I have a question about the relatively inexpensive T/C readout 
instrument sold by SM. What is the accuracy specification of the 
instrument itself [not the associated probe]? If someone still has 
whatever paperwork comes with it, could you take a look? TIA.
Doug

If that's the CE digital one, the specs say that from 32F to 932F the
accuracy is +/-0.75%+4F. Resolution is 1 degree.  Why I still have that
little folded up paper is a tad unclear. Hope that tells up what you wanted
to know.
On 7/17/06, Douglas Strait  wrote:
<Snip>

This is a multi-part message in MIME format.
Demian [or anyone else]
Does your little piece of paper make a distinction between instrument =
accuracy and probe accuracy? The reason I ask is that the information =
you provided exactly matches the industry standard for a probe made from =
"Standard Limits of Error" K type t/c wire. If the answer is Yes, does =
it also give a temperature or temperature band *of the instrument* for =
which the accuracy spec applies. Typically this will be 23C or range of =
18C to 28C. Also, does it give a temperature coefficient for additional =
error when the intrument temperature is outside of the range for which =
the basic accuracy is specified? This may be expressed in form of C/C, =
e.g., 0.1C/C. Also it is sometimes given as cold junction compensation =
temperature coefficient [tempco]. 
Thanks
Doug

Let's see.
Insturment accuracy vs probe accuracy: In the tech specs portion of
the booklet, there is a Sensor entry that specifies the k-type t/c. The
accuracy heading is on the same level as the Sensor Type. The exact text
after the accuracy is: "+/-(% of reading + number of digits) at 18C to 28C
(64F to 82F) with relative humidity to 80%."  The +/-(0.75%+4F) that I sent
you yesterday comes from the table that follows this entry. I assumed
perhaps improperly, that this was the adjusted accuracy for the insturment
within a given temperature band, but now I'm not sure and the table isn't
referenced in the "text" so it's hard to say.
A couple of pages further into the booklet, it says: "Note that in common
with other thermocouple thermometers the accuracy specification applies only
to the insturment itself and allowance must be made for limits of error
permitted in thermocouple. The relevant specifications and respective limits
for K type thermocouples are: 0-400C,  +/-3% for 400-1100C, +/-0.75%"
I don't see anything that appears to be a temp coefficient when operating
outsdie the 18-28C range.
Does that answer your question? I have to admit I'm a tad curious as to why
you're interested in these specifics given the "ballpark" nature of
temperature monitoring I do in my SCTO.
I'd be happy to copy the booklet and mail it to you or scan these two pages
and email them if you'd like to read them yourself.
Demian
On 7/18/06, Douglas Strait  wrote:
<Snip>

All this "correction factor" interest stems from the fact that Analog
Pyrometers were made from analog microammeter movements with a known
internal resistance wired with a compensating junction and a specific gauge
and length of thermocouple wire.
This was totally dependent on the potential source (Eth)at the junction in
series with the thermocouple wire plus the total internal resistance of the
meter (Rtotal). The meter responded to current flow ( I=E/R) but was
calibrated in °F, °C, K or R. (You could roast at a temperature of 860=
°R)
If you really care, Omega Engineering can send you a wheelbarrow full of
Temperature Measurement manuals. Everything else is palaver. More than you
might care or understand for a directed reading class. Brown would hate
delivering to you forever.
The high input impedance of the digital thermometer would compensate for al=
l
but very high frequency temperatures.
Cheers -RayO, aka Opa!
Use Correctol film developer if you shoot a roll of PlusX Pan film at a Tri
X meter setting-

This is a multi-part message in MIME format.
Thanks Demian. It appears that what you originally posted is in fact the =
instrument accuracy. 
Regarding your question as to why I asked the original question. There =
are two somewhat unrelated motivations. First, having had a career in =
test engineering, I am a bit of an instrumentation geek. Occasionally =
folks ask me for advice on purchases of stuff like this. For those who =
are cost sensitive, I wondered about what you get for your money with =
this one since it is notably inexpensive. My other motivation was =
stimulated by some past threads which try to compare experiences of =
different people in measuring roasting temperature mileposts such as =
temperature of 1st crack. Since many folks on the list use that =
particular thermometer I wondered to what extent it's inherent =
inaccuracy was a factor in the results posted. 
Since your last posting, another member has sent to me off-list a link =
to a .pdf of the specs. Like what you have, it do not give a tempco for =
instrument temperatures outside of 18-28C. 
My own $0.02 on the subject of temperature measurement utility as a =
roasting aid is that you need to learn what the temperatures mean for =
your specific measurement means and maintain consistancy of your means. =
Consistancy in probe placement is key IMHO. 
Doug

raymanowen wrote:
<Snip>
?? High frequency temperatures?  You mean like KT/h ?
<Snip>
Is that a joke?  You would basically have to push process by two stops.  I don't recall whether there is a developer called "Correctol".