it in the casserole.

This seems like a very promising technique for shaping these soft doughs. It's a lot more convenient than using a 20 pound iron pan, too.

Also, I am loving my new nine inch Wusthof Culinar bread knife. It replaces a Henckels that was OK, but aggravatingly short. The new santoku is also a real pleasure to use. That pretty much rounds out my Culinar collection. A tomato knife would be nice, I guess.

first Turkish coffee

I've been wanting to try Turkish coffee (making and drinking) for a long time. A few weeks ago I finally made it to a local Turkish shop, Istanbul Imports, where I bought what is very commonly called an ibrik, but which the proprietor assured me was properly called a cezve. In Turkey an ibrik is something else.

I had a very good stovetop popper moka kadir roast and the simple instructions in Kenneth Davids' Coffee book to follow. I ground the beans at the finest Turkish setting.

For two servings at light sweetness I added four rounded teaspoons of coffee and two teaspoons of sugar to the empty cezve, filled it halfway with water, then stirred to dissolve the sugar.

It went onto a burner at medium heat and in a few minutes boiled. It foamed vigorously to the top, at which point I removed it and poured into demitasse cups. The foam should completely cover the coffee, and there is apparently some art to generating and/or preserving it. I don't think mine was quite right but it doesn't seem like there's that much of a trick to it. I may be missing something. It does diminish rapidly after being taken off the heat.

It was quite tasty. I think this is probably the right sweetness for me, but I'll try it at full sweetness (twice the sugar; thrice for heavy sweet) next time. This is a nice way to get a quick coffee fix without waiting for water to boil or the espresso machine to heat.

Update: I tried again at full sweetness, and also allowed the foam to dissipate and build a few times by removing and restoring the heat, a technique which I read about somewhere. I don't know if that contributed, or if the additional sugar is a factor, but I did get a more durable head. Tracey and I both thought it was very good.

premixing and storing Five Star Star San sanitizer solution

While casting about for reasons why the witbier wasn't doing what I thought it should, I wondered whether there might be a problem with the way I was using the Five Star Chemicals Star San sanitizer. I typically mix a few gallons, use what I need, then store the rest. It occurred to me that it might lose its efficacy when stored so I went to the company web site, found the contact info for the tech support and new product development guy, and shot him an email. He replied very quickly (and gave me permission to quote him):

If you are going to do this, use DI water to make up the solution. Tap water contains metals that can cause the solution to go bad over time. If you use DI water, keep the solution out of UV light and extreme temperatures, the solution will not degrade. People then ask the question, "How do I know if it is still good?"

Answer: Is the pH below 3.5 and the solution clear. If the solution turns cloudy, it is starting to go bad.

Thanks
Jon Herskovits
Five Star Chemicals

So, I haven't been using deionized water, but neither has my solution grown cloudy in the matter of days or weeks that I've stored it. Mixing Star San is very easy, so there's really not much reason for me to make more than I need other than that I do like to keep a spray bottle handy. I'll have to check the pH and see how long it keeps with my tap water. I haven't seen DI water at the couple of grocery stores I've checked.

hot air popper roasting 4

I've been experimenting a bit more with the Popcorn Pumper hot air popper, exploring its heating behavior and maximum temperature and capacity. It can easily exceed 450 degrees (the most I can measure with a compatible thermometer), several minutes of pre-heating will get it there, and behavior changes dramatically based on bean quantity.

With the same beans (classic Italian espresso blend) I've now seen that the difference between roasting two and three ounces is dramatic. With more beans, second crack actually happens, smoke appears, and surface oil becomes evident, all with very little change in roasting time. With three ounces and a preheated popper, 8 minutes will deliver FC or greater.

At four ounces, the popper sounds like it's working a little harder but still moves the beans fine. There's much more vertical action, with beans bouncing nearly all the way up to the opening rather than just swirling. This is the most I've tried, but I think it can handle more as long as the beans aren't being prematurely ejected. Further, this can even reduce the time required to achieve a given roast level.

So my initial ideas were incorrect. Adding beans, not removing them, will increase the temperature and decrease the relative roasting times. It does make sense that the cumulative effect of the exothermic reactions in the individual beans will be greater with higher volume and density. Besides bean quantity and roasting time a significant variable is preheating time. That looks like it might provide a few minutes of control over the total time. Achieving much more control would probably require modification of the popper and/or control over input voltage. Regarding the latter, I do have some crude control given that the dining room and kitchen lights are on the same circuit that I've been using. The dining room will draw 500 watts and the kitchen 450. There's a clear effect on the popper's performance!

On 2009-03-16 I roasted four ounces of Yemen Mokha Sharasi. This is no longer available from Sweet Marias, as far as I can tell. It took only six minutes in the preheated popper to hit a raging second crack and blow off quite a bit of smoke. Roasted weight was 3.2 ounces, or 80% of original.

Not that these tasting notes matter much, since my remaining four ounces are probably all I'll ever have, but they do illustrate a few things.

Shots pulled immediately were horrible. I mean, the worst espresso I've ever had. Overpowering solvent-like flavors. There was improvement after 24 hours, with the raw gasoline flavor abating somewhat. I was unable to detect any of the theoretical fruit flavors, though. After 48 hours it had become quite good, with just a faint solvent edge and noticeable sweetness and fruitiness. It was pretty good straight and very good in a milk drink. Long, mellow aftertaste. After 72 hours, the recommended minimum rest period, it was really nice. No off flavors, some pleasant sharpness, and definite sweet and fruit notes. I would put this into my regular rotation.

So the lessons are that good results certainly are possible with this popper and I shouldn't rush to judge those results.

thermal equilibrium of hot wort and cold water

Regarding my wonderment about the calculation for final temperature of a mixture of hot wort and cold water, I dug out my old Halliday and Resnick Fundamentals of Physics (second edition!) and turned to chapter 20, Heat and the First Law of Thermodynamics. It's a matter of solving for the equivalence between the heat lost by the wort mass and the heat gained by the water mass. The specific heat of wort and water can probably be assumed to be equal and therefore be factored out. If I'm thinking straight that means that

m_wo(T_f - T_wo) = m_wa(T_wa - T_f)

where m is mass, T is temperature, wo is wort, wa is water, and f is final.

Really, when solving for final temperature, this just works out to an average of the temperatures weighted by mass:

m_woT_f - m_woT_wo = m_waT_wa - m_waT_f

m_woT_f + m_waT_f = m_waT_wa + m_woT_wo

T_f(m_wo+ m_wa) = m_waT_wa + m_woT_wo

T_f = (m_waT_wa + m_woT_wo) / (m_wo+ m_wa)

As it's the proportions that matter it's fair to use volume instead of mass.

So, if the wort boils down to 2.5 gallons and is at 212 degrees, adding it directly to 2.5 gallons of 38 degree water from the refrigerator would result in 125 degree wort in the fermenter. That is still going to take a long time to cool to yeast-pitching temperature relying solely on heat conduction through a plastic bucket into room-temperature air. That's all time during which oxidation damage can occur and off-flavors can be created.

So I think the technique espoused by The Cellar's witbier instructions is suboptimal. At a 1:1 ratio of wort to water, the wort should be cooled to about 120 degrees before being added to the refrigerator-temperature water in the fermenter in order to hit an 80 degree target.

second brew (red ale), part 9

The serendipity of both the red ale and the Arizona Cardinals being properly conditioned for Superbowl XLIII (and for the first time in the history of the universe) led to the obvious christening of the second brew as Tipsy Bird.

Photoshop 7.0 does not run at all on Mac OS X 10.5 and the fine folks at Adobe refuse to fix it so I have been without a really good photo editor for quite some time. I almost reinstalled it on the last Tiger box in the house but decided instead to give GIMP a try. I've poked at it a time or two before but never had the motivation to get past its annoyances. It's still no Photoshop but it mostly did what I wanted and I became moderately comfortable with it over a couple of hours. My main gripes in this limited engagement were that window management is awful, text capabilities are very limited, and handling of selections and layers is awkward. Oh, and the first use of a drawing tool is often accompanied by an infuriating delay.

I also tried the online SUMO Paint but found it unstable enough to prevent me from discovering very much about its capabilities. And I poked at Seashore but it crashed in short order.

In any event, GIMP let me produce something good enough for this bit of silliness and now I feel a little more comfortable turning to it for quick image editing tasks.

hot air popper roasting 2

Roasted one fourth cup instead of one half, with nothing else drawing power on the circuit. I also warmed up the popper for a few minutes before dropping the beans. The behavior was just about identical to the first experiment with respect to sound and smoke (or lack thereof) however it took just 10 minutes to achieve the same bean color.

Pulled two shots after about an hour. The ground coffee aroma is notably better than the 15 minute beans. I thought the flavor was somewhat better as well although it still has a trait that I noticed yesterday: a certain underdone quality. It's a bit like eating a roasted bean plain. It is certainly a little bit more nuanced than the 15 minute roast, but still nothing like the better stovetop roasts. The next thing to try is a stovetop roast to this same degree.

I pulled two shots of yesterday's roast. The underdone flavor has lessened, as has the bitter aftertaste. However, it is still very flat. From these two samples I think I now have a sense of the effect of the long, baking sort of roast.

If Ed is able to roast 3/4 cup in 5 to 7 minutes in a Poppery II then it appears that this Popcorn Pumper may be markedly less capable, although at a sample size of one I would hardly draw any conclusions about the two models. There could also be a voltage difference; I think the acceptable variation for house current can be as high as 20 or more volts. Maybe he'll engage in some poppery swappery for a night and we can compare.

hot air popper roasting

Score! On my occasional visits to the thrift store to buy books I always look for a cheap hot air popcorn popper but have never found one. Yesterday I at last had the long-awaited fortuitous encounter and dug out the $5 bill I've been saving for the occasion.

I am now the owner of a Wear-Ever Popcorn Pumper, seemingly in great condition other than lacking the butter cup.

I read the section on hot air roasting in Davids and dove right in. I roasted the Moka Kadir blend indoors beneath the range hood, at room temperature of about 70 degrees. There was a notable gap between what I expected and what I experienced:

• The smoke, expected at 3 or 4 minutes, did not ever appear.
• First crack did start as expected a bit after 4 minutes, but never increased in frequency and occurred sporadically for the duration of the roast.
• Medium roast was expected at 5 - 6 minutes, medium dark at 7 - 8 minutes, and dark at 9. I shut it down at 15 minutes at about full city or perhaps a bit shy. If any beans were at second crack I couldn't tell.
  

That's a pretty long roast. Probably the beans can be characterized as having baked. I wonder whether the popper was up to the task. I cooled this small quantity in just a few minutes in colander and freezer.

The beans themselves are very clean, with nearly all the chaff being blown off, and the color consistency is very good. Both are much better than what I normally experience with the stovetop popper.

I pulled two shots almost immediately, although this blend is supposed to rest for a couple of days. I have never made espresso with a roast this light, I have never roasted Moka Kadir this light, and this the first time I've used the air popper. I therefore don't have much of a basis for meaningful comparisons. The ground bean aroma was not very intense. The espresso wasn't very good, frankly. The flavors are not nearly as layered or complex as in the second Moka Kadir roast (yesterday's super dark roast can be ignored). There's a lingering bitterness in the aftertaste.

With further experimentation I should figure out what effect the long roast time had. I still want to try Moka Kadir at a full city roast, both as espresso and brew.

There are two obvious things to try next time:

• Reduce the bean quantity. I roasted half a cup, which is the volume of the hopper, but I wonder whether it should be halved.
• Don't put any other load on the circuit. The kitchen lights dim noticeably when an electric griddle or the popper is turned on. I turned the lights off, but well into the roast.

2009-03-19 update: Significantly better results.

roasting measurement thoughts

As I fool around with measuring time, temperature, and roast progress I am beginning to think a little more about tools and techniques. It's a simple fact that there are some limitations with this roasting hardware with respect to repeatability and fine-grained control. Those are tradeoffs that I accept, because I really enjoy the interactivity and physical engagement. However, there are surely ways to optimize the performance within the constraints of my abilities and tools.

There are three data types I collect or employ: time, temperature, and sensory impressions.

Elapsed time can be recorded accurately. No problem there.

Indicated temperature can be recorded accurately, but what is being measured is the air temperature of the roasting vessel rather than the internal temperature of the beans. I suspect that there is significant variability in the delta between those temperatures. Factors might be bean quantity (volume), bean temperature uniformity, air exchange rate, bean density, roast stage, and surely more. Finding ways to minimize and correct for this variability may be important. Consistent bean quantity is the easiest and most obvious change.

I don't currently have a great way of measuring the starting and ending bean temperature. Troy mentioned how much he likes his Thermapen. That would be a good way to get a speedy reading during cooling. Cooling times are presently quite inconsistent, based on ambient temperature, wind, bean volume, temperature at removal, and cooling technique (colander, roasting pan, mister, freezer, etc.).

Sensory data are bean sound, smoke aroma and volume, and bean color. These are pretty subjective. Bean color can be determined only by opening the lid, which is obviously not free of side effects. Beans are hard to see through the dense smoke. Smoke aroma and volume are easy to observe. Consistency and granularity of the observations is the main challenge. The same is true of sound.

A physical challenge in all of this is the difficulty of recording the data while cranking the popper. An audio recorder would be a big help.

I'm not sure how significant a factor the cranking rate and continuity might be.

Well, I need to cogitate further and devise a plan.

roast grapher

I whipped up a little AppleScript application with XCode and Interface Builder to turn coffee roasting data into a Google Chart. As can be seen from the screen shot, the URLs are quite awkward to construct by hand.

Now with just a few seconds of data entry I have a graph of temperature against time, along with the phases of the roast (red for first crack, orange for second crack, blue for cooling).

I'll see how this works out and modify it accordingly. I think I might have it generate a table with the data as well.

magnetic spice rack

I finally finished the implementation of my new spice storage system. The objectives were to free up some kitchen space, extend herb/spice lifespan through light and temperature control, improve accessibility, and look fairly nice. Here it is, inside a cupboard door.

This is my second try. The first attempt was a flop. The original idea was to glue strong magnets to the backs of outward-facing aluminum containers with clear glass lids. The magnets turned out to be less powerful than I expected, the epoxy did not work well at all, the lids were not nearly tight enough to reliably stay in place, and baking to cure the epoxy weakened the magnets.

Now I'm using different containers, different magnets, and am not (at this point) permanently affixing the magnets to the containers. One tradeoff is that I can't readily see into these jars when they are hanging up so some of the aesthetic charm is lost, but simply labeling them should be acceptable.

Steel sheets are from Home Depot. Sufficiently magnetic stainless steel containers with clear plastic lids and really nice silicone seals are from The Seattle Restaurant Store. Epoxy-coated rare earth magnets are from Applied Magnets.

Magnets placed in groups of four deliver very firm attachment. Even two will do but four makes me quite comfortable. I can pretty much slam the cupboard door closed without worrying about anything being dislodged. I probably won't store my liquid mercury or finely ground plutonium here, but the mass of any common spice poses no problem at all.

The whole thing is dynamically reconfigurable, just not quite as quickly as if the magnets were glued to the containers. On the other hand, the containers remain unblemished and can still be used in the rotating metal racks they came in. And it's always going to be easier to glue than to unglue, should the need arise. If the containers weren't sufficiently magnetic I'd just go ahead and glue the magnets and be just as happy.

Lessons learned from the first try:

• Neodymium magnets can be weakened or demagnetized by high temperatures.
• I would use JB Weld if I were to again glue magnets to metal.
• If the containers are going to be in a horizontal orientation then the lids need to screw on or have very tight friction fittings, or they will sooner or later pop loose and provoke foul kitchen language.
• The ineluctable fate of glass lids is to fall and shatter.

Some of the original containers are at the top left. They'll be coming down now.

This post is linked on

converting a stovetop popcorn popper for coffee roasting

When I began roasting coffee experimentally I considered the two standard low-end devices: a hot-air popcorn popper and a hand-cranked stovetop popcorn popper. I couldn't find a suitable model of the former in a nearby store or thrift shop, and I also liked the idea of being a little more hands-on, so I went with the latter.

My first attempt was with an aluminum Whirley-Pop Stovetop Popcorn Popper. I won't go into the modification details, other than that I followed Kenneth Davids' instructions in Home Coffee Roasting, because I don't think it's worth bothering with. The agitators are flimsy, the bottom deforms constantly so the agitators get stuck, and worst of all the gears are plastic. It didn't take long before I was highly annoyed, and just a bit longer before it didn't work at all. Don't use one of these.

The Back to Basics Stainless Steel Stove-Top Popper was a little more expensive but of far higher quality. There are two basic modifications you must make: install a thermometer, and replace the plastic viewing window. The tee-nut method described at Sweet Maria's is perfect. I think this is a much sturdier technique than using a spring clip, which I found unsatisfying on the Whirley-Pop. I used this Update International 550 thermometer.

The popper has a plastic viewing window that I imagine is nice for popcorn but which is rapidly rendered opaque when roasting coffee beans, so it's functionally useless. The first time I roasted, this window began to sag a bit. The second time, while the popper was pre-heating (perhaps for a few minutes longer than necessary), it melted entirely leaving a mess more difficult to remove than Tyrone Willingham.

I had on hand some thin galvanized sheet steel, which was readily snipped into the necessary semicircular shape. This must be folded over so that it can be held in place by the restraining tabs of the popper lid. One way to do this is to rest it on a workbench so that it hangs over the edge by the desired amount and gently tap it with a hammer to create a 90 degree bend. Flip the piece over and continue to tap from one side to the other until a fold is complete. Fit it to the lid, drill a hole through it for the screw that holds the wooden handle in place, and you're in business with a somewhat less transparent but significantly more heat-resistant viewing window.

I've now used the finished product for a couple of years and dozens of roasts with nary a hitch. I would highly recommend this approach if you're considering the stovetop method.