So as mentioned before there appears to be some sort of temperature sensing device built into the base that interacts with the Atmel processor. There appears to be potentially power dimming circuits in there as well but I need a multimeter to verify all that which I do not have at the moment.
I’ve mostly been roasting the past few days “by sight, smell and sound” the various coffees. I logged some details on some of the coffees I roasted and they turned out ok when I brewed them but without temperature data to go with my timings I figured reporting the results was pretty useless so I stopped mentioning the other beans and what I roasted.
Initially my plan was to roast the coffee and see which ones tasted good or “bad”. So far the only one that I have had any issues with is not necessarily “bad”…. but just not “really good”… it is only good. That of course is a horrible horrible crime in the world of coffee. [/sarcasm]. Even the “only good” coffee will be given a chance to redeem itself after reading the details provide on SweetMaria’s regarding the coffee and seeing if there are any nuances of needing to be darker or lighter or perhaps made into espresso or left for resting for additional days etc before I decide where it belongs in my ordering of new coffee for the “longer term” than the initial sampler.
I have now set up a digital gram scale (sensitivity 1.0g) with a small plastic measuring cup that I “tare” on the scale prior to pouring the beans in. I’m finding that the amount is sensitive for only a couple beans that seem to roast “weird” compared to the rest where the others do just fine with a higher amount of beans. These beans that are “weird” have little movement until the last 60-90 seconds of the roast.
The factory suggested amount is 120 grams while Tom @ Sweet Marias suggests 100 as being more realistic. While I cannot claim to be any sort of coffee snob and certainly not an expert at roasting/cupping etc I’m finding the coffees I listed in my first blog… by starting them on a completely level surface, and running the fan at 100% with the heat on low until a slighter amber color is achieved will result in pretty decent coffee with no more than one or two dark “spots” beans. It seems on very small beans that apparently get stuck in the mesh before the fan starts. They will end up scorching but the rest of them circulate quite well and explain why the “stirring trick” works due to this taking place while the fan is blowing. Once you stir all the beans free from the screen they begin to free-float in the stream like the larger ones that never got stuck. Everything else seems to roast gradually and without signs of burning though heavier beans will start to darken faster due to spending more time lower in the chamber. I believe the roaster being level to be quite important as a result and starting off with a lower temperature and the “cool trick” artificially keeping the temperature low in the startup to help limit any burning. This would side with people that declare a pre-heat cycle with the FreshRoast to be unnecessary since it would probably burn the beans worse than starting off slowly. Ultimately a much lower temperature than “low” perhaps by around 100 degrees would be better as a “dry” cycle.
Today I managed to find a temperature probe in the grill section that makes it into the upper 400s Fahrenheit. Most thermometers stop at 392 degrees or they have a lot of plastic that would prohibit you being able to have them near the heat of a roaster. It is a BBQ grill thermometer made for Kenmore and was sold at my local Sears store. It does NOT fit the roasting chamber without some convincing. Convincing involves some very gradual bends being applied to the thermometer probe to allow it ot sit further down in the chaff collector putting the tip in the middle of the roasting chamber. To convince it to turn I ended up utilizing the combination of a small gap on the rounded handle on a vise for some MINOR adjustments and the major adjustments being convinced with a large Maglight flashlight. If you have ever bent metal electrical conduit you will get the idea. You need to gradually support the pipe as it bends sit that it does not kink. The temperature probe is actually a hollow shaft with a pointed tip on it. The hollow shaft contains a set of twisted wiring that is actually a thermocouple. It connects to the braided wire cover for that wiring which leads to the socket on the display. You don’t want to kink the probe so it has to be slow bends in the probe to keep it from crimping on the thermocouple and potentially weakening things.
I have not roasted a batch using the new thermometer yet due to the fact some things came up today where I was not in the house and as a result we didnt really drink much coffee today to use the beans we’ve already roasted leaving me nowhere to place the finished roast. As a result I now have a 2 day rest on most of these beans that are in there. Today I get to see if the probe will pickup any appropriate temperature readings that match up with what would be expected at First Crack and/or Second Crack or if it is simply not sensitive enough and is actually picking up the roaster air only.
I have already attempted to run the roaster without beans for a brief time to inspect the temperature readings and see if there is any sort of ramp involved. The temperatures did not seem to pass the mid 300s no matter what the setting while on high. Apparently with that level of airflow (if things are working with the sensor) the heating elements needs to have back pressure from the beans in the roaster to reach the 400s. As a result if running too small of a batch it will likely not roast properly thus never reaching a high enough temperature like too much beans could result in too high of a temperature being applied. With the use of the Atmel processor and that possible NTC Thermistor Diode it probably means that the heater will cut out to not result in an overheat situation temperature wise BUT ultimately the beans in the roaster will probably burn without enough movement.
FYI the Atmel datasheet HERE shows that the processor has interrupt capabilties. The buttons seem to be wired to two of the pins that have interrupt capabilties. Due to there not being any chips to buffer the button pressing it would be my guess that the program made any pressing of Up/Downl to be an interrupt to the existing programming. It seems to be part of the “Analog Comparator” which treats the status of two pins as a logical on/off. If you press up it counts that button as pressed (1 or on) to the other button (down) as not pressed (0 or off). It compares the two settings and initiates an interrupt. As a result whatever signals being received for fan or heat control would be interrupted resulting in a “drop” to a minimum level that we are seeing. This interrupt stops normal interpretation of the voltage in/out on the other ports and triggers a counter to go up or down. There is circuitry built in apparently to do limited “debouncing” which is why holding the button does not make it continue to go up or down. With the use of more expensive complex circuitry the variation in speed that we see would go away but due to cost savings this is a side effect. The only way to eliminate such a pause is to use a different controller than the Atmel or potentially include extra chips to handle these changes. The length of the pause is directly related to how quickly you can trigger the button. I would suspect if you had “two bare wires” instead of the push button and you quickly tapped them together rapidly the up/down reaction would not be noticeable. Due to the type of button used there is a long “click” in the button that makes the interrupt obvious. The faster the contact and release of the button the shorter the interrupt allowing the circuit to continue on as programmed.