I don't know how much thermometry translates from a fluid bed to a drum, but the idea of a software solution -- at least as a temporary fix -- is attractive; if only for its price. But the underlying question of whether that will give you the information you want is unanswered.

Placing the probe outside the drum and in the hopper seems like a good idea to a lot of people. Some of whom (not me) know what they're talking about.

If you cared enough to invest some engineering and money into really good data, I think you'd want to use more than one probe. A relatively unsophisticated way of going about that would be to place one probe immediately outside the drum where it's isolated from local fluctuations; another in the exhaust flow (homogenized to laminarity and not subject to eddies); and average the data in real time.

At a slightly more sophisticated level you could locate another probe or two, and do something with the data more complex than a basic average -- from RMS averaging all the way to FFT analysis.

The real question is: "What's all this crap worth?"

Even relative unsophistication costs time and money -- in the form of an additional probe and a multi-event datalogger at least -- but probably wouldn't provide much in the way of extra usable information compared to siting a single probe in a good location. Also I don't think you'd get much additional from seasoning the data with a soupçon of extra math.

If I had to choose one site as most likely to be useful -- based on fluid dynamic considerations and not experience with roasters -- it would be at the bean-chute/exhaust-manifold. But, like I say, not based on experience with roasters.

Bottom Line: At the end of the day, all you really want is enough additional insight into what's happening inside the roaster so that you can add to what you already have and confidently act on their combination.

BDL

boar_d_laze wrote:I don't know how much thermometry translates from a fluid bed to a drum, but the idea of a software solution -- at least as a temporary fix -- is attractive; if only for its price. But the underlying question of whether that will give you the information you want is unanswered.

A software issue (data logging) is better addressed by the software, regardless of price. Other issues, such as (probe placement), for drums or fluid beds is a learning experience based on needs, observation, preference and lots of trial and error.

Based on simple Ozarkian logic (that may be oxymoronic), I assume that if I can get a valid environmental reading that is consistent and repeatable, I can gauge a relationship between that measure and the bean reaction. That's really what I'm after.

Before I connected the TC and added and the datalogger, I was hand recording my temps at 30 sec intervals, using the factory installed bean TC, and the analog exhaust gauge reading. My concern with using the exhaust air temp as a TC location was that at approximately FC, the BT would start to exceed the ET. While I understand there's nothing wrong with that, I didn't have a lot of confidence that I was getting a reliable ET experience.

Now that I've figured out how to correct the flat line reading on Artisan by removing the restrictive filter, I can try with the ET probe where it is and later move it to the exhaust air location and see which represents a more understandable, reliable reference.

Meanwhile, my roasting skills continue to astonish the completely uneducated local coffee crowd.

tamarian wrote:A software issue (data logging) is better addressed by the software, regardless of price. Other issues, such as (probe placement), for drums or fluid beds is a learning experience based on needs, observation, preference and lots of trial and error.

Can't disagree with that, but...

Sometimes the nature of the problem is not at all obvious and can take a little work to figure out. For instance, with Dave's roaster, although we know we don't like data which reads well above any reasonable maximal environmental temp for a drum roaster we're still at the WTF?! stage of analysis.

I suspect there's enough experience with getting useful ET readings from a USRC Sample Roaster that if relocating the probe is in order, re-inventing the wheel is less efficient than seeking out the the right people.

BDL

Did not know the location of the ET is shown on the USRC's picture already, as Tom and I had mentioned:

http://www.usroastercorp.com/sample.htm

It appears the factory ET gauge is at the location we recommended.

Generally, there is a damper, weight, or gate just below the hopper. When it is turned, moved, or lifted, the beans drop into the drum, without going into the exhaust tube. When it is let go, it closes to allow exhaust relatively unmixed air flow.

From picture of the design of the USRC sample roaster, the "ET" will be highest, and probably reproducible at the location below the hopper, below the damper/weight, just above the drum. If the temperature is measured, it likely will be stable at this location, and gradually decrease along the exhaust pipe toward the cyclone due to heat loss to the atmosphere (duh!).

Now what I have not figured out over the past three years.....which measurement will be the most helpful? Be it the negative pressure measured by the magnehelic gauge, wind speed, or amount of air by mass flow meter?

Did not know the location of the ET is shown on the USRC's picture already, as Tom and I had mentioned?

Henry, are you asking me? If so, yes I was aware that USRC typically used the flat front or flat right face of the bean-chute/exhaust-manifold for mounting the analog probe. If you're asking Dave, he is planning on using a compression fitting to make use of the analog thermometer's site.

Reliance on USRC's choices is one of several reasons I've consistently suggested that location as the first option to Dave for relocating his own thermocouple.

Having had such difficulty with a concept as simple as distance in another thread, I promise not to discuss laminar flow beyond saying that laminarity is a cheap way of getting a fair average unaffected by eddies -- and again, that the manifold/chute is likely a premiere location for mounting an ET probe. Note though: You'll want to determine if some minimal amount of fan is necessary for laminarity or whether convection is enough. Easy enough to test during dry-runs.

An ideal setup is mounting the thermocouple on the left face of the manifold, so the wire travels in the same direction as the BT's, and the analog probe on the front face -- so as to get the electronic data stream of the thermocouple and steam-punk appeal of the dial. Looks count.

BDL

I did a roast last night with the probe left where it is in the first picture and adjusted the filter on Artisan and did have success with the graphing of the roast. I would still like to find a cooler place and will try the existing gauge location as soon as the new TC and fittings arrive.

The current location of the 1/4 hole for the analog gauge is about an inch above the hole in the face of the machine's bean entry location, but I'm not punching any more holes in the machine until I've tried every other option.

Speaking of air flow, I built a frame out of 1.5" PVC to mount my cooling fan/pan and find it to be considerably faster than the cooling tray on the machine:

Interesting cooling fan. Do you stir with one hand and roast with the other?