D-Light

Start a new topic for a new roast profile. Reply to an existing topic to enter your comments on the profile, or to attach minor variants.
Geronimo
Posts: 210
Joined: Mon 03 Jun, 2019 11:10 am
x 23

Re: D-Light

#2347

Post by Geronimo »

nrdlnd wrote: Thu 30 Dec, 2021 11:17 am A couple of tips: If you make changes of a profile and don't want to rename it don't forget to reload the profile in the machine. If the profile before the changes was loaded it will go on with the profile before the changes. I have made this mistake. If you modify a profile and want to give it a new name don't forget to rename it inside the profile in the "about" section. The machine reads this and there will be some confusion otherwise. If I want to keep the original profile before making changes I save it somewhere else on the computer and after the changes and renaming of the profile I put the original profile back on the USB-stick if I still want it there. It can be a mess to remember all what you have done. I try to document the changes in the profile but it had maybe been nice to have some kind of a protocol to document what you are doing. And it's possible to restore a profile from the log of a successful roast if you have missed to save it. It can also be a good idea to move profiles that you seldom use to another place on the computer.
Hi

This is a shortened rehash of my previous reply that I lost when was asked to re-login on ipad.

Thanks for the file management tips, very handy on the loading of the profile in the machine. Would certainly be easy to get caught out on that one.

On reading your comment, it made me think it would be handy, but overly complicated to implement, along the lines in photography when you shoot RAW photos (straight off sensor, and no in-camera processing), when the photo is loaded in say photoshop or other post processing software, a "sidecar" file is created and all post processing is contained in that sidecar file, the original photo is untouched (non-destructive editing). This type of thing for the roaster software would be handy, especially if one could also create snapshots, and could roll back to any point in time as long as that sidecar file exists, and would need to be backed up also. A history bar down side would show edits and another snapshot saves, and you could in effect have multiple branches coming off snapshots for different tweaks to a profile. Fairly complicated for roasting software to make it similar to how photo RAW editing software works, but no harm in discussing of course, even if it is dreaming a bit.

With respect to power profiling, I put in a request for the ability to pick a point on the power curve, and a rubber banding line appears, and then you line up your rubber banding line to pick the straightest part of your curve and pick the second point, then displayed on screen or on side would be the relative figures required, rather than manually calculating the rise required. Its only really a time saving feature, in particular if comparing multiple logs of the same bean and profile.

I think another couple of boost/power zones would definitely be handy as you mentioned a while back.

With respect to the backup not working, I've not tested that, as I've been doing the old school copy and pasting of the folders to my backup location.

Cheers
G
nrdlnd
Posts: 211
Joined: Sun 21 Jun, 2020 12:00 pm
Location: Sweden
x 49

Re: D-Light

#2348

Post by nrdlnd »

Hi,
The only thing I have changed from the default settings is the ROR Y-axis multiplier to X 5. My ROR smoothing of logs (secs) = 30 and that was the default value.
I think this has to do with how often ROR is showed? I don't know but guess that my showing is the mean during 30 sec. In your case every second and that can maybe explain why your ROR-curve goes up and down. You could maybe just change the setting to 30 sec and see what happens with the curve if it gets smoothed. I guess the measurements already are there so I don't think you have to make a new roast to see the changes.

I'm just now on the go so I can't try your setting to see what happens with my curves. We will be away for 4 days so I can't do anything until we come back.

I wish you a nice new year 2022!
Cheers
Geronimo
Posts: 210
Joined: Mon 03 Jun, 2019 11:10 am
x 23

Re: D-Light

#2349

Post by Geronimo »

nrdlnd wrote: Fri 31 Dec, 2021 12:10 am Hi,
The only thing I have changed from the default settings is the ROR Y-axis multiplier to X 5. My ROR smoothing of logs (secs) = 30 and that was the default value.
I think this has to do with how often ROR is showed? I don't know but guess that my showing is the mean during 30 sec. In your case every second and that can maybe explain why your ROR-curve goes up and down. You could maybe just change the setting to 30 sec and see what happens with the curve if it gets smoothed. I guess the measurements already are there so I don't think you have to make a new roast to see the changes.

I'm just now on the go so I can't try your setting to see what happens with my curves. We will be away for 4 days so I can't do anything until we come back.

I wish you a nice new year 2022!
Cheers
Hi

Have a fantastic 2022.

Here are the two curves for comparison, with RoR smoothing set at 1 and 30 seconds.

Cheers
G
Attachments
Peru Organic, EY - RoR smoothing set at 1 seconds.png
Peru Organic, EY - RoR smoothing set at 1 seconds.png (69.5 KiB) Viewed 7763 times
Peru Organic, EY - RoR smoothing set at 30 seconds.png
Peru Organic, EY - RoR smoothing set at 30 seconds.png (61.25 KiB) Viewed 7763 times
nrdlnd
Posts: 211
Joined: Sun 21 Jun, 2020 12:00 pm
Location: Sweden
x 49

Re: D-Light

#2350

Post by nrdlnd »

You have a lot of very good proposals to make the life easier as a roaster with the Kaffelogic. This sidecar thing would make it much easier to try out modifications for special beans. You could have different mods for different beans naming them after the bean. This and a couple of more zones at least one could be very useful. I think also that some graphical reprecentation of what happens with the curves when you make an adjustment. Now it's more of a guess how much you should adjust and also start and end point. Of course there is an inertia in the system but it could compensate for that for example a 100g batch. I think Chris should have these proposals.
Edit: We posted at the same time! I think the 30sec smoothing makes it easier to make meaningful adjustments.
Cheers
nrdlnd
Posts: 211
Joined: Sun 21 Jun, 2020 12:00 pm
Location: Sweden
x 49

Re: D-Light

#2351

Post by nrdlnd »

Hi,
Where I am now I don't have my computer with the Kaffelogic Studio only a tablet. Just some thoughts: The magnification of the ror curve on the y-axis to x5 makes it easier to see deviations up and down. But without smoothing it's "difficult to see the forest because of all trees". Some smoothing makes it easier to see greater changes over time that corresponds to what you can adjust with zones. I think it could be useful though to check with less smoothing to see if there are some very sharp deviations that you can't see because of smoothing. To be able to adjust such changes it is important that the roasting chamber has a low mass and a smaller batch could be easier to regulate? The KL seems to have a low mass chamber. I think a high fan speed makes it easier to regulate. A high fan speed comes at a cost though that you need more heating power. An advantage the KL has is the small temp probe that can react faster to changes than a bigger probe can.
Cheers
Geronimo
Posts: 210
Joined: Mon 03 Jun, 2019 11:10 am
x 23

Re: D-Light

#2352

Post by Geronimo »

Hi

Great educational roasting tips and observations that you are putting forward. I think the y-axis multiplication factor and a bit of smoothing is more helpful than what I originally had. Looking back over a few older logs I can see the odd roast that appear to show a flick/crash.

Cheers
nrdlnd
Posts: 211
Joined: Sun 21 Jun, 2020 12:00 pm
Location: Sweden
x 49

Re: D-Light

#2353

Post by nrdlnd »

Hi again,
I found this link to a blog from Scott Rao: https://www.scottrao.com/blog/2018/2/24 ... -dont-know
He says about the settings that the smothing should not be more than 15 sec (default is 30 sec in the KL) and that it's important to have a small probe (the KL has already a very small probe). I've tried less smoothing on some of my logs and and I even went down to 1 sec smoothing like you had before. I change my mind and I think that even an 1 sec smoothing can be very useful. Actually I think that my log from 210808_Burundi_washed_ModFiltOrgP_1.5 (the one with the power zone at the end) that I linked to before is very nice even with 1 sec smoothing but some of my other logs shows tendency to crash and flicking that wasn't so obvious with more smoothing. I think you can be accustomed to less smoothing but sometimes it can be good to have more smoothing to see the bigger changes over time. This is the log from 210808 and it doesn't look to bad even with 1 sec smoothing.
210808_Burundi_washed_ModFiltOrgP_1.5.klog
(72.66 KiB) Downloaded 314 times
This evening I'm going to roast a rehydrated bean and compare it with an earlier roast with dry beans. I may start a new thread about that. https://christopherferan.com/2021/03/15 ... -protocol/ I also wonder if we should have a separate more general thread about modifying profiles? To often I'm "groping in uncertainty" how to make changes and how they may affect the profile. Do you think it's possible to get some activity around that?
Geronimo
Posts: 210
Joined: Mon 03 Jun, 2019 11:10 am
x 23

Re: D-Light

#2356

Post by Geronimo »

Hi

I think some smoothing helps to see general trends when doing back to back roasts of same beans, which I do quite a bit of. I’ve gone back thru a number of old logs with smoothing. I’ll also back off a bit as per your reference to Scott Rao. I saw your rehydration post. Going to be interesting how it he roast came out without a moisture meter. Good in you for having a go at it.

Cheers
User avatar
Coffeeforums
Posts: 2
Joined: Wed 01 Jun, 2022 11:20 pm

Re: D-Light

#2459

Post by Coffeeforums »

I am still new in this but tested D-light profile and like it at 1.5-1.8 level.
In home roast we trust!
Post Reply