Fresh Roast SR800 adding Phidget, Artisan

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
User avatar
Supporter ♡

#1: Post by MNate »

I purchased:
+Fresh Roast SR800
+12" Razzo Roast tube
+Razzo larger chaff collector
+Phidget items:

(in retrospect I think a single TC, the 20cm one, would have been plenty. And I haven't used the mounting nut)

1. Install 20cm probe from top:
A. Drill from the bottom (metal screen side) a hole with a 9/64" bit, just off center. (I chose to go from the screen side to better line up my hole and because I didn't want to end up pushing the screen away from rest of the piece, even though going from this way likely makes a rough exit the other side.)

B. Drill from the top with a 5/32" bit just part way in. This allows the probe to wedge in nicely at that little thicker part. However, I could gain a quarter inch if I made the hole a little bigger so that the next thickness when in a little instead.)
This gets the 20cm probe about 6.5" from the bottom of the 12 inch tube.

2. Install 11cm probe in chaff collector:
Basically just use a punch to make a hole in the screen in a top exhaust spot and wedge it in there. It works well enough.

3. Install Artisan and set it up for Phidgets. This was very easy and there are lots of instructions on it. There are a few things to get right, I guess.

4. I think it's worthwhile to also set it up to manually input the Air and Heat settings so that when you look back at a roast you see what you did and how it may relate to what happened with the beans. I wrote up a bit on how to do that here: Artisan Setup without Input Devices I also made the lines thicker for my heat/air by going to Events -> Style and then making Thickness and Size up to 6. You can play around with it though to display how you like.

First, how do the various temps relate? The readout on the SR800 is much much higher than either of the probes. Before installing the probes I was worried about the crazy high temps I was doing. Now they are more in line with what I read elsewhere. The probe up in the chaff collector goes similar to one closer to the beans but different enough that I think the bean one really is more reflective of the bean temp. The chaff one starts at a lower temp but then you can see it cross and ends up a good 25 degrees higher than the bean probe. I think I could do with just the bean probe. Once I get things so I can relax a little as I roast I'll also log what the SR800 has for temp along the way so someone can see how it relates.

Second, I do wish I could get that bean probe a little lower... With the 9" tube it might be perfect - roughly 3.5" off the bottom. When it starts, it is sitting about 2 inches from the top of the bean mass. As the beans expand it is sitting nicely in the top of the rolling mass. So maybe that's fine, getting a better reading when the readings are more important.

Third, I still have trouble hearing first crack (at least the end of it) and second crack... loud fan, roasting outside...

Fourth, I need to figure out how to do better at a declining RoR! But I do think the graph helps a lot with it so I'm glad I have the Artisan stuff working!

Thoughts? Anything I should have differently?


#2: Post by hercdeisel »

Man, this is an awesome write up! A few comparative notes from me:

First, I'm using the 9" tube for whatever differences that makes.

Second, I've got my thermocouple hooked up to a phidget as well and I think it's super easy to do. I didn't want to drill into anything so my alternative was to thread the tc cable through one of the notches in the chaff collector extension screen, where the top part of the chaff collector is supposed to slot in. You can still lock the top down tight as usual but the notch that doesn't have the tc cable running through it will occasionally let some chaff through. So, roasting is a bit messier but not much. Since I'm not using a rigid cable to drop through I have to get the right bend in the cable to make sure the probe sits in a consistent position in the bean pile which isn't too hard to do. The probe generally starts out sort of just into the top of the bean pile and by into FC it's completely submerged in the rotating beans for ~175g doses.

Third, I was manually writing down base read out temperatures and comparing them with the probe temperatures and the temperature gap between the base and the probe really depended on how aggressively you turn down the fan to build the heat in the roast. The gap can be anywhere from 30-80*F gap for me depending on how late in the roast and how aggressively I've been adjusting the fan down. In general I found that the highest I ever wanted or needed the base temp to get in order to get into a dark roast was about 480-485. Getting much above that is going to get a burnt or even ashy taste to it I found. In general it seems like a good idea to hit FC around the 450-460 mark on the base temp and then gap between those two temperatures will just keep narrowing more slowly.

Fourth, as for RoR...I'm still thinking through this a lot but here's where I've currently landed.

It seems pretty tough to get a smoothly declining RoR on these roasters if you're measuring temps precisely and not artificially smoothing them out in the software. The issue is that the controls are fairly coarse grained and adjusting the fan in particular causes significant temp changes in terms of the air coming into the chamber There's simply no way to drop the fan without getting bumps in the measured RoR. The general slope is definitely down but each time I drop the fan I get a little hill even if I also drop the heat (and dropping the heat actually seems worse because that'll lead to a steeper drop once you hit the peak of the RoR bump caused by your fan drop).

Related, I think part of the reason for this (and I'd be happy to be corrected on this!), is that I don't think it's possible to measure bean temp on these machines in the way that drum roasters measure bean temp. In a drum roaster that is properly dosed, I believe the beans are relatively tightly packed around the probe even though they're constantly moving around it. There's also (I'd guess) less change in the air that fills the gap between the beans in a drum roaster than in the fresh roast tubes (and probably in fluid bed roasters generally). In the fresh roast tubes large amounts of air between the beans is constantly changing over and is always freshly heated. So, for all those reasons, I'm guessing the temperature measured by a probe even in the bean pile of a Fresh Roast is generally higher than the temperature measured by a probe immersed in a drum roaster bean pile at any given stage of development because it's picking up so much more hot air than a well setup a drum roaster probe. As I said above, generally I get first crack (some nice loud snaps) when the base measures 450-460. Usually at that point my probe would be reading 400-410 (I've never seen a FC in any bean below like 398) but from all the graphs I see people posting around here they get FC far before those temps. (the probe doesn't seem to have a calibration problem as it's usually within 1 degree of several different thermostats I've measured it against when checking ambient temps). Taste and roast level confirms similar results.

While thinking through these observations, I've been trying to reduce the number of changes I make to heat and fan and let the natural temperature gradient from bean to air temperature take care of the smooth declining RoR in the beans themselves. The fewer temp and fan changes you make (my thinking goes) the smoother the declining RoR you'll generate (even if it's not easy to see that!). So, my rules of thumb the last few weeks of experimenting have been:

1. Aim for ~4min dry. To achieve that, I'll preheat the roaster to about 350 on the base readout (usually 9F fan, 7Heat gets there in a minute or two), then drop the settings to 9F, 2H and add the beans. I leave it for 3min. After 3min I raise the heat to either 7 or 8 depending on the level of roast I'm going for, 8 for dark and medium, 7 for light.

2. Depending on how long I want the ramp phase to last, I'll adjust the fan down more or less aggressively for the next 4 minutes (less time between fan adjustments for shorter ramp, longer time between adjustments for longer ramp). From there I'm trying to hit about 460 on the base temp by the time I want FC to happen. The trick, I've found, is reaching those temps by dropping the fan in increasing increments without dropping the heat (until maybe towards the end of the roast if I'm risking getting over 485). The lowest fan setting I'm going for is 6. So one common way I get there is by dropping to 8 at 4:15, 7 at 5:45, and 6 at 7:45. But this is where most of my adjustments are. Medium roast levels are achieved by going for about 2min post FC and 9ish minute total roast time. Darker roasts happen by going to 3min dev time and about 10min total roast time. Changing the fan just a few times like that without playing with the heat will keep the RoR for the probe temp in the chamber declining at a steady pace (aside from bumps from fan adjustments) and presumably the beans are riding a similarly declining RoR. The only time I adjust the heat setting is to drop it towards the end if I'm risking going past 485 or so on the base readout. That's about 4 total adjustments (one to bump heat and finish dry, 3 fan drops) in a roast.

User avatar
Supporter ❤

#3: Post by wingnutsglory »

Thank you both for taking the time to write up your setups and experiences - it's much appreciated to have such detailed accounts! I've been looking at doing something along these lines since getting at SR800 and 9" chamber.

For hercdeisel, it sounds like you're possibly using a wire thermocouple (vs the kind encased in a long stainless steel probe). Is that true? I do like the idea of using a "probe type" TC like MNate is using but I really want the tip to be embedded in the bean mass, and I'm not sure I'd have that until later in a roast with smaller quantities (e.g. 150g or less).

I think you can get longer TC probes pretty easily but I was hoping to buy everything from one spot.

User avatar
Supporter ❤

#4: Post by wingnutsglory »

hercdeisel wrote:Man, this is an awesome write up! A few comparative notes from me:

First, I'm using the 9" tube for whatever differences that makes.
Also, when did you receive your FR? The first one I received (May?) had an odd readout that always read much higher than I had expected. I found that after a replacement unit was sent (fan had stopped working on the original), the readout showed about 30-40deg less consistently. I think I'd read other reports of higher than expected temps. I'm asking because I've not seen a readout of 450 at FC on any roast so far.

User avatar
MNate (original poster)
Supporter ♡

#5: Post by MNate (original poster) »

Interesting... very interesting.

I do imagine hercdeisel has a good route. I'll try this in a day or two and post my results. Consistent heat setting does seem right to me. And yeah, adjusting the air is troublesome...

We'll get there. I'm enjoying the roasts I did the other day despite not being textbook (or actually being textbook examples of problems!)

I'm rigging up a stand too! I'll show that off soon, I think.

User avatar
Supporter ❤

#6: Post by wingnutsglory »

I tried hercdeisel's profile tonight with an Ethiopian DP (Shantawene) and got pretty good results (I think...munching a bean isn't always the best predictor)...possibly the best yet for this bean. We'll see!


#7: Post by hercdeisel »

wingnutsglory wrote:For hercdeisel, it sounds like you're possibly using a wire thermocouple
Yep, when I say 'probe' above I just mean the 1.5" or so end that does the measuring. The rest is a stainless braided cable. I definitely would prefer something rigid and fixed so I wouldn't have to fiddle with the cable each roast to get it to a consistent position.
wingnutsglory wrote:Also, when did you receive your FR? The first one I received (May?) had an odd readout
Oh that's super interesting. I got mine in mid April this year so I could easily be suffering from that kind of issue! I guess at this point I've done the translations for myself to reasonable temp levels.
MNate wrote: I'm enjoying the roasts I did the other day despite not being textbook (or actually being textbook examples of problems!)
Yeah, that's the thing...I don't mean to imply that the difficulty in generating a smooth declining RoR curve in any way means I'm unhappy with the coffee I've been getting. I've really enjoyed the coffee I've been roasting and have found been trying to get a better and better grip on the input variables I can control to create different roasts. For me, for right now, that's length of development time (easiest to control!), the heat setting for ramp and development phases, and the rate of dropping the fan speed to modulate the length of the ramp relative to the development.

It feels like a really awesome tool to use while learning about roasting...the only downside is I don't know how I'd start roasting without being able to see the coffee so easily!

And seriously, thanks again Nate. This is such a great writeup and how to to get the roaster to the next level to be able to genuinely monitor temps in a pretty damn precise way.