Cheap thermometer calibration technique?

[DaveC]

I know this subject of "cheap" and "calibrate" used in the same sentence may
well be anathema to some of you but I need to verify that either my IR temp
gun is accurate or my DMM/thermocouple is, or neither. Accuracy to 2 or 3
degrees F is fine.

I'm looking for suggestions for a simple way to provide some kind of common
temperature "standard" (I use the term loosely, here) I can compare these
against.

Thanks,

 

[MooseFET]

I know this subject of "cheap" and "calibrate" used in the same sentence may
well be anathema to some of you but I need to verify that either my IR temp
gun is accurate or my DMM/thermocouple is, or neither. Accuracy to 2 or 3
degrees F is fine.

I'm looking for suggestions for a simple way to provide some kind of common
temperature "standard" (I use the term loosely, here) I can compare these
against.

The freezing/melting point of ice is quite exact. Boiling water is
good if you correct for your local air pressure.

 

[John Larkin]

I know this subject of "cheap" and "calibrate" used in the same sentence may
well be anathema to some of you but I need to verify that either my IR temp
gun is accurate or my DMM/thermocouple is, or neither. Accuracy to 2 or 3
degrees F is fine.

I'm looking for suggestions for a simple way to provide some kind of common
temperature "standard" (I use the term loosely, here) I can compare these
against.

Thanks,

Freezing and boiling water are obvious points.

Tapwater stirred with a lot of crushed ice will get within 10-15 mK of
0 deg C. Boiling water, corrected for barometric pressure, will get
within your error budget.

You can buy a thinfilm platinum RTD, for a few dollars, that is
accurate to a fraction of a deg C, if you measure it with a good
ohmmeter. Some of the semiconductor temp sensors are better than 1 deg
C.

Fever thermometers are very good close to body temp. That and ice
point is a combination that's hard to beat.

John

 

[GregS]

I know this subject of "cheap" and "calibrate" used in the same sentence may
well be anathema to some of you but I need to verify that either my IR temp
gun is accurate or my DMM/thermocouple is, or neither. Accuracy to 2 or 3
degrees F is fine.

I'm looking for suggestions for a simple way to provide some kind of common
temperature "standard" (I use the term loosely, here) I can compare these
against.

I don't think i have boiled any water or iced a thermometer lately. I try to
check out an unknown by a known in the shop. Sometimes I may have to use a third
or fourth device to be sure. That is a problem to contend with, its very difficult to place more
than one probe in one spot. I try to test using a beaker of water being stirred very rapidly,
otherwise temperature differential will kill you. Measuring IR can be trickey, as I found out
at home measuring aluminum foil. The heat of the reflection is what you will measure.
Since I have a few thermometers on the wall at home. viewing the IR probe vs the readout
gives a feel for the accuracies involved.

greg

 

[Chuck]

Freezing and boiling water are obvious points.

Tapwater stirred with a lot of crushed ice will get within 10-15 mK of
0 deg C. Boiling water, corrected for barometric pressure, will get
within your error budget.

You can buy a thinfilm platinum RTD, for a few dollars, that is
accurate to a fraction of a deg C, if you measure it with a good
ohmmeter. Some of the semiconductor temp sensors are better than 1 deg
C.

Fever thermometers are very good close to body temp. That and ice
point is a combination that's hard to beat.

John

Don't forget that with your IR gun, the
emissivity of the source is important in
calibration. I suspect water and ice are
not friendly calibration sources. Maybe
a black anodized heat sink that has been
heated to some calibrated temperature
would be a good source. Does your gun
have an emissivity adjustment?

Chuck

 

[John Larkin]

I don't think i have boiled any water or iced a thermometer lately. I try to
check out an unknown by a known in the shop. Sometimes I may have to use a third
or fourth device to be sure. That is a problem to contend with, its very difficult to place more
than one probe in one spot. I try to test using a beaker of water being stirred very rapidly,
otherwise temperature differential will kill you. Measuring IR can be trickey, as I found out
at home measuring aluminum foil. The heat of the reflection is what you will measure.

If you IR measure shiny copper or brass, what you're measuring is your
own reflection.

I dab things with black whiteboard marker to improve the emissivity.
Apparently Scotch Magic tape has high emissivity, too.

John

 

[RST Engineering \(jw\)]

Ice mixed with distilled (or for your purposes, tap will be fine) water and
allowed to rest for a few minutes will give you 32°F within millidegrees.
Be sure to stir it every half minute or so to get the thermoclines well
distributed.

Boiling water at sea level will give you 212°F fairly accurately, less 1°F
(for your stated accuracy) for every thousand feet of altitude you are above
sea level up to a few thousand feet.

Jim

 

[John O'Flaherty]

I know this subject of "cheap" and "calibrate" used in the same sentence may
well be anathema to some of you but I need to verify that either my IR temp
gun is accurate or my DMM/thermocouple is, or neither. Accuracy to 2 or 3
degrees F is fine.

I'm looking for suggestions for a simple way to provide some kind of common
temperature "standard" (I use the term loosely, here) I can compare these
against.

Not sure what you mean by "cheap", but you can get a YSI thermilinear
network, which consists of a composite of three thermistors and some
precision resistors for about $27. The one that goes from 0-100C
claims +/- 0.15C accuracy (+/- 0.27F).
That might be expensive for one calibration, but you might use it in
the future.
Then, you could get a black anodized heat sink with one side smooth,
and drill little holes in the side, close to the flat surface, and
insert the YSI thermistor and your thermocouple, maybe with some heat
sink glop. Then point the IR detector at it. Let it go through some
temperature changes slow enough that the whole thing has time to
equlibrate, and collect some calibration data for both devices.

 

[Spehro Pefhany]

I know this subject of "cheap" and "calibrate" used in the same sentence may
well be anathema to some of you but I need to verify that either my IR temp
gun is accurate or my DMM/thermocouple is, or neither. Accuracy to 2 or 3
degrees F is fine.

I'm looking for suggestions for a simple way to provide some kind of common
temperature "standard" (I use the term loosely, here) I can compare these
against.

Thanks,

At what temperature? You have to consider emissivity with IR
measurements. If your DMM/thermocouple reads okay with the input
shorted (terminal block temperature) chances are it's accurate enough
at typical heat sink temperatures. Chances are your IR is not.


Best regards,
Spehro Pefhany

 

[DaveC]

At what temperature?

~130F, max. I am frequently asked to measure the temp of a stack of press
output (big 6-color sheet-fed presses) to determine if the IR dryer lamps are
functioning properly. Sometimes using an IR gun is fine, but some press
operators prefer the traditional method of a probe inserted halfway down the
stack. It would be nice if my tools' readings corresponded.

Re. the ice "calibration": as someone pointed out, ice & h2o isn't very
IR-emissive, it it? Would it help to mix in a little black ink? (Lots lying
around a print shop...)

Thanks,

 

[Spehro Pefhany]

Ice mixed with distilled (or for your purposes, tap will be fine) water and
allowed to rest for a few minutes will give you 32°F within millidegrees.
Be sure to stir it every half minute or so to get the thermoclines well
distributed.

Boiling water at sea level will give you 212°F fairly accurately, less 1°F
(for your stated accuracy) for every thousand feet of altitude you are above
sea level up to a few thousand feet.

Jim

It's possible to get a couple of degrees F error (high) from boiling
water, particularly if you use a smooth vessel, such as a glass
beaker, to boil the water. Maybe put something like an airstone in the
bottom of the vessel.


Best regards,
Spehro Pefhany

 

[Rich Grise]

If you IR measure shiny copper or brass, what you're measuring is your
own reflection.

I dab things with black whiteboard marker to improve the emissivity.
Apparently Scotch Magic tape has high emissivity, too.

I once used cigarette ashes and spit on a piece of aluminum.

Cheers!
Rich

 

[Rich Grise]

~130F, max. I am frequently asked to measure the temp of a stack of press
output (big 6-color sheet-fed presses) to determine if the IR dryer lamps are
functioning properly. Sometimes using an IR gun is fine, but some press
operators prefer the traditional method of a probe inserted halfway down the
stack. It would be nice if my tools' readings corresponded.

Re. the ice "calibration": as someone pointed out, ice & h2o isn't very
IR-emissive, it it? Would it help to mix in a little black ink? (Lots lying
around a print shop...)

Can you take the probe and the IR out to the line simultaneously and check
a pile of paper with both of them simultaneously? You'd need the
cooperation of the foreman and the workers, of course.

Good Luck!
Rich

 

[DaveC]

All that IR from the lamps might be reflecting around and giving you

problems too.

Measurement is taken after printing is done and lamps are cooled (they aren't
in direct "sight" of the delivery stack on the press, anyway).

It won't mix in (it's oily), that's kinda the point with offset
printing. ;-)

We've got lots of (large format) inkjet printers, too! Would h2o-soluble ink
in the ice water increase the emissivity?

 

[DaveC]

Can you take the probe and the IR out to the line simultaneously and check

a pile of paper with both of them simultaneously?

So where's the "standard" in this experiment?

 

[Spehro Pefhany]

~130F, max. I am frequently asked to measure the temp of a stack of press
output (big 6-color sheet-fed presses) to determine if the IR dryer lamps are
functioning properly. Sometimes using an IR gun is fine, but some press
operators prefer the traditional method of a probe inserted halfway down the
stack. It would be nice if my tools' readings corresponded.

All that IR from the lamps might be reflecting around and giving you
problems too.

Re. the ice "calibration": as someone pointed out, ice & h2o isn't very
IR-emissive, it it? Would it help to mix in a little black ink? (Lots lying
around a print shop...)

It won't mix in (it's oily), that's kinda the point with offset
printing. ;-)

Thanks,

Best regards,
Spehro Pefhany

 

[John Larkin]

~130F, max. I am frequently asked to measure the temp of a stack of press
output (big 6-color sheet-fed presses) to determine if the IR dryer lamps are
functioning properly. Sometimes using an IR gun is fine, but some press
operators prefer the traditional method of a probe inserted halfway down the
stack. It would be nice if my tools' readings corresponded.

Re. the ice "calibration": as someone pointed out, ice & h2o isn't very
IR-emissive, it it? Would it help to mix in a little black ink? (Lots lying
around a print shop...)

Thanks,

If you use a stainless thermos or Starbucks coffee mug thing, and fill
it with stirred crushed ice and water, and fire an ir sensor down into
that, it should be pretty good.

I've got an ice cube on a paper plate on my desk, just starting to
melt. Our FLIR thermal imager shows it at about +0.5C, and a cheap
Extech IR thermometer claims -2C.

The emissivity of the ice cube seems close to 1.0. As I image it and
wave my hand around nearby, there's no hint of thermal reflections or
temperature change, so it's almost totally black in the thermal IR.

Oh my red plastic Presidential Water Cup is on my desk, too, with
about an inch of water in the bottom. It reads 6.8C with a
thermocouple, 7.0 on the FLIR, and 4C on the Extech. Again, no hint of
reflections. So water also looks black at thermal wavelengths.

John

 

[Tim Wescott]

If you use a stainless thermos or Starbucks coffee mug thing, and fill
it with stirred crushed ice and water, and fire an ir sensor down into
that, it should be pretty good.

I've got an ice cube on a paper plate on my desk, just starting to
melt. Our FLIR thermal imager shows it at about +0.5C, and a cheap
Extech IR thermometer claims -2C.

The emissivity of the ice cube seems close to 1.0. As I image it and
wave my hand around nearby, there's no hint of thermal reflections or
temperature change, so it's almost totally black in the thermal IR.

Oh my red plastic Presidential Water Cup is on my desk, too, with
about an inch of water in the bottom. It reads 6.8C with a
thermocouple, 7.0 on the FLIR, and 4C on the Extech. Again, no hint of
reflections. So water also looks black at thermal wavelengths.

John

Is your FLIR imager one of the cryo-cooled ones, or do you have a newer
one (less than a decade old) with the uncooled detector?

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at http://www.wescottdesign.com/actfes/actfes.html

 

[RST Engineering \(jw\)]

Talk about measuring with a micrometer, marking with chalk, and cutting with
an axe!!!

The fellow is looking for a couple of °F accuracy.

Jim

 

[John Larkin]

Is your FLIR imager one of the cryo-cooled ones, or do you have a newer
one (less than a decade old) with the uncooled detector?

It's an E45, about a year old, uncooled. About every 30 seconds, it
clicks and seems to close an internal shutter for a couple of seconds,
which I assume is some sort of auto-cal of the sensor array.

It's pretty slick, except for the USB interface which is bizarre. It's
a network device, not memory like other cameras, and you have to
install their strange software to talk to it.

We can focus so close that we can image the hot-spot temp on an 0603
resistor! Imaging an operating pcb can lead to all sorts of
revelations.

We looked at the Fluke, but nobody at Fluke seemed to understand it.
Not a good sign.

John