Nordic Light

[nrdlnd]

Hi,
This is still my favorite profile. I'm working with it and I'm doing some adjustments. The NordicLight4 had a higher fan speed in the beginning of the roast to make the movement in the beginning better for bigger beans. I also lowered the batch size to 90g and lowered the preheat to avoid burning the beans. At the same time is often a higher preheat advantageous to get a "good start" of the roast. I modified the NordicLight4 with a even higher fan speed in the beginning but also a higher preheat (1000W). It's the same bean as above an Indian washed Kalledavarapura. These beans are rather big (compared with for example the smaller washed Sidamo). I call this profile NordicLight5:

NordicLight5.kpro

(3.33 KiB) Downloaded 291 times

This is the result:

230120_IndiaKalledavarapura_NordicLight5_2.5.png (76.63 KiB) Viewed 5477 times

Movement in the beginning of the roast was ok and better than before the modification.. I marked 1st crack a little late so DTR is more than 20% and the bean is fully developed but it's a very light roast.
I did also roast the same Sidamo Gr2 bean as above and used the same NordicLight2 profile as before with this bean. It had the same 1000W preheat as the Indian bean with NordicLight5. I blend these beans 50/50 and it's a very nice blend in my opinion. This is the result:

230120_SidamoGr2_NordicLight2_2.1.png (73.38 KiB) Viewed 5477 times

I think it looks ok but I may for this bean make the "boost" of zone 3 less negative. This is also a very light roast but developed. The movement in the beginning of the roast was equally good with this smaller bean and the batch size was 95g - 5g more than with the bigger bean If you compare the two logs you can see the difference with the fan profile (blue line). The NordicLight5 has a higher fan speed in the beginning. The KL is a pure "fluid bed" roaster and the movement of the beans depends on the fan speed and you have to optimize it. A very high fan speed means a higher power consumption. This can seem to be a disadvantage compared to a roaster with a rotating drum. The advantage with the KL is that it is much smaller but you have to optimize the fan speed and also the size of the batch.
EDIT: I tasted the 50/50 blend of te two roasts and measured the roast degree of the blend with the Tonino meter. It was 116. More than 115 is very light ("cinnamon") with this instrument. I tasted it as filter coffee one day after the roast and it was almost tea like but more sweet and less tannin. The acidity was balanced. It was fruity. Very nice!

[Jonk]

Hej

First of all, thank you for your detailed posts and profiles!

I tried NordicLight3 on a washed Ethiopian a while back and got very nice results for V60 (didn't stand out in cupping, my guess is that higher extraction is beneficial to these delicate roasts)

I'm curious why our roasters seem to perform quite differently though. In my own profiles I have been lowering pre-heat because as you can see there's quite a bit of overshoot in the beginning:

nordiclight3and5.png (119.9 KiB) Viewed 5460 times

These are the washed Ethiopian (regular and rehydrated - for this bean I preferred regular) with NordicLight3, as well as PNG Kuli Gap using NordicLight5. I know that the overshoot might not be critical, as long as the beans are not scorched, but still wonder if things wouldn't be better with more even application of heat.. avoiding the squiggly power corrections in the first minute as much as possible.

I do like how smooth the power is after the first couple of minutes, but for all 3 batches there's an apparent crash around FC accompanied by a flick. Have you experimented with "pushing" the beans past FC with extra power to avoid the crash? That's assuming that the beans actually experience a crash and it's not just noise caused by released moisture.

Also noticed that your last 4 logs posted have declining power during the latter part of the roast, much like ET/MET on a drum roaster. Could it have something to do with the new chaff collector?

[nrdlnd]

Thank you very much Jonk for commenting my experiments and writings. I wish there had been more comments! I think there is more to do with the profile to make it even better. For me this short and fast profile for light roasts has been some kind of a "revelation" in my coffe roasting. I have actually never tasted better filter coffee! I will later on try it as espresso also. My comments is in cursive in between.

I tried NordicLight3 on a washed Ethiopian a while back and got very nice results for V60 (didn't stand out in cupping, my guess is that higher extraction is beneficial to these delicate roasts)
I haven't cupped them only tasted tħem as filter coffee
I'm curious why our roasters seem to perform quite differently though. In my own profiles I have been lowering pre-heat because as you can see there's quite a bit of overshoot in the beginning:
nordiclight3and5.png
I don't care about the overshoots in the beginning. I've found out that often the roast follows the ROR-curve earlier when there is a higher preheat (not scorching heat!) in the beginning. I think that's more important!
These are the washed Ethiopian (regular and rehydrated - for this bean I preferred regular) with NordicLight3, as well as PNG Kuli Gap using NordicLight5. I know that the overshoot might not be critical, as long as the beans are not scorched, but still wonder if things wouldn't be better with more even application of heat.. avoiding the squiggly power corrections in the first minute as much as possible.
I'm no expert but I don't think it matters. I'm also not shure our machines are so different. Of some reason my screen doesn't show the most upper part of the curves? I'm on Linux, maybe it differs
I do like how smooth the power is after the first couple of minutes, but for all 3 batches there's an apparent crash around FC accompanied by a flick. Have you experimented with "pushing" the beans past FC with extra power to avoid the crash? That's assuming that the beans actually experience a crash and it's not just noise caused by released moisture.
I have experimented to do that with another profile but not this one. Maybe try to shorten the negative 1st zone before 1st crack and see what happens. I'm not quite sure it's a flick and crash. Everything is happening so fast with this profile. Another thing to try is to make the negative boost of zone 3 less negative. This doesn't happen with all beans only some
Also noticed that your last 4 logs posted have declining power during the latter part of the roast, much like ET/MET on a drum roaster. Could it have something to do with the new chaff collector?

Thank you for seeing that! I've seen the declining power at the end and only accepted it as something positive! I checked my roasts and you could be right. This seemed to have happened after I got the new chaff collector!
It could be that the new collector retains heat in another way than the old much lighter collector. The new collector is a massive thing with a lot of glass and thick plastic. It was a very nice observation from you! If this is the reason then there is one more advantage with the new collector I think?

[nrdlnd]

Thank you very much @Jonk for your constructive comments!
I got inspired and made a very little change to the NordicLight5. Now I call it NordicLight5B. It's for bigger beans that needs more movement in the beginning with a higher fan speed in the beginning. Preheat 1000W.

NordicLight5B.kpro

(3.46 KiB) Downloaded 327 times

The only change is that I shortened the negative Zone 1 so the PID could regulate up the heat before 1st crack and try to avoid a tendency to a crash at the end. This is the log with 90g beans:

230125_IndiaKalledavarapura_NordicLight5B_2.7.png (72.99 KiB) Viewed 5434 times

I also did something with my first variants of the profile. In this case I took away zone 2. I made zone 3 less negative and I also raised preheat to 1050W. The reason for this is that this variant is meant for smaller harder beans from higher elevations for example 95g Sidamo Grade 2 beans. I call this variant NordicLight1B:

NordicLight1B.kpro

(3.27 KiB) Downloaded 287 times

In the last paragraph I wrote that I wanted to make some further smaller changes with the profile. These changes are included in the download.Edit: You can see the result of those changes later in the thread where I compare the effect of the new chaff collector on the roasting compared with the old collector.

The log looks like this:

230125_SidamoGr2_NordicLight1B_2.4.png (72.8 KiB) Viewed 5404 times

I think it's still modifications to do with this variant (for example make zone 3 longer at the end and maybe also shorten zone 1 as with NL5B) but I think it looks better than before. I have made these changes in the download of the profile above. I start both variants with L=3.5 and mark 1st crack and look at end temperature and a DTR around 20%-25%. I went a little longer with both roasts and the bigger beans has a tendency to get a higher L at the same visual roast degree. L is not very useful for roast degree as it is relative and no actual measurement. Comments?!

[Jonk]

Cool. Yes, I'm a bit surprised that there isn't more activity on the forum - must be quite a few units spread throughout the world by now.

For me this short and fast profile for light roasts has been some kind of a "revelation" in my coffe roasting.

Sure, I like how fast profiles seem to preserve acidity. I believe rehydrating the beans might be quite helpful for light roasts as well, as FC happens at a lower temperature. The results I have gotten with rehydrated beans have been mixed so far though - sometimes the added acidity have been too much. Really depends on the beans, but it does add some kind of juicy quality.

I'm also not shure our machines are so different.

I was thinking about the first minute. I need a significantly lower preheat in order to stay near the blue profile curve (with all the beans I've used so far)

Burundi_Ngozi_NL5_comparison.png (156.5 KiB) Viewed 5422 times

In this case 700W for the other 2 profiles (using the same beans and batch size. I mistakenly called the NL5 batch Kuli Gap in the last post, it was this Burundi). But as you point out it might not make much of a difference.

This NordicLight5 batch also performed better filtered rather than immersion brewed. Note that all 3 batches experiences an apparent crash and flick near FC. Less pronounced on the slow batch, probably due to less moisture left in the beans. But it also tasted bland, mostly chocolate and nuts, while the faster roasts had plenty of floral red currant and only a faint nuttiness.

It could be that the new collector retains heat in another way than the old much lighter collector. The new collector is a massive thing with a lot of glass and thick plastic. It was a very nice observation from you! If this is the reason then there is one more advantage with the new collector I think?

Yes, I think it could be an advantage. I guess you can try to find out by directly comparing roasts with the old and the new collector as the "only" variable. As much as that is possible, one change tends to have many consequences. For example the batches in the graph above have quite a few differences:

Burundi_comparison_data.png (11.82 KiB) Viewed 5422 times

It would be neat if Kaffelogic Studio could compile this kind of summary automatically and color code by batch instead of type of measurement.

and look at end temperature and a DTR around 20%-25%

My hunch is that DTR is less important than many believe. I tend to pay more attention to the absolute development time, increase in °C from FC.. and maybe weight loss as an overall indication.

In the end I think everything has some effect. For example RoR at FC. But after more than 13 years of roasting at home I'm still mostly fumbling in the dark!

I believe that for Nordic style roasts a crash by FC is not necessarily a problem, but that the benefit of avoiding it is that a flick is less likely to happen.

[nrdlnd]

@Jonk,
Very interesting! I understood that you had a lot of roasting experience! Interesting comparison between the Adaptive Raost and NL5. DTR was much lower with the Adaptive. What about underdeveloped taste? I'm still not prepared for a low DTR especially not with a fast roast like the NL. I'm afraid to get underdeveloped roasts.

Could you see that it was less tendency with a crash after FC with NL5B? I made only a small change when I gave the PID a chance to raise the heat before FC. I think it was only about 10 seconds I shortened the zone 1.

The screenshot of the log with the NL1B is before I made a similar change with it and the crash tendency is there but I have added the change in the download. I will see if it works with this variant also! The PID seems to react very fast and also the sensitive tiny temp probe.

Besides, the old chaff collector weighs around 210g and the new 730g! A more than 0,5 kg difference! I will make a comparison even if it's difficult to exactly replicate roast sessions!

Also an interesting observation of you with more acidity in faster roasts. Yes but it's a nice acidity in my opinion. It has also to do with that it's a light roast. I measure with a Tonino and my roasts with the NL variants are always light and not seldom very light (Cinnamon). But as I said I'm more comfortable with a DTR around 20%.

[nrdlnd]

I have made a comparison between the old and the new collector. Instead of the old light plastic lid I've used a lid to a Petri bowl (as I've done the last two years). I don't think this affects how the old collector works.

I've only made two roasts and with the old collector first and with the new about 10 min later. The roaster was on all the time to get the same environment temp reading. Ideally I should have made at least another two roasts and this time started with the new collector as I think that the result can differ between the first and second roast. First I will show the two logs with the old collector first. I tried to superimpose the two screenshots but it didn't make it easier to see the differencies.

Old collector:

230127_SidamoGr2_NordicLight1B_2.4_Old.png (74.24 KiB) Viewed 5392 times

L=2.4, FC 206.9 after 4:33, End temp 214,5 after 5:46 Weight 83,0g, DTR 21.0%, Duration 1:12
Tonino=115 (border between Light and Very Light)

New collector:

230127_SidamoGr2_NordicLight1B_1.5_New.png (73.55 KiB) Viewed 5392 times

L=1.5, FC 202.9 after 4:09, End temp 213.1 after 5:20, Weight 82,7g, DTR 22.2%, Duration1:11
Tonino 121= Very Light

Both roasts were with the same Sidamo bean. 95g batch and rehydrated 12 hours with 5g water. I made the roasts as I usually do. Mark 1st crack both listening and recording with a mobile phone. I strived for a DTR around 21-22%. I started with a L=3.5 to have margin to stop the roast manually.

FC came earlier with the new collector. It's more energy efficient and the roasting environment seems to be more stable (look at the power graph!). The old collector has a very open mesh and I think the new collector creates a small overpressure in the roasting chamber and because it's so massive may also equalize temperature differencies and give more conductive heat. Interesting is also that despite a slightly lower end temp the the weight loss is about the same even a little more. Power is declining more at the end when the beans are exothermic with the new collector. I think that is a good thing!

I tasted the roast with the new collector (Tonino 121=Very Light). Very fruity, berry like. Balanced acidity. Some chocolate. Nice but I know from experience that one week rest may can make it even better!

It's not possible to make any far reaching conclusions of this little experiment. Just see it as a hypothesis!
I wonder if KL support could come in and comment if they have made any measurements and comparisons between the old and new collector?

[Jonk]

What about underdeveloped taste? I'm still not prepared for a low DTR especially not with a fast roast like the NL. I'm afraid to get underdeveloped roasts.

I'd say the NL5 batch tasted less developed than the "fast" Adaptive.. I think due to reaching FC faster (even though everything under Development would suggest otherwise) 0.6C lower drop temp would also contribute I guess. The "slow" Adaptive was a lot more developed than both, even though it's at a measly 13,4% DTR..

Same beans rehydrated to +5% moisture, using your new adjustments:

Burundi_Ngozi_NL5B.png (69.25 KiB) Viewed 5357 times

So more heat into FC can make it crack vigourosly and still crash due to the amount of moisture released. Some beans crash more than other
This was on the verge of too light even for me. About 10,5% weight loss after adjusting for the fact it was rehydrated. I like to grind by hand to be able to feel how hard the beans are, and this was tough. If you muster the courage to try it, be warned that your grinder might not like it (also it probably won't taste as nice if using traditional burrs)
I seldom have the patience to wait a day before cupping so I just let the grinds sit out for a while. The taste was dominated by a pleasant lemongrass flavor. Not as complex as the previous batch but I drank this much quicker

Thank you for testing out the difference with the chaff collectors!

Power is declining more at the end when the beans are exothermic with the new collector. I think that is a good thing!

That would be my guess as well.

[nrdlnd]

Jonk wrote: "So more heat into FC can make it crack vigourosly and still crash due to the amount of moisture released. Some beans crash more than other
This was on the verge of too light even for me. About 10,5% weight loss after adjusting for the fact it was rehydrated. I like to grind by hand to be able to feel how hard the beans are, and this was tough. If you muster the courage to try it, be warned that your grinder might not like it (also it probably won't taste as nice if using traditional burrs)"[/color

I'm sorry but I don't get this! The weight loss with the profiles NL5B and NL1B is at least 14% for me (also adjusted for the rehydration) and my DTR is 20% or more. I've measured the roast degree with the Tonino optical meter and it measures near very light or very light. They have never been hard to grind. I do always cut a bean in half to check if the color is even inside and it has always been even with the NL profiles. The beans are also always easy to cut with a knife. They taste nice as filter coffee and they don't taste underdeveloped even if I'm not quite sure how underdeveloped is tasting as it seldom happens. After a dinner today I made 750 ml (6 cups) of coffee with a Moccamaster (it's a couple of years ago I made coffee with such a machine as I usually brew manually). It was 50% India Kallevadapura roasted with NL5B and 50% Sidamo Gr2 roasted with NL1B. The roasts had rested for 3 days. It tasted excellent in my opinion and the other people in the company said the same. This is a great achivement for me; very light roasts that are fully developed! I don't remember that I've got this with any other profile. When they have measured Light/Very Light they use to be not evenly roasted inside with many other profiles. I think it's the profile or maybe the rehydration or both in combination. I have not tasted the NL profiles without rehydration though so I can't be quite sure to be honest. I rehydrate to about 16% humidity. This means that when a bean for example measure 11,5% humidity I add water 8-20 hours before the roast so the humidity becomes 16%. I can for example add 5g water to 95g beans to achieve around 16% humidity.

My preheat with the NL5B is 1000W and 1050W with the NL1B. My NL5B is not crashing and NL1B less after my latest modification. I guess you still have the old collector and that may affect it but in my comparison above I had about the same weight loss with both collectors but the new collector made the roast behave nicer in my opinion. What preheat did you use? Are our machines differently adjusted? My machine has the original calibration and I haven't recalibrated it. This could be a problem when we are comparing our results!

I'm still not convinced that a low DTR is a good thing. I have managed to get light and VERY light roasts with a DTR between 20% and 25% that in my opinion don't taste underdeveloped.
EDIT: One more thing. Our roasts differ. 1st different beans. 2nd roast time and end temperature differs. My roast 230125: End temp 214.0 your: 211.0. Time: Mine 5:25, your 5:24. FC: Mine:204.7, your: 205.7. Time FC: Mine 4:19, your 4:24. Duration: Mine 1:15, your 0:59. DTR: Mine 22.4%, your 18.4%. What mainly differs is end temp, development duration and DTR. Weight loss differs also. You calculate yours to 10.5% and I calculate mine to a little more than 14%. I'm not quite sure if we calculate it in the same way though! Can that explain the hard and underdeveloped beans of yours? It may not explain the crash that you have seen though!

[nrdlnd]

Hi again,
I made another experiment and now with a very extreme bean only to see what happens. I've preferred to use SHB beans with the NordicLight profile. At the other end there exists a bean that is exactly the opposite. It is light with a low density. It's the Monsooned Malabar. It wasn't possible to measure the moisture of the bean as the meter caracterized it as something else than coffee and showed "LO" on the screen. Anyway I measured 85g of beans and moistured with 5g of water. I roasted with NordicLight5B and stopped around 21% DTR. FC came rather late and was difficult to "nail". End weight was 77g so it could have lost a little more. I measured with the Tonino=114 so it was a light roast. L became 3.3.

230129_MonsoonedMalabar_NordicLight5B_3.3.png (74.36 KiB) Viewed 5331 times

It's not too bad. It was drinkable but I think a bit underdeveloped. The power profile is rather irregular so mabe a power zone to level some of the irregularities? Maybe also make the zone 3 less negative, -2 or -3 instead? I think other slower profiles are better for this very special bean. But the profile worked anyway even if not very well then it could work ok for light roasts with medium elevation beans and maybe even Brazilians?