Dr Gary, congratulations on getting such a good result against professional roasters

Luca, I just wanted to thank you for your straightforward and honest opinion of equipment, especially the Kaffelogic. I roast with a Coretto and was looking for something with greater repeatability and automation. Considered Bullet, Kaleido, Ikawa and Kaffelogic. The Kaffelogic seemed - via reviews - to be a reasonable choice so proceeded to read everything on it. It was yours and SteveC comments on both the Kaffelogic forum a Coffeesnobs which were very helpful. It was even helpful just to read the help (or lack of it) on their own forum. That said there seems to be many happy users, I think your best advice was to find someone with similar tastes and see what they use. I wouldn't pretend to have a palate like yours and like more medium coffees, but as I was purchasing this just for the repeatability I am grateful for your freely given advice and thoughts. Thank you very much for everything you do for the coffee community.

LuckyMark wrote:Dr Gary, congratulations on getting such a good result against professional roasters

Actually, believe it or not, I'm actually under-selling Dr Gary on this one. We went to a cupping workshop thing that Dr Gary organised with a pro roaster who generously hosted us and put some of their own coffees on the table, as well as opening it to the floor to bring in their own roasts. My absolute favourite coffee was from another home roaster; my second favourite, not far behind, was Dr Gary's. And the generous pro whose coffees were beaten? None other than famed international roasting consultant Rob Hoos at Nossa Familia. Now, to be fair to Rob and Nossa Familia, from memory that is exactly the result that should have happened and is not a knock on Rob or Nossa Familia's coffee because the green that Dr Gary and the other roaster (whose name I can't remember) had used was very high quality delicious washed Ethiopian coffee, whilst the commercial samples were green that would naturally score lower. And I guess the further caveat for people reading along is that these were home roasts on small commercial drum roasters, so I'm sorry if that kind of puts a bit of a dampener on it!

LuckyMark wrote:Luca, I just wanted to thank you for your straightforward and honest opinion of equipment, especially the Kaffelogic. I roast with a Coretto and was looking for something with greater repeatability and automation. Considered Bullet, Kaleido, Ikawa and Kaffelogic. The Kaffelogic seemed - via reviews - to be a reasonable choice so proceeded to read everything on it. It was yours and SteveC comments on both the Kaffelogic forum a Coffeesnobs which were very helpful. It was even helpful just to read the help (or lack of it) on their own forum. That said there seems to be many happy users, I think your best advice was to find someone with similar tastes and see what they use. I wouldn't pretend to have a palate like yours and like more medium coffees, but as I was purchasing this just for the repeatability I am grateful for your freely given advice and thoughts. Thank you very much for everything you do for the coffee community.

Thanks, but I'm just an opinionated nut job who knows what he likes and knows what he hates. I never built a full on corretto, but I had a lot of fun fooling around with heat gun dog bowl. Actually, a friend of mine who is an engineer made what I guess, in hindsight, was sort of a souped up coretto for me - he split wired a heat gun and we used that as the heat source for my Quest M3, not using the element in the Quest at all. So I guess the quest drum was basically working like a bread maker, to some extent. My friend even split wired the HG so that we could control it via TC4 and artisan. We actually thought that roaster worked really well, but it was like a benchtop full of breadboards to pull out and pack away every time, so that was way too much work, it was next to impossible to get good probes into it, and because we totally redid the airflow path, basically the chaff all lodged outside the drum, so we'd have to strip it down and vacuum it out every roast session. So we abandoned it, but not for lack of roasting quality or ability, funnily enough.

All of which is to say that the mighty heatgun seems to have tonnes of power and the capacity to roast to whatever profile you want, but the difficulty is getting information and controlling it repeatably. Which is kind of funny, since I also think that the KL probably has the capacity to deliver whatever sort of roasts you like, but I also found it difficult to control.

Morten Munchow did some research to show that colour is probably the biggest predictor of roast qualities; I think that this is probably the best article of his on that point:

https://www.mdpi.com/2306-5710/6/2/29/htm

So it's good to see that KL have the clear roast chamber now so that you can at least see the roast colour. That has to give you a much better chance of getting the first roast right by being able to manually end it by looking at the colour, rather than not being able to see the roast progression and just having to guess. Which always felt pretty stupid to me, when this seemed to be a solved problem on pretty much every other roaster around, usually just requiring some glass in the machine somewhere.

Luca, thanks very much for that link, very interesting. A fairly detailed piece written so well that even someone like me can follow it. Confirms that you need to see the colour of the beans!

Your M3 story would bring a wry smile to anyone with a homebuilt roaster, after a few times of adding more devices and pulling out all the breadboards and associated paraphernalia I can see the allure of a store bought solution would be strong, especially something like the Kaffelogic. Nice to hear that you were able to make the M3 consistent and a good roaster. I hope your Roest is going gangbusters for you, the fact you bought one made me look up their price. I looked fondly at my Corretto after that

Yeah the roest felt kind of ridiculous when I got it, and I certainly did not pay full price - I had a nice opportunity come up to buy it from a friend, kind of stemming from over 20 years playing around in coffee. Very lucky! Much about the machine is overkill, and I think that part of our thinking in getting it was just that we have all used so many roasting machines that have some sort of niggling irritation to them (to say the least) that we were all a bit exasperated. Yes, my friends and I very much felt that there is actually so much work in engineering all of this stuff well that we kind of looked at how much work it would have been to fully mod the quest to get it to work how we wanted that buying a used roest started to look reasonable. I also live in a very small apartment, so I wouldn't be able to house something like a nice small gas drum, which may be available cheaper.

A very good example of the work that you either want to do properly, or have done by the manufacturers for you, but never be surprised by, is probes. Scott Rao will tell you about the importance of this at length. You can't just assume that any temperature reading is useful just because you have it. One of the many details about probes that is a total PITA is electrical interference, which can screw up the readings. Roest have been great to deal with recently; I asked them about working something out on temperature and got a custom probe from them. They made a custom non-conductive ceramic housing for it. It turns out that their engineers had a quest as well and when I praised them for how much work had obviously gone into a simple probe, they confessed that they have spent hundreds of hours just working on probes.

One thing that is really important is repeatability. The graph I'm posting below is a little hard to tell on, but it's actually a graph of five roasts. It's hard to tell that it's a graph of five roasts because the three temperature lines are pretty close to carbon copies, because I wanted them to be. The bottom blue line is the external stationary drum temperature probe. Using this, I think I have been able to get to a repeatable starting position. The drum temp probe sort of follows bean temperature. The red is BT, but how well it correlates to a traditional drum BT I don't know, and perhaps I don't care, as long as I can get results that I want to. The DT and the BT are dependent variables. This particular profiles is controlled at the top blue point, which they call the air temperature, which is the independent variable that I controlled (as well as RPM and airflow, but I kept those constant to make things easy). Let's not get into too much detail about air temperature at the moment. The point of this graph is that I was able to roast two coffees on one day and decide that I liked one roast enough to repeat it on another day (you'll see that there is one roast that went much longer than the others that I didn't like). Many roasters tell you that the first roast is always slower when the machine is colder, but with a proper warmup and between batch protocol, I think I was able to eliminate even that as a variable. Setting aside whether or not these roasts were any good, the point is that I was at least able to repeat them easily. The web portal enables the roast profiles to be edited easily, so if I have repeatability, I can work on the roast profiles incrementally. I like to think of my roast style is progressively failing in the right direction. The ability to have this level of confidence in repeatability, and to be able to get this data fairly easily, and to be able to tweak the roast conditions pretty easily, and to not have to spend hundreds of thousands of dollars and tens or hundreds of hours drilling, tapping and fitting probes, are all features of roasters that are worth something, I think.

Luca, you are right I could not pick that graph as 5 roasts till I read your text and expanded the graph. That is impressive repeatability! To be honest if you bought the Roest at somewhat less than retail it makes it look very reasonable compared to the other alternatives, it has a lot going for it. Nice they will make custom probes, that and their attitude just adds to the ownership experience.

Looking good and congrats on the new rig. Any chance you could repost that comparative roast graph but change temp axis to 0-300 and RoR axis 0-25 and get rid of all of that text with the black background around roast end?

Sorry, this is about the best I can do easily. Can't redo ROR axis, can't reduce temp axis to max < 325C. (I had 500C max for the experimental inlet temp probe.) I found an extra roast that I did on this profile, so this is now six roasts. I also added exhaust temp as well as "air temp". But I guess the point that I want to make is that this isn't necessarily a good roast, and I don't know that any of it would really translate over to much out even if it is. The point of this graph is really about the repeatability of the roasts, based on the probe measurements that the machine gives you. (I'm taking it that this isn't too off topic, since this is basically some lessons I've learnt from using what I'm pretty much thinking of as a fluid bed roaster at the moment ;P)

I understand how it shows repeatability, something I'm hoping to improve with my own roasting. Those large axis ranges can "smooth" data a bit and I just wanted a closer look. Thanks.