HomeRoast Digest

Topic: Thermocouple (33 msgs / 821 lines)
1) From: David Lang
I use a USB LabJack.http://www.labjack.com/labjack_u12.htmlMy datalogger is some visual basic programming in Excel.
The thermocouples were acquired on eBay and I built an amp using a small 
PCB from herehttp://www.embeddedtronics.com/ Select signal conditioning

2) From: Michael Dhabolt
 Did you use the thermocouple amplifier from embeddedtronics based on the
Analog Devices AD595? Great link BTW. Reading the specs on the AD595
indicate its suggested uses are limited to 50 degrees C, I couldn't figure
out if they are talking about environment temp for the chip or temperature
indication range. how is it performing for you in the roasting temperature
 Mike (just plain)
 On 10/6/05, David Lang  wrote:

3) From: J.W.Bullfrog
Michael Dhabolt wrote:
There is a theory which states that if ever anyone discovers 
exactly what the Universe is for and why it is here, it will 
instantly disappear and be replaced by something even more 
bizarre and inexplicable.
There is another theory which states that this has already happened.

4) From: Michael Dhabolt
 Re-reading the description rather than writing, should have been my first
move. Too many espresso's this evening, or not enough.
 Mike (just plain)
 On 10/6/05, J.W.Bullfrog  wrote:

5) From: David Lang
I think that  the 50 is just the ambient temperature spec for the chip.  
The 2nd page of the datasheet gives the voltage/temperature calibration 
which goes much higher.  The one thing I had to do was power it from 2  
9 volt batteries wired to give me +/- 9volt,  Couldn't make it work on 
only 1 battery.  Web site gave reasons I didn't understand (I'm a 
chemical engineer not electrical).
I got started on this project and for data logging it worked fine.  The 
Poppery II that I was using was fine without control so I never finished 
the project  i.e. added control software.  Now I'm using a Poppery I 
which is too fast so I'm thinking of reviving it again.  Have solid 
state relays ready to go. for heater control..
Mike (just plain) mwrote:
 Did you use the thermocouple amplifier from embeddedtronics based on the
Analog Devices AD595? Great link BTW. Reading the specs on the AD595
indicate its suggested uses are limited to 50 degrees C, I couldn't figure
out if they are talking about environment temp for the chip or temperature
indication range. how is it performing for you in the roasting temperature
 Mike (just plain)

6) From: Michael Dhabolt
 The question was from one of my "DUH" moments, I re-read the specs and it
was pretty clear. JW recognized the moment and jumped on it.
 I use a modified P1 also, you are going to enjoy the project. The P1 is a
great platform for mods and is capable of excellent roasts. If I can be of
any help, don't hesitate emailing off list. I'll also be interested in
following your project - always something to be learned.
 Mike (just plain)

7) From: Matthew Price
On 10/7/05, David Lang  wrote:
Since thermocouples give temp differential in +/- volts you will need
a +/- supply.  Most times it's only necesary to have the negative
supply if you want temps lower than reference ( the meter's side) but
that chip might be picky.

8) From: Scott Marquardt
Tom's description for the nice TC he sells includes this: "You can replace
the thermocouple with any K-type or J-type (different temperature potential)
probe, available at any electonics store like a Radio Shack or even a Sears
Um . . . I don't think so. At least, *I* haven't found one that locally.
Has anyone found suitable replacements anywhere?

9) From: M. McCandless
You might try:http://www.techbuys.net/tl-108.htmlMcSparky
At 08:43 AM 7/10/2006 -0500, you wrote:

10) From: Matt Henkel
On Mon, Jul 10, 2006 at 08:43:42AM -0500, Scott Marquardt wrote:
RadioShack ahas certainly gone to the dogs.  I've seen the thermocouples
with mini connectors at Home Depot.  Check building supply stores like
that which cater to electricians (you'll generally find thermocouples
with the multimeters).

11) From: David Springston
    Anytime I need items like a thermocouple I use Omega: http://www.omega.com/ppt/pptsc.asp?refIN&Nav=tema07 I usually get the 
1/16 inch dia, grounded and stainless steel sheath.  I have never had a 
problem with this company, and their support is very helpful.  If you do 
order one of these be sure to order the connector for the end of the wire.

12) From: raymanowen
I just saw this.
You cannot just swap type K thermocouples with type J thermocouples. They're
totally different- K is Chromel/ Alumel wires, and J is Iron/ Constantan
Just so clods like me don't confuse them, the plug and jack connectors of
type K t/c's are yellow and the J is black.
The type K can generally be used up to 2000 degrees F, while the pure Iron
conductor of the J would rapidly oxidize at that temperature. In addition to
the color thing, their electrical response is completely different.
The meter is calibrated to read out in degrees F or C, and you can indeed
use any type K thermocouple. Several, if you have an external selector
switch, or you can just connect the one you want to read.
Thermocouples are not inherently accurate. If you have a thermodynamically
poor installation, the readings will be guaranteed Useless Lies.
Cheers -RayO, aka Opa!
Espresso is a Best Use of Water resources-

13) From: David B. Westebbe
This is a multi-part message in MIME format.
Thermocouples are not inherently accurate. If you have a thermodynamically
poor installation, the readings will be guaranteed Useless Lies. 
 I'm not sure what you are saying.  Is it more than "improper installation
will yield incorrect readings"? A re you saying that a proerly installed
tcouple will nevertheless  yield inaccurate results?

14) From: Michael Wade
This is a multi-part message in MIME format.
MessageDavid, I've heard (read) Ray expound before on the idea that heat =
is not static in a thermocouple; while heat is being absorbed into the =
bead, it is also being conducted away down the leads, or in the case of =
a sheathed TC, down the sheath.  If the TC is mounted in such a way that =
the tip (bead) is exposed to heat, but the leads/sheath are not, you are =
reading some temperature related to these heat flow factors.  This may =
be what he is referring to.
IMHO, I dont't think it matters.  It's true you're never going to get =
absolutely accurate bean temperatures, but all you need are relative =
temperatures.  You are going to build an experience base of how your =
coffee tastes at the temperatures you read on your particular TC =
installation, and that is going to be the truth for you.  
Michael Wade

15) From: Justin Marquez
On 7/13/06, Michael Wade  wrote:
I agree!  As it is functionally impossible to measure the internal
bean temp, just go with the air temp (which we can measure with
reasonable repeatability and some accuracy) and use our experience and
observation to relate it to our desired roast level.
In fact... most of us using HG/DB don't even try to measure temps.
I'll bet the same is true of the "wokkies", too.
Safe Journeys and Sweet Music
Justin Marquez (Snyder, TX)

16) From: Aaron
Thermocouples, contrary to popular belief, some of which may be from non 
understanding the principles in their operation ARE quite  accurate.
One can get a TC that is accurate down to 1/100th of a degree, and even 
the cheap ones are well within a degree (more than adequate for our needs).
Now you have to remember that the TC is only measuring what is at that 
very small spot where it's measuring bead is at.  A coffee bean is 
larger than that so very well may have temperature differences across 
and through it.  Airflow is turbulent by nature and not uniform,  there 
are 'drafts', 'channels', 'breezes',  whatever you want to call them of 
hotter air and cooler air.  Especially when it is not allowed a 'free 
flow' and has to force it's way through a moving mass (like your roaster 
full of beans for example).
Coffee beans are NOT uniform either.  some are wetter, some are drier, 
some are denser, some are lighter.  You get the picture.
So even IF you could get a perfect uniformly temperatured airflow, 
hitting each bean exactly the same, and measure it with a perfect device 
that is measuring exactly the core temperature.... Bean A is going to 
absorb more heat than Bean B and the temps will be different.
This is not a .... hobby.... profession...... however you consider it... 
to get overly anal retentetive about super precision temperatures on. 
Coffee Roasting is NOT an exact science.
Given the nature of how this all works, ANY temperature measurement we 
take is going to be an approximate anyways.  Are we measuring air, are 
we measuring bean core temperature.  What is the difference, how BIG is 
the difference, how accurate is my reading .vs. what is really in the 
bean, how uniform is my reading across all the beans,  what is the 
variation between the beans, what is the deviation.  Is this from my 
error or just part of the process.... we can go on and on and on.
It was already said,  you use your devices, and you tweak and experiment 
and find what works for you.  Just because joe blow's thermometer reads 
430 when his beans start cracking doesn't mean yours will.  His numbers 
are a starting point, a 'point of interest' for you, your 'process 
points' should be what your stuff reads at X starting to happen.
Im not trying to be an A hole about this, but just trying to explain 
that there is no set in stone rule book.  What works for one person 
might not be the same for another person, even if they have the exact 
same setup, same roaster, same TC probe etc.
This, I believe is what makes this such a great thing to be into, 
because you can play around, and can change things very easily, and are 
able to experiment very widely.  I can tell you how to roast a bean and 
not burn the house down.  I can tell you about how to get it to where 
it's close to it's best spot... if you like lighter coffees.   Jimmy can 
tell you the profile if you like darker roasts...  in essence, we are 
giving you starting points, to begin your journey at, and not giving you 
final destinations where you MUST end at.... that point... is entirely 
up to you and your taste preferences and your equipment operating 

17) From: Brett Mason
Well stated... constantly in pursuit of the perfect cup.
On 7/13/06, Aaron  wrote:
Brett Mason

18) From: miKe mcKoffee
I've experimented with a TC and my Rosto roasting. Many of the reasons below
precisely why I prefer using a simple bimetal. It's slower response not
reading rapidly bouncing temps like TC but more an average of slow bean mass
and surrounding air. I've seen bimetals sticking out of manually controlled
Probats and Diedrichs too. The super fast response of a TC not necessary to
Artisanally roast coffee.
Kona Konnaisseur miKe mcKoffee
URL to Rosto mods, FrankenFormer, some recipes etc:http://mdmint.home.comcast.net/coffee/Rosto_mod.htmUltimately the quest for Koffee Nirvana is a solitary path. To know I must
first not know. And in knowing know I know not. Each Personal enlightenment
found exploring the many divergent foot steps of Those who have gone before.

19) From: raymanowen
"Are you saying that a properly installed tcouple will nevertheless yield
inaccurate results?"
Not at all, David- the opposite is true: "improper installation will yield
incorrect readings..."
Physically, the readings you obtain from a thermocouple are electrical
potentials [in the micro- to milli- volt range for the type K] are direct
functions of the temperature of the junction of the wires. The meter will
indicate the temperature of the junction.
If the junction is heated by a high temperature air stream [fluid bed in a
roaster], and cooled by the lead wires and s/s or Inconel sheath that has a
temperature gradient of -400 degrees in a distance of a couple inches or
less, the meter indication will be misleading.
The installation will be drastically improved [and the readings more
accurate] if you maximize the exposure of the lead wires and the sheath to
the same heat you're measuring at the junction.
Michael and Justin have the right idea- you build an experience base for
your roaster, instrumentation and technique. Temperature [What temperature?]
is only a relative indicator. No two roasters will get the same readings
with different equipment and methods.
If you really don't want to know what's going on, use a type J thermocouple
with your type K calibrated meter.
Cheers -RayO, aka Opa!
Eschew obfuscation-

20) From: Scott Marquardt
This is all a great discussion.
My philosophy with my modded Pumper has been to place the TC near the bottom
of the descending mass (I'm using the long 3" pyrex tube, allowing for great
9 oz. roasts of late). My thought is this: I want to avoid any spot where
hot air is percolating upward in significant amounts, and I want to be low
enough among the descending beans that by the time the beans have gotten to
the TC, their outer temperature has had at least as much time as possible to
equalize inward.
Intuition plays a big role, short of a battery of empirical testing that
would drive me nuts and probably only vary intuition's placement by mere
millimeters and a degree or so.
In a year of roasting, I've found that when you're not chaining the roast to
a PID (I still want to do that!), it's less about "where it is" than "where
it's going" (what's the thermal momentum of the system at this point).
Something else interesting; I believe I've posted about how I've chained a
Salamander ceramic infrared emitter to the side of the 3" tube I'm using on
the Pumper? Well, I've found another benefit to that for my large roasts
that ascend high enough in the column, as the beans expand, to be
well-exposed to that emitter. You know how, in some cases, you wonder if a
largish Variac change might risk slowing a roast too far? Not a problem,
now. I've created a scenario where as the beans fully expand, they
automatically become subject to a "this roast simply cannot stall" scenario.
Dramatic changes to the Variac (slowing the roast) are buffered by the
constant IR, making the process much more forgiving.
Having said that, last night was very muggy -- though not terribly hot --
and the roasts seemed to want to go a bit faster than I'm accustomed to
seeing. Is this because the air is more dense on account of the humidity,
and thus more efficiently transfers heat? It's either a puzzler or I'm
imagining things.

21) From: Tom & Maria - Sweet Maria's Coffee
Agree with Mike - and I have given up on the empirical method in 
terms of thermocouples ... well, what I mean is that I can't get all 
my various roasters, from L12 Probat to my Poppery 1, to callibrate 
in terms of measureing bean surface temp. I say, get a 
easy-to-find/easy-to-replace low cost exposed thermocouple (not a 
thermo-well) and get it into the bean mass. Then calibrate *yourself* 
to the temperatures/corresponding crack temps/roast results. I know, 
this will make the engineers cringe, but there are SO many variables 
in roasting it would mean quitting your job to study physics and 
electonics to actually do it "right". As long as you can see the 
patterns develop by watching the temps and by watching the coffee and 
by TASTING the results, then you are indeed getting all the benefits 
a thermocouple has to offer. Once again, my .02 cents.
By the way, can we make sure we change the subject lines on threads 
when the topics change - lots of OT talk disguised under 
coffee-relate subject headings...
                   "Great coffee comes from tiny roasters"
            Sweet Maria's Home Coffee Roasting  -  Tom & Maria
                      http://www.sweetmarias.com                Thompson Owen george_at_sweetmarias.com
     Sweet Maria's Coffee - 1115 21st Street, Oakland, CA 94607 - USA
             phone/fax: 888 876 5917 - tom_at_sweetmarias.com

22) From: Philip Keleshian
I am an engineer and this does not make me cringe.  In both my Z&D and
my HWP I have a guide tube thru which I insert a simple K-type bead
probe into the bean mass.  I believe this is about as good as can be
Within the limits of my taste sensitivity I believe that the results
from each machine are similar if I match profiles.  Both machines are
modified to control temperature.  The control of the HWP is a bit
skittish compared to the very tractable Z&D but I can achieve similar
profiles with each machine.

23) From: raymanowen
Scott, it sounds like you've got it by the horns. I like your Pumper mods.
I totally agree that intuition plays a big role. If you depend upon a
thermocouple sensor for data input, even a non-stellar placement is useful
if it stays in  the same place from roast to roast.
Otherwise it would be like studying for a 4-year technical degree in the
native language in a different country every year- like the US, Germany,
Czechoslavakia and the PRC. Start yesterday on the Czech and Chinese, even
if German is their technical language.
Cheers -RayO, aka Opa!
Espresso is a Best Use for Water-

24) From: Rob Stewart
This is a multi-part message in MIME format.
I prefer to use an ungrounded TC probe and the temps don't change as =
fast and bounce all over.  Ungrounded means the bead is isolated from =
the wall of the probe as opposed to touching it   I was not happy with =
bi-metal dial types.  The ones I had would stick and then jump ahead . I =
resorted to frequently tapping them.
An Omega type K ungrounded    khss-18u-rsc-12  has worked nice for me.   =http://www.omega.com/ppt/pptsc.asp?ref=JHIN-RSC  and plugs into the =
meter SMs sells.    

25) From: Scott Marquardt
Yah. When I was a youngster learning to fly, I had three instructors. The
"what the heck're you doin'!?" factor drove me -- and them -- nuts.
Definitely want to keep the probe well-placed.
You're going to REALLY love how I place the probe(s) in the Open Sky
roaster. Roasting again tomorrow.
Say, I never read the FAQ --can we post pics to the list?
- Scott
On 7/14/06, raymanowen  wrote:

26) From: Michael Dhabolt
fast and bounce all over.  Ungrounded means the bead is isolated from
the wall of the probe as opposed to touching it  <
Your experience mirrors mine,  I clicked on the wrong item when
ordering from Omega one time and recieved a grounded probe - - that TC
probe is still in my TC stuff junk box.
Mike (just plain)

27) From: Douglas Strait
This is a multi-part message in MIME format.
Scott M asked:
Having said that, last night was very muggy -- though not terribly hot =
-- and the roasts seemed to want to go a bit faster than I'm accustomed =
to seeing. Is this because the air is more dense on account of the =
humidity, and thus more efficiently transfers heat? It's either a =
puzzler or I'm imagining things. 
Actually Scott, humid air is less dense since H2O has a  molecular =
weight of 18 while the weighed average of the other constituents of air =
is around 29. As a general rule, for a given pressure, a gas with a =
lower molecular weight has better heat transfer properties than one with =
a higher molecular weight.  That said, I would be surprised if you would =
notice the difference in your particular situation but perhaps you =
Back to the very original topic of suitable thermocouples, I prefer a =
moderately fast response, faster that I am likely to see with a sheathed =
t/c. My favored solution for placement in the airstream/bean mass is to =
form an exposed t/c junction [by simply twisting the pair together] =
using a larger gauge wire [20 gauge] that is in a shielded pair with =
drain wire. Between the heavier gauge an the added shield wire, the =
resulting t/c is stiff enough to remain immobile in the agitated bean =
mass. I have modded several Pumpers this way with good results. I use a =
20 gauge K type, Kapton [DuPont tm for a polyimide film] insulated =
shielded wire with drain. The stranded drain wire can be used to secure =
the Kapton outer layer as one might whip the end of a rope. Kapton is =
orders of magnitude more abrasion resistant than the fiberglass =
insulation that I believe was the subject of the post that started this =
thread. Kapton is rated to 600F continuous which is more than adequate =
for our needs. With my setup I have measured a response time constant on =
the order of 4 to 5 seconds with the probe in the center of the bean =
mass and typical air flow rate. I don't see any "noise" in the readings =
with this time constant yet believe it to be sufficiently fast.

28) From: Scott Marquardt
Something like these?http://www.omega.com/Temperature/pdf/KK_WIRE.pdfYou using something like the KK-K-20?
On 7/16/06, Douglas Strait  wrote:

29) From: Douglas Strait
Scott M asks:
Something like these?http://www.omega.com/Temperature/pdf/KK_WIRE.pdfYou using something like the KK-K-20?
That is the insulation but not the wire construction. Apparently Omega 
does not make what I described as a standard stock item. The closest 
they make would be in PFA Teflon, Omega #TT-K-20-TWSH. It is rated for 
use to 500F which is also adequate and is significantly cheaper than 
the Kapton. Here is the applicable Omega link:http://www.omega.com/ppt/pptsc.asp?ref-J-TWSH&Nav=temh06Per the Omega insulation guide, the PFA insulation, like Kapton, also 
has excellent abrasion resistance. Here is the applicable link:http://www.omega.com/temperature/pdf/tc_insul_guide.pdfThe wire I used was a Watlow Gordon division product but I failed to 
find a link after a search. It has been a few years since I purchased 
it. Watlow may have divested itself of Gordon since then. Their 
Serv-rite division seems to be making their t/c wire now.
Unlike a SS sheathed probe which buckles if bent very much, this wire 
can be routed as needed by bending in radii down to less that 0.5 
inch. If you would like to try some of the wire I'm using on your next 
Pumper mod, I'd be happy to fabricate a t/c from it and send it to 
you. If interested email me your mailing address.

30) From: Scott Marquardt
I might take you up on that. But let me know if anything on these pages
looks promising. I note at least one of the links may offer some very heavy
gage wire indeed. Any judgments you'd offer on these items would be as
appreciated as any actual thermocouple you might send!http://www.industronics.com/thermwire.asphttp://www.accuglassproducts.com/In-VacWiring/ThermocoupleCable.htm">http://www.thermo-electric.nl/pro.wir.tef.phphttp://snipurl.com/tnqrhttp://www.industronics.com/thermwire.asphttp://www.accuglassproducts.com/In-VacWiring/ThermocoupleCable.htm
On 7/17/06, Douglas Strait  wrote:

31) From: Douglas Strait
This is a multi-part message in MIME format.

32) From: Aaron Boothe
I need some help picking out a new thermocouple for my poppery.  My last one
died and I am hoping to get one that connects directly to my lap top.  Does
anyone have any good (and cheap) ideas?  Thanks for the help.

33) From: raymanowen
Your last thermocouple is not dead, and never connected directly to your
laptop. The laptop would be an expensive readout device even if somebody
gave you the adaptor.
If you want to use the laptop as a controller, it needs a few other things,
like the control program itself, a/d, d/a converters, SSR for power control=
Jeffrey Pawlan has already invented that wheel.
Cheers -RayO, aka Opa!
*Disclaimer:* *Statements contained herein have not been evaluated by the
SCAA.* This free advice is worth every penny and is not intended to
diagnose, treat, cure or ameliorate any SNAFU. Always consult your
professional cupper to see if it's right for you… Discontinue use and tel=
your Cupper if the coffee develops off-flavors. Side effects are usually
mild and result in Dissing some coffee shops…
25-Mar-07 01:20

HomeRoast Digest