Environment temperature effects chemical reactions within the bean and the amount of energy available for the roasting process. With most sugars hitting their melting point in the 320-400F range and many of the chemical reactions that happen within the bean occurring in the 356F range I would assume that you would want your ET to drop no lower than the 360F range, 375F being a safer marker at the expense of sugar development.

At the same time the cellulose of a coffee bean is damaged at 446 to 536F. At that point the cellulose matrix of the bean breaks down and you loos complexity and physically damage the bean. That would suggest you do not want your MET to rise above the 450-500F range with 560+ essentially nuking the bean and you have charcoal briquettes.

The best cup characteristics are created when the degradation of Trigoneline (melts at 424F with most of the compound burning off at 445F, this is where a lot of bitters come from) and the derivation of Nicotinac Acid (which melts at 457F and is responsible for acidity) are linear. I found a note that taking these variables and adding in the transfer efficiency of the roaster you get your Best Reaction Ratio with an ET of 401-424F.

Now armed with that information, I would think the magic to creating a controlled dark roast would be not allowing your MET to exceed 446F which is where you start to actually burn the bean verses roasting. At the same time, creating a light roast you should not allow your ET to drop below 375F or you will stall the chemical reaction happening in the bean with the 'safe zone' for your ET being 401-424.

Now, in the process of roasting, and controlling the first to second crack transition and overall bean development I want a BMT max of say 420F (just a random number). To get a slow bean mass increase to develop the bean complexity and caramelizing its sugars without burning off any acidity I would need to maintain a minimum ET of 375F'ish to prevent stalling the roast. If I allow my ET to drop to say 350F for a moment even though my bean mass temp continues a slow but steady incline to my target point, how much damage would be done to the roast during that short time? Inversely, how much damage is done when the MET is allowed to momentary exceed the cellulose safe mark of 446F.

What is your time tolerance range before you do harm to your roast?

Thanks for this information. I've been thinking a lot about the question of max MET since I'm trying to roast fast and light to preserve acidity, but, of course, fight with the problem of scorching from overly high METs. Establishing a max MET seems to be really important, but it does vary depending on roaster and bean type... Realistically, based on the 500 max MET I use for high-grown, washed coffees, I have trouble finishing a roast in less than around 10:30 if I follow Rao's 20% development ratio, though I should be able to get times down to <10:00 with ~1 minute of development.

So my concern is with your max MET figure. If I maxed out MET at 446, I'd be likely to stall during 1C. Generally, I risk stalls below around 450-460 F during 1C on my roaster.

Your discussion of min. temp is quite interesting as well. It seems that the Nordic profiles tend to be quite linear - which means slow start, fast finish. In this case, they're certainly flirting with your minimum acceptable ET figure. This allows more time for "drying" and a faster and lighter drop (as light as ~+15 F over 1Cs). On the other hand, Rao seems to advocate a hot charge and fast dry to get the bean center up to temp, but I have not been very successful with that approach (undevelopment with a drop of +20 F above 1Cs, 2 min. after 1Cs; acceptable development at +25 F over 1Cs, but that's too dark for my palate).

It's probably not the answer you're looking for but I'd recommend that you experiment and see first hand what the effects are of dropping ET slightly while BT continues to rise.

According to Rao's philosophy, one of his three commandments is "the bean temperature progression shalt always decelerate." A rise in ET without a subsequent rise (or leveling) of the BT RoR does not violate that commandment. Also keep in mind he says that the commandments aren't written in stone and in some cases may be bent or even broken. The truth is always in the cup.

It's been my experience that subtly influencing ET for short bursts of time , a technique I've heard referred to as "feathering," is more effective than than big, aggressive changes. For example, at first crack it's standard to back off the energy to coast through the beginning of first crack then adding energy again partway through first crack to "regain control" of the roast (a W. Boot theory, I believe). To achieve that, you can increase energy for 10 or 15 seconds then back off for 10 or 15 seconds. You may try that style of feathering and see if it allows you to change ET slightly while not drastically impacting the BT RoR.

I often will use feathering to try and hold ET in the range of 401° - 424°F for the same reason. I'm a big proponent of Staub's theory. But having said that, you should also remember that ET readings will be different between roasters and even within the same roaster. An ET reading taken at the very top of the drum will be higher than ET read just above the bean mass.

A little off topic, but it would be better if we all took temperature readings from multiple locations and averaged them, then used that average when we talk about ET or BT, rather than using only readings from a single point. It'd be great if logging apps could do that by allowing one to identify probes as either ET or BT and then take readings from each and calculate an average.

mcwresearch wrote:A little off topic, but it would be better if we all took temperature readings from multiple locations and averaged them, then used that average when we talk about ET or BT, rather than using only readings from a single point. It'd be great if logging apps could do that by allowing one to identify probes as either ET or BT and then take readings from each and calculate an average.

Can be done in Artisan. Just record several "ET" signals using some Extra Devices and add a symbolic formula that takes those values and computes the average live. Easy.

Can be done in Artisan. Just record several "ET" signals using some Extra Devices and add a symbolic formula that takes those values and computes the average live. Easy.

Great! I recently ordered the TC4 so while I'm in there I'll add more probes and give it a shot!

MaKoMo wrote:Can be done in Artisan. Just record several "ET" signals using some Extra Devices and add a symbolic formula that takes those values and computes the average live. Easy.

I am doing bean mass, ET measured at mid drum and exhaust temp at the intake of the exhaust port and have one channel left for future use. I was thinking of using it for ambient air temp. The Phidget 1084 has an on-board ambient temp sensor as well so you could get 5 inputs if you could access that on-board TC for logging in Artisen.

I was asking about this because I just went through a major roaster upgrade from my trusty HotTop. I have a lot of tools at my disposal now so I am trying to wrap my mind around all of them so I can make the best use of the information.

Hi Dave - You have thrown a lot of information out there most of which pertains to the Agtron Scientific Roast Method AKA Kinetic Function Roasting. The exerpt on the SM web is a bit dated and much has been changed although the basic method has not. I attended Carl's roasting workshop in Reno at Agtron in 2007 and subsequently purchased a 3 K Diedrich which has been re-fitted for this roasting method. A picture is worth a thousand words ....

Most of that is from an Agtron article which I knew was old which is why I asked. I am only 4 or 5 roasts in on my new roaster but starting to get the hang of the controls and avoiding the EKG looking ET. I added an exhaust TC right at the drum exhaust opening so I am logging the bean mass, ET at mid drum, exhaust at the exhaust output which is top of chamber and the ambient temp for good measure. My last roast looked much better.

Looks like you are killing the heat at charge to stretch the drying phase. I ran a lower gas but closed the damper down to 10% to keep the ET level. Then once I had the drying finished increased the gas flow and opened the damper to around 25% to slightly increase the ET and get a steady rate of rise up to first crack. Then lowered the gas and opened the damper to around 35% during first. Then increased the heat for the stretch with no damper adjustment to give the bean adequate development time. I will be cupping this one tomorrow. That was only a 215g charge so I ran my ET in the low/mid 400F. A larger charge takes a higher ET.

Lot more control then my HotTop. The first couple roasts were all over the board. I was use to adjusting for the lag with the electric heat. Gas is almost instant so I have had to get use to making smaller, more controlled adjustments through the process.

Hey Dave,

I just spent 45 minutes replying to your post and when I uploaded an image, something flushed my efforts and left me in another thread. $%^#@**&&& ARG.... Not the first time this has happened. Maybe I'll have the patience to try again tomorrow.

Did 2 roasts this afternoon to the full 3 kilo capacity. Times to first varied by 12 seconds and 1 degree. Dropped immediately at the end of first (same temps both times) and total roast times were within 9 seconds. Pretty consistent. It would be very easy for you to implement this system on your gas roaster.

You would need two correctly placed sensors, a PID controller, and a modulating gas valve.

Let me know if you want to know more. Cheers - Bill

HB somehow loaded the first picture twice - I guess I made a mistake. Should have been ....