And how are you going to implement this camera in a heated environment? 200C? you must stick the lens inside.

Also about roasting color on the outside/whole bean, think about a bread. You put it in the oven, you burst the heat to max with air/ventilation, the outside gets done fast it looks crispy but the inside doesnt inflate and remains a bit behind. The same applies to coffee, coffee color is important, but it wont give you good roasts just like that.

you think there is any merit in coupling a lens like this to the little sightglass on the roaster - or this wont give us any meaningful data ?

I am sure some of the more experienced folk here are familiar with this, however thought I would share these sensors here:

https://www.tipteh.rs/pdf/Micro-Epsilon ... OR--en.pdf

Some of them can measure via fibre optic cables and I am sure I have seen high temperature fibre optics somewhere. I will dig them down.

How do you think the commercial products work? Maybe a little triple-paned window does the trick. Or you may resort to a probe lens; but that's a little more pricey than the camera module.

https://www.laowalenses.ca/product/24mm-f-14-probe/

I have a prototype from Madison Instruments that's gone through a few updates. I don't think they've turned it into a proper product yet, but hardware-wise it's basically a camera connected to a Raspberry Pi doing some calculations with a pretty short program written in Python (which I've modified extensively). There have been challenges, but I've gotten some interesting data out of it. I wouldn't want it as my primary data series to decide when to end the roast, but as another data series for trying to figure out what's going on in a roast there's some real value there, especially for decafs. Some day I hope to have the time to ditch the Pi/Python side of this and get richer data out. If building something similar yourself, my advice would be to not underestimate thermal issues as the early iterations had big problems with the sensor saturating.

I'd imagine that getting consistent illumination would be a pretty big problem with chaff and smoke flying around. I'd be amazed if it were possible to get anything like a tonino type measurement accuracy in this environment. Even disregarding price, as a pure experiment, I'm pretty skeptical - consider how dark some naturals can be on the outside, even when roasted lightly. Commercially producing, and supporting such a system is a bit of a nightmare in my opinion, I'd be very surprised if we see widespread adoption.

N3Roaster wrote:I have a prototype from Madison Instruments that's gone through a few updates. I don't think they've turned it into a proper product yet, but hardware-wise it's basically a camera connected to a Raspberry Pi doing some calculations with a pretty short program written in Python (which I've modified extensively)..........

Neal, This camera is "focused" at the window or the lens is inside the beans like a thermocouple?

Is this good enough? saw this 1 year ago but forgot about it:

OP mentioned colortrack. I think they are $7K+ to get going. There is a monthly fee one that is cheaper but no clue what the monthly rate is.

stoned_survival wrote:Neal, This camera is "focused" at the window or the lens is inside the beans like a thermocouple?

I've been asked to not say too much about the inside of the optical portion so I won't go into depth on what all is there, but it's in a tube that's a few inches long mounted outside of the sight glass, which on a Diedrich IR-12 is pretty usable for that purpose.