FreshRoast SR800, Phidgets, and Artisan

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
travis_rh

#1: Post by travis_rh »

Hey!

I've been a lurker for a little while, and finally wanted to get some feedback on my roasts and have some discussion around roasting on a FreshRoast SR800.

I'm running two thermocouples to Artisan and logging those values. Here's a photo of my setup:



I'm getting good BT curves, but some of my roasts are coming out horribly unevenly - all splotchy and mottled. A washed Guatemalan that I had was the worst, but I'm seeing it to a lesser degree in other roasts. Obviously, air flow is the key element to balance - enough for good bean movement, but not so much that you inhibit heat transfer.

Here's an example of that Guatemalan:


I'm also struggling with a smoothly declining Rate of Rise. I tend to get large increases in RoR when I lower the fan setting throughout a roast. That's all based on washed coffees. I have a natural Ethiopian, and while I've only done 3 batches, smoothly controlling RoR has been horribly difficult - much harder than the washed coffees I've roasted.

Here are a few roast curves:
The uneven Guatemala (Proyecto Xinabajul Chalum from Sweet Maria's)


A not too bad Rwanda (Nyamasheke Gitwe from CoffeeShrub / Sweet Maria's)


Two of the three not so great natural Ethiopian batches (Hambela Goro from Sweet Maria's)



I've tried batch sizes from 159g up to 215g. I've settled on the smaller side at 170g (basically, 6 ounces) in an attempt to improve airflow and bean movement at lower fan settings.

I've played around with a few different approaches, but generally I leave the heat at 9 for the entire roast and start with the fan at 9. Then I back down the fan one setting at a time, about every minute or so, until I end up at 6 for a 170g batch. I found that if I stayed at a fan setting of 7, I could get to FC, but the roast would often flat-line.

I was aiming for about 17-19% development time, but found those roasts to be super light and a little under-developed (though, I don't know how accurate any of my numbers are). Now, I'm aiming more for 19-22%. Generally, I'm seeing about a 11-13% drop in weight.

Here are my questions:
  • What size batches are you generally using?
  • How are you minimizing RoR spikes that result from lowering the fan?
  • What, if anything, do you do differently based on processing method (washed, natural, honey, etc.)?
  • How short is too short of a roast? How long is too long? Depending on how I was messing with heat and fan settings, I was hitting FC at ~5:00 or 5:30 on some roasts, but can lengthen out to as much as 6:30 or 7:30 before hitting FC on others
Artisan
  • The Delta Span setting in Curvers >> Filters is what controls the time interval for Artisan to calculate RoR / ∆BT, correct?
  • And to make sure I understand the math correctly, it's averaging the rate of change over that 15-second span (or whatever Delta Span is set to), or is it simply calculating the delta between the data point at 0 seconds and the data point at 15 seconds? What's the default setting?
Thanks!

MikeTheBlueCow

#2: Post by MikeTheBlueCow »

So, I have an SR700, but can answer some of the questions.
travis_rh wrote:
  • What size batches are you generally using?

    800 users might answer this better, but with the 700 I can at least relate that more weight means needing to keep a higher fan and will lead to higher temps (versus the same settings with a lower weight).
  • How are you minimizing RoR spikes that result from lowering the fan?

    This is going to happen no matter what. You should be seeing big changes with fan and heat setting changes. It's a biproduct of the fact that the probe is sensing air temp as much as it's sensing bean temp. You have to adjust fan to help drive the roast, so I don't think you'll ever get rid of this, you just have to learn how to read the graph while ignoring the humps because they're "fake". This is when it's important to remember that a pretty graph does not necessarily make a tasty coffee. You have to use intuition and your knowledge of what is occurring in the roast, what is causing the readings, etc.
  • What, if anything, do you do differently based on processing method (washed, natural, honey, etc.)?

    Good question, and I think one that is hard to get great information on with air roasters, as most of the info is for drums and not really applicable. For any coffee, I'm adjusting my starting heat setting, and simply learning by watching the graph to see what changes I should make when.
  • How short is too short of a roast? How long is too long? Depending on how I was messing with heat and fan settings, I was hitting FC at ~5:00 or 5:30 on some roasts, but can lengthen out to as much as 6:30 or 7:30 before hitting FC on others

    I have found a longer dry phase beneficial (4-5 minutes), then the maillard and development phase times I'll adjust based on the bean. I'm usually doing in total 8-10.5 minute roasts for a City to City+, and I would venture into 12 minutes for City+ to Full City style roasts would be acceptable. In all, I've found a 6 minute roast is a little quick and unevenly developed.
Artisan
  • The Delta Span setting in Curvers >> Filters is what controls the time interval for Artisan to calculate RoR / ∆BT, correct?
  • And to make sure I understand the math correctly, it's averaging the rate of change over that 15-second span (or whatever Delta Span is set to), or is it simply calculating the delta between the data point at 0 seconds and the data point at 15 seconds? What's the default setting?

    After playing with settings I have mine set at 10s. You don't want any Smoothing, and I would recommend changing the delta span timing just to minimize it being too jumpy, but sensitive enough that you can react with settings changes. 15s is fine (in fact I'm going to try my next roast out on that), and it does depend on your probe and sampling rate, I believe, as to what the ideal settings are.
By the way, I find the unevenness of your roast very similar to what I would see on something like a natural Ethiopian, and something I have seen improvement on with longer dry times, but it's never a fully even roast when going for City to early City+. I would focus on making fan changes solely to keep a good, fluid, even bean movement and then adjust heat up as necessary, and that should improve the evenness of the roast a bit as well. If you see the beans just start jumping/hopping up and down, you need to lower the fan immediately, as that leads to the bottom being heated more than the top and pretty clear uneven roasting. Also longer development times really improve the evenness. I would target 20-25% DTR as a minimum, and find that helps improve the taste as well by preventing some of the under-roasting flavors. By playing around with drying time, I found that *total roast time* is also a key factor of development, and that it's best to add the extra time to the drying phase vs after the crack. Especially because I see you hitting 300F BT pretty consistently at 2min, that's much too early IMO.

MNate
Supporter ♡

#3: Post by MNate »

travis_rh wrote:Hey!


Here are my questions:
  • What size batches are you generally using?
  • How are you minimizing RoR spikes that result from lowering the fan?
  • What, if anything, do you do differently based on processing method (washed, natural, honey, etc.)?
  • How short is too short of a roast? How long is too long? Depending on how I was messing with heat and fan settings, I was hitting FC at ~5:00 or 5:30 on some roasts, but can lengthen out to as much as 6:30 or 7:30 before hitting FC on others
Artisan
  • The Delta Span setting in Curvers >> Filters is what controls the time interval for Artisan to calculate RoR / ∆BT, correct?
  • And to make sure I understand the math correctly, it's averaging the rate of change over that 15-second span (or whatever Delta Span is set to), or is it simply calculating the delta between the data point at 0 seconds and the data point at 15 seconds? What's the default setting?
Thanks!
Glad to have you joining us. I definitely think it will take a bunch of us trying a bunch of things to really get some understanding.
I can't answer many of your questions, and I keep thinking I'll post an update to my process on the other SR800 thread I was using... on the other hand maybe some of these need their own thread so it gets manageable unlike the massive Robot thread... Anyway...

I'm on about my 4th different approach with the SR800 (my first roaster), and I have the 12" tube so my method is likely a little different. I tried the "Roast at heat 9" method and quickly didn't like it. It was too fast, gave me too many scorch marks, and if I went with a lighter roast I got splotches like you are.

I just updated my approach on that other thread: Fresh Roast SR800 - 1/2 lb Air Roaster but just say here that I think you're roasting much too fast - try 6 or even 5 instead of 9 so that your roasts take 10 or 11 minutes instead.

But other answers to your above... I'm keeping my fan setting lower so the beans are just shuffling around nicely, but never getting any loft. So I turn the fan down more quickly. I find the roast is even this way with no scorching still.

But I'm still learning and experimenting!

travis_rh

#4: Post by travis_rh »

Thanks for the input!

So, it sounds like lengthening the drying phase and using a little less aggressive heat would be good starting points? Also, I think MNate kind of clarified this one for me, but you're both aiming for fan settings that result in an obvious turnover of the bed, but without beans jumping off of the top of the bed?

MikeTheBlueCow, is there anything special about 300F in particular? Do you aim to reach 300 in a certain amount of time?

Development time is also something else I'm trying to improve. My tastes lie in the City to City+ area, and I have no interest in going anywhere near 2nd crack. With that being said, trying to get good, even development without the roast flat-lining and baking is somewhat difficult. Any tips there? Do you aim for a specific rate of rise or overall temperature change from FC to dump?

MNate
Supporter ♡

#5: Post by MNate »

travis_rh wrote:Thanks for the input!

So, it sounds like lengthening the drying phase and using a little less aggressive heat would be good starting points? Also, I think MNate kind of clarified this one for me, but you're both aiming for fan settings that result in an obvious turnover of the bed, but without beans jumping off of the top of the bed?
I don't know if it's right! But yes, that's my new theory. I have heard echoes of it among the masters too... but much of it seems theory.

And for your later question, I'd first try just doing a single heat setting for your whole roast and see what things look like. Maybe a 5 first since you want a lighter roast. Need to stretch things further do a 4. Etc. from there maybe we could decide we want more or less heat in a given phase so we bump it up or down during that phase. I haven't found much stalling this way through the lighter roasts.

MikeTheBlueCow

#6: Post by MikeTheBlueCow »

I have Artisan auto-mark DE at 302F, just to keep things consistent. I see my beans Yellow at anywhere from 305-345 depending on my fan and heat settings. I've looked into discussions here and it really seems to me that actual "Yellowing" is actually the start of caramelization, but Maillard reactions start before that, at 302, so I'm just marking that instead. It is still influenced by the heat and fan settings, but there's not a lot to do about that unfortunately.

Yes, I keep my bean movement so that it's turning at a quick pace. I don't mind if it's a little aggressive, but it is important to prevent the "hopping" that you'll see, where the bean mass really splits in two and a whole bunch hop up at once while the others stay at the bottom of the chamber.

With my approach (and I'm not saying I'm getting perfectly wonderful and amazing roasts, it's still a challenge, but I'm dealing with some extra issues due to my unit and the way it's controlled)... Since I'm drawing out Drying, I am starting low heat and max fan. I up the heat a couple times, down the fan when it needs it, and continue like that until I increase the heat to a max setting (probably about 80% of the max power) and then only lower the fan as needed. The heat reaches it's max setting at about Yellow. The fan gets down to maybe setting 3 by around FCs then I let it ride. I'm not sure how well this transfers to the 800, since that has a more powerful fan. But I'm mostly just paying attention to Artisan to guide when to make adjustments, based on ROR.

Bluenoser

#7: Post by Bluenoser »

In your picture, it looks like some of the unevenness is because some of the chaff is still sticking.. I get this on certain beans... but there is some obvious differences in roast levels as well. I find the chamber for the SR800 was a bad choice.. it is too wide and too thin. I just roasted some decaf Peru beans at 15C outside temp and I had to turn the fan down to 3 to get some good heat build up, but then the beans just don't move.. And I use a 4-5 minute drying phase. I can do most washed fine.. Brazil Santos works well. I'm finding that the Razzo roasting tubes might be the only way to overcome this design flaw. (However, Razzo Roasting is on vacation now). There is another glass tube that FreshRoast makes that goes inside the stock tube that might help give you more even roasts.. but its not an elegant solution. I wish FreshRoast had charged more and done a better job with creating a tube that would promote more bean movement and improved heat retention. I tried insulating the bottom half of the tube with some ceramic paper, but it didn't have much of an effect.

I agree that the ROR curves are pretty hard to get BT on in an air roaster. I currently don't have external probes on my SR800. I want to display the curves on iOS devices so can't use artisan.. Still researching some app (RoastMaster?) to use.