100g sample roasting

[Yarumo]

Hello all, first time posting so please forgive terminology missteps

I purchased the Huky specifically for sample roasting, ideally 100g batches since I often don't have much to work with, just wanted to share my experience so far and see if anyone has feedback or knows a better way:

Upon first roast it was apparent BT is not reliable below 300g batch.  Relocating BT probe is not an option due to blade height.

However, with 300g, dialing in an ideal gas setting based on batch size and adjusting only airflow (with dimmer) during the roast, I was able to get an acceptable BT curve and proper development for cupping.

Based on this, I have been able to reduce gas for 100g, leave it constant throughout the roast, and use the same airflow management as I did with 300g, basically fly blind, and replicate the same time and development result with 100g batch size.

It's not a perfect solution, but allows for a much better result than the maddening electric machines I have used in the past. Leaving constant gas and adjusting only airflow is the way most exporters' QC labs I have seen here in Colombia roast on barrel machines specifically designed for samples, so I don't feel so bad doing it this way.

Has anyone come up with a good way to roast 100g for cupping for green coffee quality?

Thanks

[edtbjon]

Hi and welcome.
Thank you for a very specific and well phrased question (well, subject... ). Well, I have done some roasting with 125g samples following the general guidelines for sample roasting (as used for comparing green beans, possibly a lot of different green beans). While I did these tests with a perforated drum setup, my observations still apply to solid drum roasting. That thread can be found here: Sample Roasting with the Huky.

While sample roasting per se isn't about finding the best possible profile for a bean, I still think that I should develop the best possible common profile for all the beans that I'm supposed to try out. ("common in regards to my control profile, i.e how I set my heat and air and at what time I change them etc trying to produce a similar roast for all of the beans.)

I've read about a few attempts of using the Huky in this specific role, some successful and others just suffering through the whole roasting session(s). In my mind it all comes down to experience and knowledge. The Huky is up to the task, but you have to know about roasting in general and the Huky in particular.
In general it seems like you're doing very good. There are a few things that I personally would do different though. I guess that the machine(s) you're comparing with are "barrel" sample roasters, which have a very open airflow to begin with. The Huky is more traditional in how it works (with airflow and heat), so I'm not sure if "constant heat" is the best approach. I'd probably use a scaled down 1-2-3 airflow - 3-2-1 heat approach, i.e the same that I use for a normal roast.
This is a bit like "reverse engineering". If I have a bean for which I've found a good (production) roasting profile which is "in the window" of my intended sample roasting parameters (roasting time, Agtron/Tonino values etc.), I'd use that same bean for experimenting, trying to find a good control profile (how to manage heat and air) with the 100g samples.
While I didn't pay to much attention to the ET probe while doing my tests back then, I'd use that as my guide, trying to mimic what happens to the ET during the roast when comparing my "normal" roast and my sample roast. My reasoning for this is simply that the ET is in the open gas (heat) flow in both the reference (in your case 300g) roast and the 100g sample roast.

For the sake of completeness, I guess that I would re-make my profile if I had to sample roast a bunch of e.g Brazils after having done my Colombians. (This is where roasting experience and knowledge plays a part. A low grown natural Brazil does roast very different from a high grown washed Colombian...)

[hankua]

Sounds like a good project!
I would imagine the ET would need to be in the 200* range, using the projected LCD in Artisan to regulate air.

[SusanJoM]

Here's the YouTube video that Peter/Bean Animal made of the HUKY with jut 100 gm of beans in the drum.
https://www.youtube.com/watch?v=JKBKOkNMxAU`

[beananimal]

You can get Clear BT reading for 100g.
You need to know, that the less beans are in, the more they are concentrated at the front. So you need a shorter probe.
Additionally a shorter probe you can place close to the blades.
In" my little fat Huky" the 17mm long probe is at a 40mm radius at a 9 o'clock position and fully in the beans.
( I am writing on the mobile, therefore no Clear Referencing  ) typing correction is a pain

[edtbjon]

You can get Clear BT reading for 100g.
You need to know, that the less beans are in, the more they are concentrated at the front. So you need a shorter probe.
Additionally a shorter probe you can place close to the blades.
In" my little fat Huky" the 17mm long probe is at a 40mm radius at a 9 o'clock position and fully in the beans.
( I am writing on the mobile, therefore no Clear Referencing  ) typing correction is a pain

... which of course is easy to find out if you've run the tests with the plexiglass front.
I measure my normal 400g roasts in a transparent jug and they always levels around the 500ml mark (a little bit more than a US pint). Doing the math, 100g of greens is about 125ml (4.5 fl oz ?) which for sure doesn't make up a lot of mass for a probe with its tip protruding 50mm or so into the roaster.
Knowing that the fins in the drum pushes the beans towards the front it's easy to imagine how this works, but with BeAnAnimal's videos things became much more clear. (Or should that read "Bean Animal"... as we are dealing with coffee beans. )

I've been thinking about experimenting with the TCs, but I guess that I'm more focused on profiling the roast instead of the roaster... It's a great tip though and I think that a shorter tip does give more accurate readings even with bigger charges.

[SusanJoM]

You can get Clear BT reading for 100g.
You need to know, that the less beans are in, the more they are concentrated at the front. So you need a shorter probe.
Additionally a shorter probe you can place close to the blades.
In" my little fat Huky" the 17mm long probe is at a 40mm radius at a 9 o'clock position and fully in the beans.
( I am writing on the mobile, therefore no Clear Referencing  ) typing correction is a pain

... which of course is easy to find out if you've run the tests with the plexiglass front.

We are lucky Peter did all those tests for us, so all we have to do is refer to the links he posted in this thread for each of the usual charge sizes.

https://www.hukyforum.com/index.php/topic,1213.msg14141.html#msg14141

Scroll down in this post for pictures of those short probes he mentions:
https://www.hukyforum.com/index.php/topic,1213.msg14141.html#msg14141

Remind me, Peter:  did that short BT probe go into a brand new location or is it in the original BT hole?   

[beananimal]

In" my little fat Huky" the 17mm long probe is at a 40mm radius at a 9 o'clock position, looking at the front.
This is very close to the blades, 38mm is save. And would work with less than 100g. With 500g (only) the filling level is even.

[SusanJoM]

So I went back to your pictures and think you made a new hole for that short/fat T/C and that that hole is right to the side of the original BT T/C ?

I guess my real question (as in what can I and what can't I get away with) is:  could that short T/C be put IN the BT location and still miss the blades and be useful?  I'm disclined to drill more holes in mine, but it would still be a fun mod on one of the HUKYs.

[beananimal]

Yes Susan , you can use any shorter TC in the original location,
But not the original 70mm probe in the new place.

It should be closer to the real BT, but I am not sure, as there is sort of a hole, free of beans in that area. It may depend on the tryer, installed or not.

[hankua]

I tried this a few times, using mostly fixed gas and air. The first batch was charged with the gas off and the second with the gas on. Not sure about the consistency and if it can be done ending at a cupping finish. First crack ended up at @ 220c on both roasts, using 1kPa and 25V on the Variac. Some fine tuning might help, such as a lower gas setting.