Ypuh wrote:Why don't you contact the manufacturer? They're quite responsive at Kaleido and eager for feedback.

I'm pretty sure you're right that the mechanic/engineers are not expert coffee roasters. If they can experience the problem, they likely have the capacity to fix it or figure out which parameters to use to dampen the effect.

Actually, it was the other way around. The manufacturer read my comments, and they promised to release a video with a decent PID roast. So, let us just wait.

Ypuh wrote:For me I'm just using manual mode. There's too many factors involved (ET, outside temp, pre heat etc.) and my quantities are too small to really benefit from the auto roast feature. Eventually I may get there, but that's still months out.

It is like manual vs automatic transmission, I guess. Yes, some people (in fact, professional) prefer stick shift.

With PID control we
- can achieve a perfect repeatability
- will not be subject to mistakes, less bad roasts
- can quickly switch from one profile to another without experimenting

Like with stick shift, some people (including myself) like it, but the vast majority people in the USA cannot drive manual cars. I guess the manufacturer wants exactly that - increase the pool of customers. Only a relatively small number of people know the art of roasting.

I posted quite a few comments today I just wanted to explain that some of the answers to my yesterday comment were incorrect. People wrote that the example with the Kaleido roast was OK, the two curves are close, the PID did the work well. It is simply not the case.

Haha, yeah. Same as with cars; if you drive every day and/or long hours, then automatic is definitely preferred. If you're just playing and fooling around a bit and the terrain is still varying, manual is also fine/better.

If you roast once a week 1 batch, then there's little use for auto-roasting due to the changes in the environment. If you plan to do 3-4 roasts back-to-back, or 1 roast every day, then it already makes a lot more sense.

I want it to work too (also a nerd, so if it can it must), but like I said I'm not really there yet. If they need a guinea pig I might be tempted.

Ypuh wrote:
If you roast once a week 1 batch, then there's little use for auto-roasting due to the changes in the environment.

It is actually the other way around. If you set up water temperature for espresso machine to be 199, the PID maintains that temperature despite any changes in environmental variables.

The same applies to PID roasting. The PID should reproduce the BT curve in any case. Probably you were talking about automatic roasting using pre-set alarms. Then, indeed, the curve will be different as the same inputs may yield different outputs.

In fact, PID roasting is good for those who roast rarely and may forget their tricks from the last time.

One roast from today in manual mode 200g Brazil Yellow Bourbon from NL supplier

I received a reply from Kaleido. They sent me a video of an automated PID roast with M10. I can see that the BT curve follows the background curve well, and that the RoR curve is quite OK.

Unfortunately,
- the resolution of this video is very low, and the details are missing.
- It looks like PID controls the ET only, not the fan speed.
- no explanations, no information on setting the three PID parameters.

So, even though I see that PID works in this particular example, but the repeatability is still in question. They did in in this example, but how I would do it, how I would use their technique - I have no idea.

But I am more optimistic now.

Thanks for following up Daumier!

That looks promising. There's 2 minor hiccups at the 2 and 6 minute mark, but those get evened out pretty fast. You'd have a tough time reproducing the same results manually with the human response time and guestimating the amount of power you'd need. Overall it follows better than my initial expectation. I assume this is where the M10 also scores better than the M2 due to the larger mass and more powerful heating elements.

I still need some more experiences, get a few decent roast profiles that I'm happy with, but already looking forward to exploring this feature in the near future. For now I need more customers (or smaller batches). I have some time on my hands the next months before my studies start again so might just give it a go.

Kaleido has to tell us the P and I and D parameters they used. Otherwise it looks like - hey! We did it. We won't tell you how. Go figure it yourself. Just know it is possible.

Guatemala San Marcos Finca Sola bio 250g looks and smells very promising, manual filter roast.
https://www.facebook.com/photo/?fbid=446744707456596
https://www.facebook.com/photo?fbid=446744780789922

Autoroast with Kaleido m10 all-in-one controlled by Artisan.

Pretty impressive control. How much smoothing do you have on?