---

The amount of heat available from the TO is closely related to the overall volume of the roast chamber for best roast results

Greater volume will result in baking not roasting
To small a volume and the roast will burn

So look at my thread for all the details regarding ideal volume

KK

---

Indeed, volume is a key factor, especially for a majoritar convective roasting, where heat losses are significant, and walls are mandatory to be insulated.
And also the agitation is critical, I got uneven roasts easily until the paddle was been finely tuned for the available RPM.

For IR roasting, because energy hits the beans directly, insulation is less critical, and roast is much more uniform, if the coffee bed is not higher than 3-4 layers, which equates to one pound greens in a 23 cm dia roasting area.
When roasting using IR mainly convection role is more to ensure an uniform ambient, hot air is a lot less hotter than in case of convection roasting, at least 10 C degrees lower, that can be seen on chaff color, honey gold, instead dark brown.

---

Lamp to beans is 12.7 cm. The screen is 90+ % open design with a very thin wire and openings around 4.76mm wide. There will be no appreciable loss of IR pass through with this screen. I've removed the stainless shroud.

I'm not sure if I mentioned this but I have a radiant heater below the bean disk so if needed, I can also add some under-bean heat if the roaster lacks sufficient heating power.

Interesting observations you've made with aerodynamics of vortex in a cylinder produced by a turbine.

---

So you will have "contact" transfer too
I'm envious ...

---

Ok, diameter of pot is 10". Depth from top rim to bean disk is 7.5". Comes out to 8.013 liters. in reading one of your earlier threads on TO design, ideal volume is 7-9 liters.

---

That volume should be ideal at 8 lt
Don’t forget to calculate the TO dome as part of the roast chamber

KK

---

Installed bean dump door and ET/BT sensors, BT thermocouple is attached to curved return vane and is hooked up to TC4 for monitoring time/BT/RoR, ET sensor hooked to Fuji PID controller and roast is controlled via programmed ET ramp from 300 F to 470 F in 5 minutes resulting in a 8.5 minute roast. Without sound absorbing layer attached to the stock pot exterior, I'm not able to hear first crack but am ending roast via bean temp which is giving a reliable end of roast indicator. First roast cup quality is excellent at 1 hour post roast. I will report back on cup quality as I run more beans through it.

---

Ok, second roast, a nice colombian, didn't turn out so nice. Very flat and underdeveloped in the cup and noticed the usual nice aroma of the beans when transferring to a mason jar after roasting was missing. So, need help with best practice for ET in a turbo oven roaster that uses halogen lamp for heat source. ET sensor is 2" above bean mass, 2" out from stainless pot wall

Here's my ET/BT results from the roast after preheat at 150C for 3 min.

min 1 ET 137 BT 88
min 2 ET 149 BT 114
min 3 ET 174 BT 134
min 4 ET 198 BT 150
min 5 ET 218 BT 167
min 6 ET 232 BT 182
min 7 ET 245 BT 195
min 8 ET 246 BT 206
min 9 ET 246 BT 212
min 10 ET 246 BT 216
min 10:10 ET 246 BT 218 End of Roast. Level of roast = medium roast.

---

What was the green bean weight ?

I normally pre heat at full heat limit of the TO

And lastly - we don’t talk as much about another, but most important part of the roast
That is the degas period
Depending on the bean origin - The degas period ranges from 5 days to 14 days before the roast can develop its flavour

KK

---

How was the second roast different from the first (excellent cupping), as timing and power levels?

Are you reducing power when browning starts ?

Attached are two roasts of the same Eth coffee, with or without lamp shield plate attached.
You can notice that the IR version has 5 C lower ET, and needs about 5% power reduction in the Maillard phase, even if the top power for both roasts in the dry end is close, in the 76-77% ballpark.
~~~

---

I'm doing all of my testing at 1 lb charge. I can try higher temp preheat but I find the roaster more than capable of moving the beans through any fast profile desired, in fact, much faster than desired if so programed and this is with a much lower preheat temp.

As you and I have been roasting for literally decades, no need to discuss optimum post roast wait. Most of my roasts cup very nicely within hours of a roast and do not change that much each day after roasting.

What do you find is a minimum roast time that still gives you optimal results?

---

For second roast, I slowed rate of rise slightly during drying phase to give it at least 4 minutes from start of roast to end of drying. Otherwise, pretty much the same curve. EOD for my setup occurs at BT of 152 C. From your experience, what is the minimum time you've found to reach EOD in a TO roaster? I'm wondering if I'm going too fast through this phase.

I did not reduce power when browning starts as I noticed a naturally occurring falling RoR without lowering power level but will give that a try on next roast.

---

Minimum was around 4:30 for washed coffees, because so is programmed my roast automation.
But I noticed that some naturals are drying very fast, around 3:45, with the lowest limits I programmed into that roast logic.

---

I normally do 1 kg green now
Roasting for drinking
Minimum roast time is 9 min to first crack
However I much prefer longer at around 12 min to first crack

Roasting for Cupping
Around 5 min is what I prefer

KK

---

Found one of the primary problems after re-reading your post on IR/convection roles. Instead of using power levels and watching beans RoR, with my setup having radiant as the dominant heat source, I was primarily using ET as my guide which did not correlate with changes in beans RoR. Couldn't figure out why no matter how low I kept ET at start of roast, the RoR stayed higher than desired. This makes sense now and I will have to scrap my Fuji ramp soak PID controller and switch to my large variac and will mark power levels on the dial faceplate in %. Controlling a roaster that has radiant as the primary energy transfer is challenging.

---

My preferred time to first crack is closer to 10 min on my drum roaster so we've got similar roast stage times. I will experiment with these times on my next testing runs.

Is your heat transfer primarily radiant with convection coming in 2nd?

---

Because the fan is on from start to finish
It’s convection all the way

KK

---

If the lamp shield plate is removed, and distance is less than 10 cm, the equation changes significantly... radiation becomes at least equal to convection, if not prevalent.
Measurements shows this very well, in the case of radiant roasting, in the browning phase the RoR skyrocket if power level is preserved all the way the same.

---

Since last post, I've reconfigured various parts of the roaster and am working out details for a variable speed drive motor to allow a more gentle agitation which will also lessen noise. RoR will be monitored using a TC4 but will be controlled via changing power % throughout the roast.