HomeRoast Digest

Topic: USB Data Loggers (42 msgs / 954 lines)
1) From: Matt Henkel
Lately I've been playing with my roasting times and curves and a
realization has come my way:  I really should know what the curves are.
I've been using my digital camera to record my thermocouple so I can
find temp/time but it's a bit tedious so I'm looking for a data logger.
I know a few people on this list use computer data loggers and I recall
there has been previous mention of the TC-08 (too expensive) and Go!
(too proprietary) loggers, but I'd like something that will accept a
standard K-type thermocouple (preferably mini plug, but hollow
would be okay) and costs under $100.  Is this a pipe dream?
I've done some google/froogle research but haven't had any success;
does anybody here have any suggestions?  The OS isn't an issue: I have
an old laptop that I can load with *nix or Windows (if I must).  There
is no IR or serial port on this laptop, only parallel and USB.
Hoping that a data logger will prevent me from torturing another batch
of Monte Leon "Miel" (although it's still pretty good)

2) From: Jim Anderson
On 3/14/06, Matt Henkel  wrote:
Try these folks. I have quite a few of their loggers and they work well.

3) From: Obrien, Haskell W.
Check out the 1-wire stuff from dallas. You can get some samples to try
You have to poke around a bit to find the sensors with high enough
range, but they are there.  You can use a serial or usb 1-wire interface
to get the data into the computer.

4) From: Tom & Maria - Sweet Maria's Coffee
Matt - can you tell me what you find out, or give a URL for a USB 
datalogger. I am interested too.
                   "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

5) From: Matt Henkel
I'll definitely let you, and the list, know what I end up doing.  I
checked out the HOBO units Jim pointed out earlier but their software is
$100 and the google shows that they don't play nice with OSS.  Right now
I'm leaning towards the PhidgetSensor Thermocouple
http://www.phidgetsusa.com/cat/viewproduct.asp?category&subcategoryA00&SKU51).I'll have to throw it in a project box but it should be trivial to write
a C-app which queries the temperature every second and stores it in a
text file; I should also be able to feed the data stream to gnu graph
for visualization during the roast.
There are certainly serial devices which do more out of the box, but as
far as USB devices go this is the least expensive and is extremely
developer friendly.  Even so I'm still poking around, I'll let you know
what I end up with.
On Wed, Mar 15, 2006 at 01:13:25PM -0800, Tom & Maria - Sweet Maria's Coffee wrote:

6) From: Demian Ebert
We've used HOBOs (mostly for water temperature monitoring) for years. Their
software is intuitive, the units are rugged, durable, and pretty darn cheap=
I suppose if you only need one datalogger then the cost of software makes
the unit unreasonably priced, but if you needed a dozen dataloggers, the
cost of software becomes minor.
Not knowing what OSS is I'm not sure how to interpret that.
My two cents worth.
On 3/15/06, Matt Henkel  wrote:

7) From: Matt Henkel
On Wed, Mar 15, 2006 at 02:17:09PM -0800, Demian Ebert wrote:
Agreed. I'll most likely purchase one datalogger and nothing more after
My apologies: OSS stands for Open Source Software.  While I'm sure
(from your comments and the reviews I've read) the HOBOs are great they just aren't
feasible if I can't get them to work with OSS (such as gnu plotutils).

8) From: scotthal
You could take a look at the Parallax thermocouple hobby kit; cost is <$40, and interfacing the 1-wire output to a standard parallel port is pretty simplistic (tho' i don't have a link to the relevant app note)

9) From: dballard
http://www.hobbyengineering.com/H1611.htmlIs this the one you were talking about?
scotthal wrote:

10) From: Matt Henkel
On Wed, Mar 15, 2006 at 05:33:46PM -0600, dballard wrote:
Directly from Parallax it's $29.95:http://www.parallax.com/detail.asp?product_id(022Of course the drawback here is that I'd still need a BASIC chip and USB
interface...  While this would be a nice toy I think it's overkill for
my simple data collection needs.

11) From: Coffee Dork
Demian Ebert wrote:
Agreed. I have used those in the past for stream surveys for ODFW and 
those things rock. But, the total package is probably a little cost 
prohibitive for what any of us want to do as far as roasting goes. Also, 
the Hobo probes generally have their own built in intelligence and data 
logging which drives their cost up. What I think any of us need is a 
more direct connection to a computer and do all the intelligence and 
logging there.
Open Souce Software/Solutions - some might say Open Sores.
Junior Roaster

12) From: Michael Wascher
Thermocouples are analog, low level signals, very non linear, and they need
to be referenced to a second thermocouple with a cold junction. No way are
you going to connect that directly to a PC. Single-chip processors are
cheap, and can be bought with interface hardware on-board . It's an
excellent solution.
And, cost? Maybe it wouldn't be an issue if you hadn't donated to Mr. Gates=
On 3/15/06, Coffee Dork  wrote:
"Life is just one damned thing after another."
  - Elbert Hubbard

13) From: Aaron
as for the data logger and temperature.  has anyone looked into a simple 
A/D converter, maybe a steppable one, and then just plug it through 
either a usb or rs232 (serial)  input...  it's kind of a bitch to set up 
but even microsoft excel with the odbc cube's or modules or whatever 
they call it, can take live readings and log them and put them into a 
spread sheet and store them.
I wrote one algorhythm / routine  a few years ago I use to keep track of 
kilowatt hours of electricity I generated with my solar panels, and it's 
a simple raw datafeed into the serial port and just manip the numbers 
like you'd do any other number if excel.
once could even build their own a-d converter  using an el cheapo op amp 
setup through one of the run of the mill a/d converters you can buy at 
any electronics outlet.
Just something else to think about.

14) From: Aaron
If you don't mind spending a grand or so, get a fluke documenting 
process meter,  I think it's model 721, ill have to dig it out and look 
at it tomorrow to be sure.
This thing is great, it will swap from any value to any value, and can 
simulate as well...  you have a 4 - 20 mA input from device A you want 
to change to 0-5 volt output to device B, tell it to do so, plug the in, 
and the out leads and you are set.
It will take any RTD type input,  K,R, N you name it, RTE's as well,  
you can sepecify PT100 or whatever discipline you are using as well...  
Probes are cheap too.  Given you are not trying to control down to 
1/100th of a degree, you dont need platinum...a standard 10 dollar run 
of the mill K probe will do you fine.  These also log data and can 
output to a computer via a serial port.  so you can in essence use the 
meter to 'convert' to what you need and squirt it out the data port.
Just another expensive toy if you really want to play hard core,

15) From: Michael Wascher
Thermocouples aren't that easy. The voltages that they output are very low
level, they need amplification,  and need to be referenced to a cold
junction. Now you have a signal you can run into an A/D.
Now you can log the data, adjust for the non-linearity of the thermocouple,
and finally have usable temperatures.
On 3/15/06, Aaron  wrote:
"Life is just one damned thing after another."
  - Elbert Hubbard

16) From: Aaron
thermocouples are a pain correct, but RTD/RTE's are very predictable and 
a lot easier to work with.  Probably a bit cheaper too.

17) From: Ebeneneezer Shay
As far as most data loggers of any kind, they generally use a serial
connection all the way up to the very nice stuff like National
However this is do-able for a descent price for descent hardware! Take
a peek here for a pretty nice logger that is capable of reading 2
probes simultaneously.http://www.extech-direct.com/421509_7_Thermocouple_Datalogger_w_Alarm_p/421=509.htm
and  thenhttp://www.extech-direct.com/TP875_High_Temperature_Probe_with_Ba=nana_Clip_p/tp875.htm
temp  for the probe.
The above set up offers bi-directional serial interface with windows
compatible software (All). The T-Logger is $159.00 with software. The
high temp probe (1000 F) is $29.00. The T-Logger comes with a few
probes, but the highest temp one is rated for only up to 482 F. If
that is comfortable, then the additional probe is unnecessary I
suppose. Finally here  is a rs-232 to usb adapter, there are cheaper
ones but I like the design of this one personally.
but something in my gut is telling me that the db9 connection on the
adapter may be insufficient to run this T-Logger. You could probably
call them. Most of the other quality thermocouple devices will run on
a usb but the price goes up to the 300 to 400 dollar range.http://sine.ni.com/nips/cds/view/p/lang/en/nid/201881--Ebeneezer
Date: Wed, 15 Mar 2006 13:13:25 -0800
To: homeroast
From: Tom & Maria - Sweet Maria's Coffee 
Subject: Re: +USB Data Loggers
Reply-To: homeroast
Matt - can you tell me what you find out, or give a URL for a USB
datalogger. I am interested too.

18) From: Nelson, Frank
	Thanks for your input.  I have been looking at the Extech 421509
7 Thermocouple Datalogger and the Extech EA15 Thermocouple Datalogger.
I restrained from stating anything because I could only find the EA15
yesterday.  Do you have any insight into the major differences between
these two models?  The only thing I can tell is the EA15 has more
possible samples ~8000 vs. ~1000 on the 421509.  At ~1000 samples and a
13 min roast I could take a reading every 8 s, which should be

19) From: scotthal
I'd originally suggested the Parallax kit, because it was a cheap implementation of a Maxim appnote (AN991) for a 1-wire hosted thermocouple. Kinda fun, quite accurate, and easily hosted by OSS solutions (see owfs.sourceforge.net). An alternative that can support either thermocouple or a couple NTC thermistor channels could easily be built around a $20 TI eZ430-F2013 USB Development stick. Relevant links would be:
www.ti.com/ez430 - USB hosted MSP430 uC development kit
focus.ti.com/lit/an/slaa216/slaa216.pdf - explaination & code for thermistor:thermocouple temperature sensing with the MSP430 uC
www.sensorsmag.com/articles/0405/21/main.shtml - using a PGA (the MSP430 has one onchip) to improve sensitivity (autoranging, w/hysteresis)

20) From: Ebeneneezer Shay
The 421509 7 model is aimed at remote data collection as it's primary
purpose, much like the HOBO units mentioned before. While I am sure
you would be able to get good readings on an entire roast with it at
3sec intervals as its lowest setting. The primary difference is that
the rs-232 connection for the 421509 7 is one way com. The EA15 has a
two way com so you can read real time. Funny enough that the real time
monitor can do both and is almost half the price, the tradeoff is that
the EA15 has less memory, yet still enough to profile a roast I would
I may just try this myself, that EA15 is a nice price! I don't have an
rs-232 on my laptop (I am guessing this is Tom's predicament as well) 
so I haven't decided if I want to bust the bucks on this project yet,
I will call them and see about the usb interface issue. That might be
the kick over the edge for me on this.   : )
Subject: RE: +USB Data Loggers
Date: Thu, 16 Mar 2006 09:17:31 -0500
From: "Nelson, Frank" 
Reply-To: homeroast
       Thanks for your input.  I have been looking at the Extech 421509
7 Thermocouple Datalogger and the Extech EA15 Thermocouple Datalogger.
I restrained from stating anything because I could only find the EA15
yesterday.  Do you have any insight into the major differences between
these two models?  The only thing I can tell is the EA15 has more
possible samples ~8000 vs. ~1000 on the 421509.  At ~1000 samples and a
13 min roast I could take a reading every 8 s, which should be

21) From: Nelson, Frank
Please let me know about the usb interface issue.

22) From: Matt Henkel
Wow!  I greatly appreciate the links, this is phenomonal information.
Unfortunately my Cognitive Studies degree never allowed me to take
circuits.  Although I'd love to get into hardware now I think it's going
to have to wait, I hope to get into this stuff when I'm "grown up".
On Thu, Mar 16, 2006 at 09:08:55AM -0800, scotthal wrote:

23) From: dballard
Does anyone know if this can be hooked to a PC.http://www.inxs-inc.com/cgi-bin/menu.cgi?a=view_prod&idv47Nelson, Frank wrote:

24) From: Cameron Forde
Keyspan makes a serial/USB adapter:
Probably not the best solution, but handy if you need it.

25) From: Michael Dhabolt
Cameron and All,
The USB communication format has had poor buy-in, so far, from
professional analysis equipment manufacturers.  On the other hand high
quality laptops have bought into USB to the exclusion of serial ports
and a good USB to 232 adapter had eluded me until I stumbled across
the keyspan unit.  The Keyspan product you provided the link for is
the USB to RS232 adapter that I have absolutely no trouble with. 
Unfortunately, I purchased several cheaper models from other
manufacturers and had compatability etc. issues with all of them.
Good product.
Mike (just plain)

26) From: Oolan Zimmer
On 3/16/06, dballard  wrote:
Short answer:  no.
Long answer:  That's a mux board that plugs into a larger data acquisition
system (Keithley 705). It's just a relay board with a thermistor at the
junction block so the controller can perform cold junction compensation.
This is what the controller looks like:http://www.atecorp.com/Equipment/Keithley/705.htmIt uses a proprietary connection back to the controller.   If you were
willing to reverse engineer the interface and design enough electronics to
CJC, A/D, and sample the values and put them on some sort of recognizable
data stream back to a computer, then you could hook that up to a PC.  But a=
that point, why bother with this board?
Oolan Zimmer

27) From: Oolan Zimmer
RTD's are expensive, fragile, and usually only go up to 600F or
thereabouts.  However, they're very fast acting and are usually accurate.
Thermocouples are cheap, can measure much higher temperatures, and can be
robust, but need conversion electronics, can be less accurate, and react
more slowly.
On 3/15/06, Aaron  wrote:
Oolan Zimmer

28) From: Aaron
OOlan you get a P100 rtd for example and it can go up to several 
thousand degrees.... it's a platinum element and is very linear... even 
the standard K type non plat  i believe will read up to 1400 degrees I 
believe is it's top.
they are actually very robust, save for bending them up they can take 
quite an amount of abuse.  Now granted if you are going to beat the hell 
out of a piece of equipment it is not going to last as long, but when 
you compare the construction of an RTD/RTE versus a thermocouple, they 
are both about the same.... the RTD probably is a bit more sturdy 
because you don't have the dissimilar metal thing with different rates 
of expansion / contraction and hence stress at those points.
The speed of reading, in the way they would be used, would be fairly 
negligible.  you don't need something that will react in half a 
second..  The temp is not going to shoot from 70 degrees to 600 degrees 
in 2.4 seconds either really.  get a thinner stem for the sensing part 
if you need quicker reaction time.

29) From: Michael Wascher
The dissimilar metals are just twisted together, so they don't really
produce a lot of stress. And, if something happens to it, you snip it off,
strip back the insulation, twist them together, and you're good to go again=
On 3/16/06, Aaron  wrote:
"Life is just one damned thing after another."
  - Elbert Hubbard

30) From: Oolan Zimmer
On 3/16/06, Aaron  wrote:
J and K-type thermocouples can read up to kiln temperatures and cost
$20-50/sensor.  All of the ones I've used have stood up to ridiculous
amounts of abuse.
The only RTD's I've seen that go over 600F cost over $100/sensor.  Those
went to about 1500F.  Admittedly, thermocouples that can be submersed in,
say, molten glass are also very expensive.
they are actually very robust, save for bending them up they can take
Your experience.  Admittedly,  I have much less experience with RTD's than
thermocouples, but mine is that they tend to be finer probes that are easy
to physically damage.
Now granted if you are going to beat the hell
I consider thermocouples to be disposable.  They get used up (by, say, high
temperature corrosive materials), then replaced.  I've never managed to
break one, they bend first.
The speed of reading, in the way they would be used, would be fairly
By response times, I'm talking about milliseconds.
Oolan Zimmer

31) From: dballard
I had that feeling but it is good to know for sure.
I think I will follow the rest of you on RTDs
Oolan Zimmer wrote:

32) From: Aaron
ok michael now you are just talking twisted pairs of wires... while 
technically that is a thermocouple, most 'professional ones' oh god 
here's that word again... are encased in something to protect that 
junction somehow.....  yes you can buy TC wire and just twist it and let 
er rip... very cheaply too.  and for those who worry, the response time 
will be very rapid as well.... this might actually be a bit better way 
to go, because of the slight resiliance of the wire, it'll bump and 
bounce around with the beans a bit better.  This is the cheaper way to 
go up front but, still you have to have additional 'stuff' to compensate 
and 'figure it out' since the output may not exactly be linear.

33) From: Ebeneneezer Shay
The rs-232 connection for the 421509 7 is one way com. The EA15 has a
two way com so you can read real time. IS ACTUALLY BACKWARDS, I
fubared the model numbers here, sorry.
It is the 421509 7 is the one with the RTA capabilities.
Frank, I didn't get a chance to call today. I plan on trying them
tomorrow, one of my co-workers has taken a position elsewhere so we
are throwing him a going away party tomorrow and there should be time
amidst the frivolity of the day as we won't be working.... well not
too much!
BTW, these are thermocouple loggers, this 421509 7 is a logger with
RTA. It is aimed more at the field tech rather than the QC or
As far as the usb to rs-232 adapter, it is not really a matter of
system incompatibility it is that when you use a USB adapter of any
type the system assigns generic standard for addressing the I/O. This
is fine for system peripherals that are designed to communicate using
these generic I/O standards. The issue comes in that these types of
devices are not designed to be a genericly assigned system peripheral,
instead they use specialized drivers to communicate with their devices
through the rs-232 as a specific I/O port. (i.e. it is the software
for the RTA that has the issues with the usb converter). That is not
to say that it cannot be done either, but a fair proficiency in C/C++
programming is necessary to write drivers for sindows (not a typo),
linux/unix, or OSS. There are programs like Labview that take much of
the pain from this type of exercise and as with anything else the
price escalates sharply parallel to stabilty and ease of use. Further
it takes a perticular type of person to be at home writing raw device
drivers (excentric, intelligent, and slightly massochistic), this is
not an activity for the faint of heart.....lol
Ha ha, I am a plumber or rather I was a plumber and I don't like being
refered to as that either...lol

34) From: Michael Wascher
Professional? Last time I used thermocouples professionally they were just
as I described. I have a 10 input logger, use the thermocouples to validate
performance of electronics for use in telephone switching offices, and
previously used them during development of aerospace electronics. I've been
making my living as an engineer for 30 years now, have a BS & MS in
Oh, and I hate the way professional is used these days. Last night I saw an
infomercial about a professional salad shooter. Give me a break!
On 3/16/06, Aaron  wrote:
"Life is just one damned thing after another."
  - Elbert Hubbard

35) From: Brett Mason
Professional Salad Shooter?
  "Boy's let's have it out with the Antipasti gang.  Red, you take the
lettuce, I'll aim for the pepperoncini.  And watch out for the onion,
he's got good aim..."
images are sometimes frightening...
  Zass - the Professional Bean Grinder
   (Yours isn't working?  Maybe you don't have professional beans...)
On 3/17/06, Michael Wascher  wrote:
Brett Mason
   _(( )_  Please don't spill the coffee!

36) From: Aaron
Oh, and I hate the way professional is used these days. Last night I saw 
an infomercial about a professional salad shooter. Give me a break!...
So you mean, like the 'professonal size' can of WD-40 I just bought not 
too long ago?   That kind of use for professional you don't like heh?  
But doesn't that make me a professional now because the can of goop i 
use says I am?   That's kind of the point I was trying to make, what you 
said earlier... words are thrown around now that don't have any meaning 
anymore, they have been turned into marketing tools and are nothing 
more.  Whether the choice of words are intended to make the product 
sound somewho superior / the user superior because they are 'hip' enough 
to buy that overpriced product, or serve that product, or whatever.. the 
true or perhaps I should say, original meanings of some words, are now 
BTW, what does the 'non professional' or perhaps its called ameteur 
salad shooter look like?  Somehow im picturing a spud gun on steroids here.

37) From: kofi
On 3/16/06, Oolan Zimmer  wrote:
http://tinyurl.com/ecm6e   for the **HEL-707-U-1-12-00*
-75 C to +540 C (~1000F)   ~$22 at newark or $21.10 herehttp://www.onlinecomponents.com/:-)

38) From: Oolan Zimmer
not working for NASA.  :-)
If you know of any RTD's that are accurate with < 1% error from 400C-1000C
and a 50ms (or less) response time, and can tolerate 1200C oxidizing gas an=
0C with moisture, and cost $60 or less per sensor, I'd be very interested.
I'm currently using type K thermocouples in stainless sheaths.
Oolan Zimmer

39) From: Matt Henkel
To all interested parties:  I ended up going with the
PhidgetTemperatureSensor and I must say I'm quite pleased.  Although
I'll need to put it in an enclosure it's extremely programmer friendly.
While I intend to use an old linux laptop permanently it's currently
having hardware difficulties so I slapped together a VBA program that
uses excel to collect the data, and the probe is great.
If anybody's interested in the app it's quick & dirty (ugly, no error
checking) but it's functional.  If you connect a
PhidgetTemperatureSensor this app will record the thermocouple couple
temperature and graph it in real time along with the ambient (cold
junction) temperature.  There are also buttons to record the time &
temperature of first & second crack.  A save button lets you (and you
need to) save the data to a new excel workbook without the macros (if
you try to save the application workbook with the roast data in it the macro will
erase the data the next time it's launched).  Since I don't plan on
continuing to use Excel for data collection I don't plan on doing
anything else with the data, but invite you all to build on it if you'd

40) From: M. McCandless
I would like a copy please.
At 08:27 PM 3/27/2006 -0700, you wrote:

41) From: Scott Marquardt
Ya. me too. Good footwork.
On 3/27/06, M. McCandless  wrote:

42) From: Matt Henkel
Okay, there have been a number of requests for my excel sheet so first
of all let me stress this:  It doesn't even deserve the title of alpha
software.  The only currently known crash situation is if you try to
start logging data before the application has announced that it's
recognized the temperature sensor and you then press the button to stop
logging data.  This would be an easy fix (switch from an toggle button 
to a normal button which checks the state of pTempSensor and toggles
it's caption based on that.  If this change is made the application
should be stable, if ugly.  I've released the code under the BSD license
so feel free to modify it and redistribute it.
On that note:
If you have a PhidgetTemperatureSensor
http://www.phidgets.com/index.php?module=pncommerce&func=itemview&KID4352974470.58.25.82&IIDafor Canadians, or PhidgetsUSA.com for Americans; unfortunately
PhidgetsUSA.com is down right now so I can't get you a direct link) all
you need is a K-Type thermocouple with a sub-mini connector
http://www.phidgets.com/index.php?module=pncommerce&func=itemview&KID4352974470.58.25.82&IIDdtry to find a different source, you should be able to find one for under
$15USD) templogger.xls allows you to record a roast profile and export
it to an xls file.
templogger.xls http://coffee.menagerie.cc/templogger.xls)has been
developed on Excel 2003 and requires both Phidget.msi
http://www.phidgets.com/modules.php?op=modload&name=Downloads&file=index&req=getit&lid)and Tushar Mehta's VBA Timer
http://www.tushar-mehta.com/excel/software/vba_timer/). If you connect
a PhidgetTemperatureSensor templogger.xls will automatically detect it
and notify you.  Two fields on the interface form allow you to enter the
bean type (Harar Horse, etc) and bean mass (I use this for start mass).
Entering these values changes the title of the graph.
If you then press "Start Collecting" once a second a
temperature will be recorded in "A" column.  This value is calculated by
finding the average of 16 temperature readings taken every 62.5 milliseconds (it should be taking readings every 50 milliseconds but reality is giving me 62.5 milliseconds).  These 1 second readings are graphed on a real-time auto-scaling line graph along with the ambient temperature (cold junction temperature), which is written to the "B" column, taken every second (no averaging, this is a single reading taken every second; the cold junction isn't nearly as sensitive as the thermocouple).
The thermocouple and cold junction temperatures are also displayed on
the interface form along with the elapsed time.  While recording data there are two other buttons:  "First Crack" and "Second Crack".  Pressing either button records the current elapsed time and current thermocouple temperature.  When you want to stop the collection press the "Stop Collecting" button (it was previously the "Start Collecting" button.  This stops the data collection and fixes the graph.  It also writes the total duration of the roast to the sheet.
When you are done you need to hit the "save" button.  This will open
another window asking you what file you want to save to.  You NEED to do
this.  If you leave the data in templogger.xls it will be erased the
next time you run the macro (usually the next time you open the file).
Choosing to save the data copies sheet1 (the only sheet in
templogger.xls, don't add another sheet and expect it to be copied as
well) to the new xls document WITHOUT THE MACRO.  This means you get a
static record of the roast.
If you have any questions or problems let me know, I'll do my best to
help out.

HomeRoast Digest