Agilent U1186 Thermocouple +3C accuracy instead of ±1.1°C advertised

[smile]

Hi, I just got my new proble for Agilent U1252B DMM, tested it with mercury thermometers (two of them) show room tem at 24C, the DMM proble shows as 21C.

So that gives +3 Accuaracy???

The included "Operating instructions with the probe"

List Accuracy of U1186A K type thermocouple as ±1.1°C

So I should get something like 25.1at high end, and 22.9 at low end if the room temp is 24C, my DMM shows 21C.
Should I return the probe as faulty???

[HLA-27b]

What are the accuracies of the mercury thermometers ? What were you measuring? That probe is tiny and the mercury thermometers are big maybe there was a difference in coupling to the heat source. For probes this tiny the air currents make a difference. Take this into account before deeming it inaccurate.

[saturation]

Yes, to eliminate geometry errors use a water bath.

[smile]

What are the accuracies of the mercury thermometers ? What were you measuring? That probe is tiny and the mercury thermometers are big maybe there was a difference in coupling to the heat source. For probes this tiny the air currents make a difference. Take this into account before deeming it inaccurate.

As I said I use two lab mercury thermometers that show the same temperature, they are 1 degree accurate, I also got 0.5 degree mercury thermometer (but actually cheaper one) that shows 0.5 degree more than the two I mentioned.

So at worst case there is 0.5 degree fault on mercury thermometer side.
The DMM proble shows 3 degrees less, I even tried to measure human body temperature, and at 33.5 that can't be right when you add 3 degrees you get 36.5 normal body temperature.

Initially I was measuring air temperature in closed vessel, no air movement there.

[smile]

Yes, to eliminate geometry errors use a water bath.

OK I re tested, the error is still 3 degrees, mercury at 23 degrees, DMM at 20.2 degrees
As you know mercury doesn't measure negative temperatures so I wonder what would be negative temperature accuracy? Or is it linear?

Dave could make a temperature probe shootout to review various including this model thermocouple probes !!!

[HLA-27b]

Mercury thermometers are usually calibrated in water at a certain immersion depth. This depth is usually marked on the thermometer. Anything except measuring liquid temperature at this exact depth may be introducing error.

Besides that, the way the k junction is connected to the DMM may be introducing an offset voltage which needs to be accounted for. Ideally you should make yourself an offsets table by calibrating the thermocouple at 0°C and 100°C.

I am saying all these because there isn't much in a way of inaccuracy which can be attributed to a manufacturer. It is hard to manufacture a thermocouple wrongly.

[Richard W.]

For measurements with K-type sensors you must be shure, that the multimeter and the plugs on the leads have room temperature!
If the multimeter or the plugs are hotter or colder than room temperature, the cold junction compensation doesn't work correct.
(Thermocouples are measuring the temperature difference between hot and cold junction. The compensation is necessary to get the absolute temperature.)

Usually there is a platinum temperature sensor in the meter to measure the temperature of the cold junction.

For testing the cj-compensation you have to short circuit the input terminals of the meter.
The value on the display is the actual value for the compensation. This should be room temperature.

 If not, wait some hours. If it's still not room temperature -> faulty meter

The procedures with the two thermal fixed points can be found in the internet. This would also be worth a try.
Maybe you have to consider the height above sea level. depending where you are living 

BTW: The Agilent is also capable to do measurements with J-type sensors. The resulting error when the wrong sensor is selected would only be a few degrees...

[smile]

For measurements with K-type sensors you must be shure, that the multimeter and the plugs on the leads have room temperature!

Yes they are is the room for a whole day so certainly room temperature.

For testing the cj-compensation you have to short circuit the input terminals of the meter.
The value on the display is the actual value for the compensation. This should be room temperature.

 If not, wait some hours. If it's still not room temperature -> faulty meter

My meter has internal thermometer displayed too, its 20.8 degrees
I have short circuited the meter and the values is 20.9 - 21.0

The mercury thermometer shows 23 degrees.
 

BTW: The Agilent is also capable to do measurements with J-type sensors. The resulting error when the wrong sensor is selected would only be a few degrees...

I have the yellow probe, clearly written K on it, the default for agilent is K type probe. K symbol is displayed during measurements.

[smile]

Here is photo from my test:

[Kiriakos-GR]

temperature probe shootout to review various including this model thermocouple probes !!!

Here is one made before five minutes. 

The gladiators : Fluke & Agilent & a 23 years old (made in Taiwan) K probe,
that came with my first (Good one ) DMM.

[saturation]

Hmmm, impressive.  Test it now against water as calibration, ice bath water and boiling water.  Ice water is ~ 0C while boiling is 100C.  Note in real life there can be minor variations due to the contaminants in fresh water [ ideal should be distilled] and your absolute ambient pressure but for 1C degree accuracy those two should suffice.  Whomwever shows the proper value is likely the right measure and the others are off.

FWIW I have the same 1186a model, 2 in fact, a 1252a and several 1272a under tests; I also have a Fluke K thermocouple and a 1000F version, they are all within 1-2F of ice and boiling water.  However, the Agilent models are the fastest responding of all.

Yes, to eliminate geometry errors use a water bath.

OK I re tested, the error is still 3 degrees, mercury at 23 degrees, DMM at 20.2 degrees
As you know mercury doesn't measure negative temperatures so I wonder what would be negative temperature accuracy? Or is it linear?

Dave could make a temperature probe shootout to review various including this model thermocouple probes !!!

[nukie]

Get an Omega with proper termination for thermocouple mini plug. With all the uncompensated adapters and connectors you are introducing thermal error. Heck even my  3x 187/9 flukes are not accurate. They only give you a ball park figure. Measuring temperature is not hard but measuring accurate temperate is another game. Your lab thermometer, do they come with a cal certificate? If not, don't even trust it.

[smile]

Get an Omega with proper termination for thermocouple mini plug.

Any direct URL to product or name please.

With all the uncompensated adapters and connectors you are introducing thermal error. Heck even my  3x 187/9 flukes are not accurate. They only give you a ball park figure. Measuring temperature is not hard but measuring accurate temperate is another game.

Well seems you did not read all thread, human body temp is 36.5 degrees so my 2 mercury thermometers measure like it should be 36.5 degrees, my DMM shows 3 degrees less.

Your lab thermometer, do they come with a cal certificate? If not, don't even trust it.

No I don't have measurement certificate, but thermometers are from lab hardware that cost allot of money, so they should be accurate, both show the same temp so that is alone 99% guarantee that they do not lie. Also they show human body temp OK, so seems to be accurate unlike the DMM.

Also my DMM when shorted shows incorrect temperature, see attached picture, so it is faulty? The internal DMM thermometer and the value shorted however seems very close from each other.

[Richard W.]

Also my DMM when shorted shows incorrect temperature, see attached picture, so it is faulty? The internal DMM thermometer and the value shorted however seems very close from each other.

The internal thermometer IS the sensor for the cold-junction-compensation. So the values should not only be close, they should be equal.

[Kiriakos-GR]

@smile
Is this the way that they act in your village when some one helps them out.
I do not know if it sounds crazy but I was expecting a tiny thanks word.

Even so, when some one comments accuracy issues, I expect that he had clean first the banana plugs of the DMM unit,
with medical alcohol or other cleaning fluid with out oil in it.
 

[smile]

Also my DMM when shorted shows incorrect temperature, see attached picture, so it is faulty? The internal DMM thermometer and the value shorted however seems very close from each other.

The internal thermometer IS the sensor for the cold-junction-compensation. So the values should not only be close, they should be equal.

When one values changes the other doesn't so that makes them not the same.

[smile]

@smile
Is this the way that they act in your village when some one helps them out.
I do not know if it sounds crazy but I was expecting a tiny thanks word.

Thank you for your help, I did not notice that I have not replied to your post.

You did not answer my initial question by posting different values for different probes (very likely to happen and normal because they are different brand probes).

I was expecting same manufacturer and same type probe comparison, like agilent K trype 3 probes compared between each other. Then we could talk about probe accuracy, comparing different manufacturer probes that are different type K, J etc. only shows that reading can differ.

Even so, when some one comments accuracy issues, I expect that he had clean first the banana plugs of the DMM unit,
with medical alcohol or other cleaning fluid with out oil in it.

The meter and thermocouple is brand new

[IanB]

I don't think it has been mentioned strongly enough in this thread that thermocouples measure relative temperature, not absolute temperature. There is an independent temperature sensor inside the meter using a thermistor or some other device to establish a reference point, and the thermocouple probe temperature reading is based off this reference.

So if your meter's internal temperature reference says room temperature is 21 C, then the thermocouple will read room temperature as 21 C also. It can do nothing else.

So if you think your room temperature is really 24 C and the meter reads it as 21 C, then the meter's internal ambient temperature sensor might be out of calibration.

[Richard W.]

Also my DMM when shorted shows incorrect temperature, see attached picture, so it is faulty? The internal DMM thermometer and the value shorted however seems very close from each other.

The internal thermometer IS the sensor for the cold-junction-compensation. So the values should not only be close, they should be equal.

When one values changes the other doesn't so that makes them not the same.

My meter has internal thermometer displayed too, its 20.8 degrees
I have short circuited the meter and the values is 20.9 - 21.0

Thats a deviation of 0.2°C, caused by 8 microvolts on the input terminals. Thats negligible.
(measuring microvolts with a handheld DMM is not the best idea, but that's what you are doing when you measure with k-type sesors)

As IanB said:
The internal measured temperature is the reference for the k-Type measurement. Wrong reference, wrong k-type measurement.


By the way:

Please have a look at the second table on page 6.  0,3% + 3°C  accuracy of the meter
You have to add this to the +/- 1,1°C accuracy of the sensor.

If the real temperature is 24,0°C the displayed value can be from about 20 to 28°C (approximately)

Problem solved (maybe) 

[Zero999]

just one question: why are thermocouples used?

Surely a thermistor or temperature sensing IC are better options?

[IanB]

just one question: why are thermocouples used?

Surely a thermistor or temperature sensing IC are better options?

I think because a thermocouple probe is small and robust and has a fast time constant. You can put the probe in hostile or hard to reach places on the end of a long wire. The probe can withstand temperatures of several hundred degrees. When suitably encased the probe can be placed in hot liquids or in corrosive environments. The reasons are many...

[Kiriakos-GR]

Its true that for every type of application, there is specialized thermal probes.

And according to that rule,  the basic K type probe that come with the DMM, or you additionally get,
it serves basic measurements that high accuracy and fast temperature changes its not important. 

The topic was looking fishy from the start, because the expectations was not logical if we think that some one expects the ultimate accuracy by a 10$ probe.

Lets hope that Agilent will offer to me the true professional K type probe for review.
It retails about 180$ , and yes this one it supposed to be accurate all the way for a Year time.

The low cost basic probe it supposed to be used for heatsink temperature measurements,
somewhere between 30 to 80C .
No one cares if the heatsink  would be hot as 51 or 53C, and there is no critical difference to that, 
and so the low cost K type probes are good enough about this task.

And now I will offer one practically valuable tip.
And this called as K type probe matching.
You need to have  a minimum set of three identical K-Probes, 
You test all three , and by this comparison, you accept as best the one who gives the middle reading.
And you mark the difference with a permanent marker  on the other probes like ( example  at 20C)  +1 ( 21)  and - 0.5 ( 19.5).
And now you can use any of the three, and you known what to expect from its one.

     

 

[ejeffrey]

just one question: why are thermocouples used?

Surely a thermistor or temperature sensing IC are better options?

Thermocouples have much bigger temperature range than NTC thermistors and temperature sensing ICs.  They can be dunked in a liquid (even a conductive liquid) and still work, and they are almost indestructable due to their welded construction.  Platinum PTC thermistors have a similarly large operating temperature range, but they require an excitation current and often a 4-wire resistance measurement due to their low dR/dT (around 0.4 ohm/kelvin for a Pt100). 

The only disadvantage of thermocouples is that they have relatively low accuracy, especially when using sub-optimal cold junction compensation that you will find in many DMMs.  This is usually not as big of a problem as people think.  In measurements of gases or solids, other factors will often dominate.  Temperature gradients and fluctuations mean the temperature at the probe tip may not be particularly meaningful, and direct heating or cooling of the probe tip by evaporation, air currents, light absorption, or conduction along the probe body can all create intrinsic errors in the measurement.  PTCs and ICs can have their otherwise better accuracy compromised by self heating due to the operating current.  Liquid measurements are much better behaved, at least if the bath is stirred sufficiently, but can still be problematic.

[greatseo]

Is this the way that they act in your village when some one helps them out.
I do not know if it sounds crazy but I was expecting a tiny thanks word.

Even so, when some one comments accuracy issues, I expect that he had clean first the banana plugs of the DMM unit,
with medical alcohol or other cleaning fluid with out oil in it.

[SgtRock]

Dear Smile:

--You might want to generate plotted curves showing the Agilent reading vs your Mercury reading. If the space between the two curves is relatively consistent, adjusting the calibrated temperature should help. If the lines diverge or cross you will just have to do the best curve fitting you can, making the part of the range you use the most as accurate as you can, and using a fudge factor for the rest.

--You might also consider getting one of these "Temperature range: -58 to 482 °F(-50 to 250 °C); Accuracy: +/- 1C (between -20, 80 °C), or +/- 2 ° (out side of -20, 80 °C)", see below. They are quite handy in the kitchen. BTW, I am betting DJ uses a Fluke to check the baby bottle temperature. :

http://www.ebay.com/itm/Digital-Thermometer-SS-Sensor-Memory-Kitchen-Food-HVAC-/370378620912?pt=LH_DefaultDomain_0&hash=item563c493ff0

To Greatseo:

--"immigration attorney"

--Welcome to the EEVBLOG Parade Grounds. I just wanted to tell you, that you are starting off on the right foot, that is, the wrong foot. In the Air Force we used to make troops that started of on the right foot carry a 5 pound rock in their left hand for a day or two. In other words Sir or Ma'am as the case may be, please try to get with the program.

--This is truly an international forum. Many of the people posting here are Electrical Engineers and Electronics Professionals for whom English is a second or third language. Now DJ, Simon and the other administrators are not likely to censor you, but you are not going to make many friends, belittling folks  for  their  nationality  or  ethnic  origins.  I  will  let  it  go  this  time  and  wish  you  a  hearty   "- - - - .   - - . -"


Best Regards
Clear Ether