Choosing thermometry equipment for home roasting

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
Louis

Postby Louis » Jul 14, 2011, 2:20 pm

Note: This was first posted inside another thread, but I realise it will have more visibility as a separate thread.

I have to admit that the urge to buy thermocouples/data logger has taken me, as I realise that what is needed to use temperature as a roast progress indicator is really the bean temperature (BT). Else, I can only use indirect indicators: environemental temp (ET), color, sound, smell, smoke and past experience as guidelines to know where I am in the roasting process and when to end it.


Choosing a datalogger

There a a lot of different choices available, with different levels of functionality, user friendliness, construction quality. Note: As I'm in Canada, I've looked at Canadian prices, which may differ from US prices. All prices cited are in CAD.

All models referenced here offer: the same precision (0.3°C), 2 channels, wide support of thermocouple (TC) types, and max sampling frequency of 1 per sec.

Features wanted:
- 2 channels, to record both BT and ET.
- Ability to record data for the duration of a roast (sufficient memory).
- Ability to datalog in real-time to a connected PC, to eventually use software such as Artisan during the roast.
- Good quality/functionality - price ratio.

Fluke 54 II
Active123.com
$548 +$170 for FlukeView software + IR/USB cable + TC probes

(+) Good build quality.
(+) Easy to use.
(-) No real-time logging to computer (can only transfer pre-recorded data).
(-) Memory limited to 499 data points.
(-) 2 channels must use the same thermocouple (TC) type.
(-) Pricey (especially with the software).

Omega HH506RA or Extech 421509
Omega model at omega.ca $169
Extech model at active123.com $209
+ $30 for serial to USB cable/adaptor & basic software + TC probes

(+) 2 channels, each can accommodate different TC types.
(+) Allow real-time logging to computer (although data points maybe dropped by serial link).
(+) Cheap.
(~) 1,024 data point memory.
(-) Less easy to use but can be managed.
(-) Lesser quality compared to Fluke.

Omega HH806, models W, AU or AW
omega.ca
- W (Wireless, no USB) version : $199. 1,024 data points memory. Batt. only. 900MHz Wireless (require USB dongle).
- AU (USB) version : $210. 16,000 data points memory. AC adaptor avail. USB port.
- AW (USB + wireless) : $315. 16,000 data points memory. AC adaptor avail. USB port. 900MHz Wireless (require USB dongle).
+ $22 for USB cable/software or $40 for optional wireless receiver/software + $10 for optional AC adaptor

(+) Large memory (AU and AW models).
(+) AC adaptor avail. (AU and AW models).
(+) Integrated USB port.
(+) Wireless option (W and AW models), which would allow me to use software such as Artisan on my desktop computer while the roaster/datalogger is kept under the range hood.
(+) Still cheaper than a Fluke, even with the premium for the AW model.
(-) Lesser quality compared to Fluke.
(-) W and AW not compatible yet with Artisan but the dev. team is open. I would expect at least the AW to work when plugged in via USB. If the wireless works too; great.

Data logger prefence for...

The Omega HH806AW or the HH806AU (w/ | w/o wireless).

I've also considered the Thermoworks TW-USB-TC-LCD ($99), but it only supports one TC.


Setting up the TC probes in the Hottop

Basically, two options:
A. through the loading chute cover (Randy Glass' page)
(+) Temporary. The roaster can be brought back to its original state by buying a new chute cover.
(+) With Randy Glass' work, I know exactly where to drill and how.
(-) Looks like an after-thought. Have to manipulate the whole rig to load beans.

B. through a drilled hole at the back end of the roasting chamber
(+) Permanent. A more professional solution.
(-) Need to be sure where to drill holes for BT probe to keep it in contact with moving beans.
(-) Need to fix the probes in some way to keep them from moving back and forth.

I still need to decide which I will go with...

Q. Comments? Suggestions?


Choosing the thermocouples

Omega.ca has some well written documentation on thermocouples (how they work, the different types, etc.).

TC type
Most seem to use K type TCs, although T type are more accurate. When I use the Omega Probe Configurator, K and T type probes are the same price.

I will then order T types probes, as I see no reason not to do so.

Q. Have I overlooked something?

TC build
I will go with a rugged transition joint with stainless steel overbraid wire, stainless steel sheath, 6" length (to be confirmed if I chose to install them through the back wall of the roaster), 1/8" diameter.

TC junction type
Three choices here:
A. Exposed. Very fast response time. The TC junction is naked and unprotected. Not adequate for my use.
B. Grounded. Fast response time. The TC wires are physically attached to the inside wall of the probe. Probe is subject to electrical ground loop, creating noise on the line, affecting the temperature reading.
C. Ungrounded. Slower response time. The TC junction is detached from the probe wall. Offers electrical isolation. Longer lifetime.

Q. As the probe would be in contact with the roaster frame (which will be grounded to the electrical outlet, while the data logger will not...?), would the ground loop issue affect the temperature reading or not?

Regarding response time, Omega provides a comparison table (in air) but it only seem to compare exposed to grounded probes. Another chart compares exposed, grounded and ungrounded probes response time in water. According to this table, a 1/8" probe would result in 350ms response time (to reach 63% of the temperature change) for a grounded probe, compared to 550ms for an ungrounded probe. To reach near to 100% of the temperature change, this needs to be multiplied by 5, hence 1,75s vs 2,75s. Again, this is in water. In air, response time for the same grounded probe seem to be 10s (or 50s for 100%) (from the first chart)?

Q. So... what is best: grounded or ungrounded?


Last but not least, I need to see what kind of probe I need to buy (a third one) to replace our always broken cheap electronic cooking meat thermometers (we have gone through 5 of them in the last few years)! K or T type... grounded or ungrounded? ;)


Q. Any comments or suggestions?

User avatar
Arpi

Postby Arpi » Jul 14, 2011, 5:00 pm

My two cents:

two probes are better than one:

In a kitchen oven, when baking bread, there are two important temperatures. You have the controlling temperature of the oven (Environmental T) and the temperature inside the loaf of bread (BT). The same applies when roasting coffee. You have the temperature of the drum (or environment), and then you have the temperature of the beans. BT takes a while to catch up (or react) to ET. Some people try to control the roaster by using BT. But that would be like controlling a kitchen oven by using the temperature in the center of a loaf of bread. The outer surface of the bread would burn by the time it reads what you want.

A thermocouple probe (needle) is better than a naked thermocouple for roasting:

A thick probe works as a natural averaging filter. A naked thermocouple reacts very quickly and adds noise. If the thermocouple is placed in a draft of air, then it will add noise (up and downs). Noise is bad for pids and makes delta Temperatures difficult to read.


Cheers

User avatar
Jeff

Postby Jeff » Jul 14, 2011, 8:02 pm

If you're a tinkerer, you might also consider the TC4 board as well as many of the freely available roasting software products that supports it (and probably the data loggers you listed as well).

I've got an old Omega HH506 that has served well. Once you add in the USB adapter (unless you have a serial port), it is about the same price as the newer HH806 unit.

User avatar
MaKoMo

Postby MaKoMo » Jul 15, 2011, 3:50 am

(+) Allow real-time logging to computer (although data points maybe dropped by serial link).


These drops are not an issue anymore with Artisan. So the Omega HH506RA works well in this setting. However, I updated to a 4-channel Voltkraft K204 (without memory for the cheaper price) just because I wanted to see more ET!

Now I am starting to use both in parallel, this gives me 6 curves in Artisan. The two extra once I spent for Ambient temperature and cooling air temperature.
M.

User avatar
cafeIKE

Postby cafeIKE » Jul 15, 2011, 8:57 am

The Extech uses an optical data link, so grounded probe does not create a loop

With any meter, check that it functions properly with NO Battery at a wide temperature range when run on an A/C adapter. Some meters use the battery for CJC and while the LoBat warning is not visible, a low battery affects accuracy.

JimG

Postby JimG » Jul 15, 2011, 5:53 pm

cafeIKE wrote:The Extech uses an optical data link, so grounded probe does not create a loop

As does its sister model, the Omega HH506RA. I can confirm that the Omega model also works well with grounded probes, whether or not it is tethered to a PC.

There remains a minor problem with infrequently dropped data over the serial line, at least with the OEM software. I don't know if this is on the sending or receiving end, and I've never tried it with Artisan.

Jim

User avatar
cafeIKE

Postby cafeIKE » Jul 15, 2011, 6:54 pm

It's the receiving software. Roll your own works flawlessly.

Bad news is that it was reported to the vendor years ago and will never be fixed.

Louis

Postby Louis » Jul 18, 2011, 11:19 am

cafeIKE wrote:The Extech uses an optical data link, so grounded probe does not create a loop.

JimG wrote:As does its sister model, the Omega HH506RA. I can confirm that the Omega model also works well with grounded probes, whether or not it is tethered to a PC.

Interesting. So the IR link on the Fluke, the Extech and the Omega HH506RA is not simply an outdated technology but rather a feature to ensure electrical isolation of the meter?

Do you think I should fear such interference if I go with the HH806AU/AW on an integrated USB port?

Louis

Postby Louis » Jul 18, 2011, 11:55 am

Jeff wrote:If you're a tinkerer, you might also consider the TC4 board as well as many of the freely available roasting software products that supports it (and probably the data loggers you listed as well).

That would be fun and really interesting feature-wise, but I probably won't have the required free time to tinker with this, configure it, learn it, etc.

Maybe for a next iteration!

User avatar
cafeIKE

Postby cafeIKE » Jul 18, 2011, 1:17 pm

Manufacturers are well aware of the problems associated with grounded probes.

I'd lay better than even odds that a USB interface is opto-isolated.