Kaffelogic for the USA is on Indiegogo

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
Frenchman
Posts: 402
Joined: 14 years ago

#1: Post by Frenchman »

There's very little information on it here, especially when it comes to comparing with the Ikawa offerings. I hope we can start a discussion and maybe be lucky enough to have Australian users provide feedback on their units. The thread I found here was pretty short.

https://www.indiegogo.com/projects/kaff ... oming_soon
LMWDP #712

LuckyMark
Posts: 211
Joined: 4 years ago

#2: Post by LuckyMark »

This post (no 60) by Luca gives a decent rundown - Some lessons after using a fluid bed roaster for two weeks

Fair to say others love it but they may not be as experienced or discriminating as Luca

User avatar
luca
Team HB
Posts: 1134
Joined: 19 years ago

#3: Post by luca »

Crikey, they're not shy to lay it on thick! Check this out from the indiegogo:
No matter what brew method, Kaffelogic will roast perfectly to your needs.
They haven't exactly left themselves much room to move if you're unsatisfied with it! Maybe they could at least edit this claim to change "will" to "can"?

Anyway, I just popped in here to add that my bad experiences with the KL may well have been due to wanting to get aromatic and clean light roasts; it may be that people that don't mind a bit of roast character in their coffee have had more success with it. I gave up on it because I was only borrowing a (very generous) friend's, so it wasn't really worth my spending any more time or effort on it (though I would have bought one had I had more success). I should also add that I did have some exchanges with KL that suggested that the calibration in the unit that I was using may have been off; I went and tested it against a unit from a local distributor, who happens to be a friend of mine, and indeed the calibration was a little off, in that my roasts on the same profile as the reference unit seemed slightly more developed. My distributing friend offered to recalibrate the unit, but I doubted it would have made any difference (since I had tried a wide variety of settings that would have spanned whatever the calibration would have adjusted the profile to), so I didn't trouble him to do it.

My friends still have KLs, so if anyone happens to have some profiles or whatever that work well for aromatic and clean light roasts, I'd love to pass them on to them so that they can try them. (Having said that, if you are going to say something to the effect of "this profile duplicates the graph you get from a drum roast" or "Scott Rao says that you should have a constant linearly declining ROR and I have cleverly made a profile that has a derivative that does that", I'll let you know that it seems like every few months, a breathless newbie emerges, Prometheus-like, to grace us with these revelations.)

Like I said in the other thread, the machine has heaps of power; it wouldn't surprise me if you can manipulate it somehow to get roasts that you like, and I hope that people get roasts that they like at a reasonable price with the machine.
LMWDP #034 | 2011: Q Exam, WBrC #3, Aus Cup Tasting #1 | Insta: @lucacoffeenotes
★ Helpful

User avatar
ducats
Posts: 141
Joined: 9 years ago

#4: Post by ducats »

Got a USA beta version about a week ago. Half dozen roasts into it or so. Here's 4 from yesterday from same bean. Data exported out of Kaffelogic Studio and imported to Artisan to make this infograph. Delta Span for BT is 30 seconds, this is one noisy probe, exposed k-type in the bean mass. 90g batches with about 12% loss +/- 0.2. Fan calibration done for these beans, 0.98.

All roasts were of SMs Kenya Nyeri Ichuga Factory from '21. It's a mix of PB and AB, some user discretion needed for calling 1C but I tried to be consistent.

Roast A was a "learn the probe" roast and manually ended. Roast B was a modified version to get 4 | 3 | 1:10 for about 10F. Roast C was the same goal but with a linear fan profile, default fan profile decreases at about 3:30. Roast D was a fine-tuning of Roast C.

I plan to cup tomorrow but I won't false-promise that I'll come back with an expansive write up for that. But I thought I'd get some data out there. Still early stages...


If anyone knows how to change the labels for the 3 axis to white please let me know.
luca wrote:My friends still have KLs, so if anyone happens to have some profiles or whatever that work well for aromatic and clean light roasts, I'd love to pass them on to them so that they can try them.
What's your definition for light?

User avatar
luca
Team HB
Posts: 1134
Joined: 19 years ago

#5: Post by luca »

Oh cool, this is great. I guess one thing that would be cool to know for 110V people is how fast you can roast with the lower voltage. If you had some junk coffee that you were happy burning it would be good to see how fast you could roast something. I know in the other thread I posted some roasts around 6.5-7 mins, but at the moment, I'm leaning towards the view that you get better filter/cupping roasts in these sorts of roasters, with these batch sizes, at quite a fast roast time, like 5 mins. You'll see that tonnes of Ikawa profiles are around that time, Matt Winton's profiles are around that time (or even shorter) for Roest, I think TW was also lower in roast time. The idea with these roasts is to have something that is very acidic, very sweet, very clean, zero bitterness, zero roast flavour (eg. char or toast/bake) and well-developed acidity. My usual roast defect is underdevelopment.
ducats wrote:Roast C was the same goal but with a linear fan profile, default fan profile decreases at about 3:30.
This was one of the things that sort of did my head in a bit about the KL. They say that the airflow must change because you need more air to yeet the green around when it is dense and unroasted, then it decreases so that it isn't hurling the coffee around too high later (perhaps hurling the coffee too high when it's less dense would loft it into lower temperature zones, too). But if you have a temperature reading of X at an airflow of Y and a temperature reading of X at an airflow of 2Y, presumably the latter has a greater capacity to impart thermal energy into the beans because there is more mass of air at that temperature. But I can't remember any discussion of this point, and when I bring it up, I usually get people scratching their heads and looking at me like I'm a martian.
ducats wrote:What's your definition for light?
Ideally maybe just a tad more developed than Passenger, Sey, and (maybe) Apollon's Gold; I guess ideally Tim Wendelboe and Manhattan filter or just a tad lighter. That's probably stylistically the most useful touchstones, hopefully you've tried one or more of those?
LMWDP #034 | 2011: Q Exam, WBrC #3, Aus Cup Tasting #1 | Insta: @lucacoffeenotes

User avatar
ducats
Posts: 141
Joined: 9 years ago

#6: Post by ducats »

luca wrote: If you had some junk coffee that you were happy burning it would be good to see how fast you could roast something. I know in the other thread I posted some roasts around 6.5-7 mins, but at the moment, I'm leaning towards the view that you get better filter/cupping roasts in these sorts of roasters, with these batch sizes, at quite a fast roast time, like 5 mins.
I do.

5 minute roasts remind me of my poppery roasting days where all I had to lengthen the roast was a combo of extension cords. Even then I don't think I ever saw 5:30 or longer. From those roast I remember their peak flavor lasted maybe 48 hours.
luca wrote:You'll see that tonnes of Ikawa profiles are around that time, Matt Winton's profiles are around that time (or even shorter) for Roest, I think TW was also lower in roast time. The idea with these roasts is to have something that is very acidic, very sweet, very clean, zero bitterness, zero roast flavour (eg. char or toast/bake) and well-developed acidity. My usual roast defect is underdevelopment.
With the Kaffelogic the charge temp is the ambient temp. Is that how you did your profiles? I talked to them about this a while back and they said forcing charge temp to be ambient temp was the only way they could get back-to-back roast consistency. But a 5 minute roast would probably benefit from a high charge temp.

luca wrote:This was one of the things that sort of did my head in a bit about the KL. They say that the airflow must change because you need more air to yeet the green around when it is dense and unroasted, then it decreases so that it isn't hurling the coffee around too high later (perhaps hurling the coffee too high when it's less dense would loft it into lower temperature zones, too). But if you have a temperature reading of X at an airflow of Y and a temperature reading of X at an airflow of 2Y, presumably the latter has a greater capacity to impart thermal energy into the beans because there is more mass of air at that temperature. But I can't remember any discussion of this point, and when I bring it up, I usually get people scratching their heads and looking at me like I'm a martian.
I think I follow you. Normal fan would be like a normal garden hose, and double fan would be like a fire department hose. Even though the hoses output water at the same temp the fire department hose has more "velocity" for a lack of a better word. Most of my roasts have been on a modified Poppery 1 popcorn popper with wiring split off for independent heat and fan controls. What I found with that setup was that a constant fan setting gave me my best results. Too much fan at the end would "strip" the coffee and make it hollow. Too little amplified heat adjustments and made it hard/impossible to control Development RoR. Fluctuating during the roast always seemed to have a negative impact, I would guess from altering the convection/conduction ratio. As an aside, the current default Kaffelogic fan profile starts at 14700 and ends at 13200, an 11% drop so nothing drastic as far as I can tell, but the audible difference is noticeable. Lowering fan at the end should allow the bean mass to move the same at the start and end but it should also lessen the demand on the heater and make it easier to roast darker, for whoever wants to do that :wink:
luca wrote:Ideally maybe just a tad more developed than Passenger, Sey, and (maybe) Apollon's Gold; I guess ideally Tim Wendelboe and Manhattan filter or just a tad lighter. That's probably stylistically the most useful touchstones, hopefully you've tried one or more of those?
I was drinking Wendelboe's Finca Timana while I wrote my previous post. I had some Sey last year, and I should have some April coming today. One thing I haven't figured out is if these 5ish minute roasts are used for production roasts at the "nordic" roasters. So when you get their roasts and compare them to your own 5 minute roasts it's not really apples to apples.

I found this ROEST profile for Matt Winton. https://www.roestcoffee.com/roestblog/d ... att-winton Since you have a ROEST can you look at it in greater detail and tell me exactly how long development is and how many degrees? The article says you can copy the profile to a ROEST machine.

I roughly made the equivalent in Kaffelogic Studio, and theoretically I think I have enough power.
I did it in Celsius and RoR multiplier is at 20, so at first crack the RoR is 6.25.
Not sure if pic will be legible, so...
Dry End at 1:55
First Crack at 4:15
Drop at 5:02, though I'd probably manually override it to drop at 5min exactly. 3.9C increase in Development.
*The fan curve is linear though the graph makes that look like a lie.
I should be able to run this within a couple days.

Milligan
Supporter ❤
Posts: 1502
Joined: 2 years ago

#7: Post by Milligan »

luca wrote: This was one of the things that sort of did my head in a bit about the KL. They say that the airflow must change because you need more air to yeet the green around when it is dense and unroasted, then it decreases so that it isn't hurling the coffee around too high later (perhaps hurling the coffee too high when it's less dense would loft it into lower temperature zones, too). But if you have a temperature reading of X at an airflow of Y and a temperature reading of X at an airflow of 2Y, presumably the latter has a greater capacity to impart thermal energy into the beans because there is more mass of air at that temperature. But I can't remember any discussion of this point, and when I bring it up, I usually get people scratching their heads and looking at me like I'm a martian.
That had me thinking a bit... I need to check my Ikawa curves to see because Ikawa profiles nearly always start off high fan then lower throughout the roast as the beans dry and the roast slows. I would think that with a decreased airflow the delta between the inlet air and the exit air would be larger for the same energy imparted to the beans compared to high air flow. As in, the inlet air would need to be hotter at lower air speeds to maintain the same temperature rise in the bean mass. At least for the Ikawa, if I ran a profile and then did the same profile again but lowered the fan, I suspect the Ikawa would increase inlet temps. Would be interesting to try.

I'm only speculating here but that is interesting to think about.

User avatar
luca
Team HB
Posts: 1134
Joined: 19 years ago

#8: Post by luca »

ducats wrote:With the Kaffelogic the charge temp is the ambient temp. Is that how you did your profiles? I talked to them about this a while back and they said forcing charge temp to be ambient temp was the only way they could get back-to-back roast consistency. But a 5 minute roast would probably benefit from a high charge temp.
Yeah, the machine auto cools, or at least it did when I had it, so you can't really start with a hot roast chamber. I mean as long as you are starting with a chamber that is the same temperature each time, maybe it can provide consistency. In hindsight, I don't think I ever did a whole bunch of roasts back to back to check consistency and repeatability on the KL. So there's a job for you if you're looking for more work ;P
ducats wrote:I think I follow you. Normal fan would be like a normal garden hose, and double fan would be like a fire department hose.
Maybe a better example is frying. Let's say you get a really heavy, massive, cast iron skillet and you heat it to, say, 250C. And you also get a piece of aluminium foil and heat that to 250c. Then you cut the heat and put a steak on top of each. Both are at 250C, yet clearly the greater mass of the cast iron skillet is going to cook the steak more than the foil.
ducats wrote:Too much fan at the end would "strip" the coffee and make it hollow.
Thanks for the reminder. I have heard people say this, but haven't methodically looked into it myself. My current roest profile is high fan, because I wanted to try to get it working more like an air roaster, but I ought to drop the fan and keep everything else the same and see what, if any, difference it makes. There's always some additional experiment that you haven't done in the world of coffee.
ducats wrote:I was drinking Wendelboe's Finca Timana while I wrote my previous post. I had some Sey last year, and I should have some April coming today. One thing I haven't figured out is if these 5ish minute roasts are used for production roasts at the "nordic" roasters. So when you get their roasts and compare them to your own 5 minute roasts it's not really apples to apples.
The ~ 5 min ikawa/popper roasts are not the roast profiles that nordic roasters are using for sample roasts. I think that TW's production loring roasts are in the 9-10.5 minute range, give or take, and his sample roasts on the roest are in the 5-6.5 minute range. I don't know what's relevant to compare, really. I mean, colour is probably the most important thing to compare. Of course, the first thing every breathless newbie does is to try and make a graph that looks like the graph from a commercial roaster that they have seen. I think kaffelogic have some profile called "firestarter" that was one that Rob Hoos came up with at a conference when he basically tried to get the same roast time as a drum roaster and first crack starting at the same time, so you can look that one up and use that as a starting point to form your own opinions if you want to look into trying to replicate drum roasts.
ducats wrote:I found this ROEST profile for Matt Winton. https://www.roestcoffee.com/roestblog/d ... att-winton Since you have a ROEST can you look at it in greater detail and tell me exactly how long development is and how many degrees? The article says you can copy the profile to a ROEST machine.
Oh yeah, this is a good point. So I have used matt's profiles and they are power profiles, from memory; they're just power at X% for like 4 mins and then X-Y% for the next step. So they don't use any temp input at all. Plus the roest uses "air temp" as the control point, which is different from what the KL uses. Let's not get into that. I think I asked Matt a bit about his roasts a little while back, actually. Anyway, I think I was stopping the roasts maybe 50s after FCS, but I don't know that that was necessarily right. I think TW's rule of thumb for the roest is that he likes about 45s DT for naturals, 55s for pulp naturals and 65s for washed, but that may be on a slightly slower profile. As for temperature ... well ... that's kind of a tricky question. The answer is actually that it depends which temperatures and probably which coffees. For example, on the roest, for some reason, with most kenyan coffees, once you hit first the BT does not go up, no matter how much heat you put in. But for colombian coffees, the BT might skyrocket.

Anyway, as you point out, the objective of the exercise isn't to get good coffee; it's to find out how much power the 110V KL has and if power might be a limiting factor, so you don't really need a great profile for this. If you just make it slightly unrealistically hot, you should be able to get it to run and then what will happen is towards the end if there isn't enough power, the machine will just give it max power and the power will flatline, like in the graph I've posted below. Maybe just make sure the fan speed isn't too high so that it's not excessive fan speed that is preventing it reaching its targets. So what we want to know is if you can get a roast of less than, say, 5 or 4.5 minutes and have the coffee sort of popping along aggressively as opposed to barely cracking. If it can do that, then the voltage on the US version probably isn't going to be an issue. If it can't do that, then there might be some roast styles that the 110V KL struggles to deliver. Of course, there's a complicating issue here, as always, that I gather the KL is quite sensitive to local voltage variations at each individual place at which it is used.


At least for the Ikawa, if I ran a profile and then did the same profile again but lowered the fan, I suspect the Ikawa would increase inlet temps. Would be interesting to try.
Good point; I should be able to generate some information on this, too. I could run a few "air temperature" profiles with the same air temp, but different fan speeds, and log inlet, air, outlet and bean temperatures and then cup the results. I'll add that to the list of experiments I'll probably never get around to!
LMWDP #034 | 2011: Q Exam, WBrC #3, Aus Cup Tasting #1 | Insta: @lucacoffeenotes

User avatar
ducats
Posts: 141
Joined: 9 years ago

#9: Post by ducats »

Quick reply.



Did a 5 minute roast. Manually ended it at 5 min but Studio says it ended at 4:50, not sure what's going on there.
Fan Calibration at 0.98 and fan profile is linear at 1370
In: 90g Ethiopia Worka Wuri from SMs July '21; washed; Appearance .2 d/300gr, 14 Screen
Out: 81.2g
%loss 9.8
Dry End at 1:52
1 Crack at 4:07
End at 4:50, but see above. Went 5.8F in Dev.




Artisan of same roast



I felt better about calling first crack with this bean than the kenyan of the above post. So first crack at 4:07 into the roast. Max power consumption looks to be at 1:44 measuring 1.315kW....I think it's supposed to go up to 1.400kW.

User avatar
ducats
Posts: 141
Joined: 9 years ago

#10: Post by ducats »

ducats wrote:I plan to cup tomorrow but I won't false-promise that I'll come back with an expansive write up for that.
Responding to myself. Did some cupping. Linear fan for the win. Hard edge to default fan profile and roasty aftertaste. Linear fan was more transparent, sweeter and with a cleaner finish.

Post Reply