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

[IanB]

To Greatseo: ...

Did you notice that this post was a spam post? It is an exact copy-and-paste of what Kiriakos wrote a few posts back... 

[Kiriakos-GR]

With off topic gossip  the post counter rises fast UV.
Keep up the good work.  LOL

[SgtRock]

Dear IanB:
--No I did not realize it was a spam post. Now I see that you could be correct. Just how does that work anyway. Do they have a Spambot that looks for imperfect English grammar and syntax and the joins forums in order to post spam links. Or do the have humans posting spam for a dollar a lick or what. Just how do they do this sort of thing. I am not being sarcastic. It certainly looks like spam, now that I think about it, but I cannot figure out how they do it profitably.

--Back on topic; I have two Fluke 87s with no temperature capability, that is why I bought a kitchen thermometer with remote sensing probe. They are cheap, reasonably accurate, and free up your DMM to do something else at the same time. And most of them even have an alarm to let you know when you have reached the temperature.  Most also function as stopwatches and alarm timers. One should be very helpful on any bench.

"Three weeks in the lab will save you a day in the library every time" Stanley Williams, HP Labs

Best Regards
Clear Ether

[smile]

Thanks for everyone help, I try to get some more thermometers to compare to.

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) 

Yes the deviation of the meter is like you said 0,3% + 3°C + +/- 1,1°C accuracy of the sensor, that really is crap for such an expensive meter that is. And what is the point of using expensive (like @Kiriakos-GR said) probe for 180$ if the meter will be the bottleneck then ?

Still better to use mercury thermometers to measure accurately then what a nonsense in 21st century, and inconvenience.

[ejeffrey]

Yes the deviation of the meter is like you said 0,3% + 3°C + +/- 1,1°C accuracy of the sensor, that really is crap for such an expensive meter that is. And what is the point of using expensive (like @Kiriakos-GR said) probe for 180$ if the meter will be the bottleneck then ?

Damn.  I knew that thermocouple measurement was a bit of a 'we can do that free' feature on DMMs, but I didn't expect a 3C offset.  An omega temperature controller will do a much better job, of course at a cost.  If you want to measure temperature more accurately on your DMM you could get a Pt1000 and measure it on ohms mode.  You will have to look up the conversion yourself.  Pt100 is more common but when measured with a standard DMM will have larger errors due to contact resistance.  Even a Pt100 will be good to 0.5 C or so.

[alm]

Yes the deviation of the meter is like you said 0,3% + 3°C + +/- 1,1°C accuracy of the sensor, that really is crap for such an expensive meter that is. And what is the point of using expensive (like @Kiriakos-GR said) probe for 180$ if the meter will be the bottleneck then ?

Yeah, temperature measurement is crap, even in expensive meters. Same for capacitance, actually. Or frequency. Or inductance, if available at all. Basically anything except the classic VDC/VAC (low frequency)/R/IDC/IAC/diode/continuity features.

Fluke (and some others) also make dedicated thermometers, my guess is that those are better, although K-type has always been the cheap, not so accurate option.

[baljemmett]

--No I did not realize it was a spam post. Now I see that you could be correct. Just how does that work anyway. Do they have a Spambot that looks for imperfect English grammar and syntax and the joins forums in order to post spam links. Or do the have humans posting spam for a dollar a lick or what. Just how do they do this sort of thing. I am not being sarcastic. It certainly looks like spam, now that I think about it, but I cannot figure out how they do it profitably.

There's no 'could' about it -- it's a spam post.  The bot algorithm is (roughly) join forum, find active discussion, copy an extract from a little way back in the thread, post it as fresh reply with spammer's link in the footer.  It works because people don't always notice it's just parroting something already said, since (usually!) the 'new' contribution will be related to the subject being discussed, and so the link gets to sit there gathering search engine legitimacy and maybe even the odd click.  Setting 'nofollow' on links helps with the search engine bit but I don't think SMF implements that.

Unfortunately in this case the bot happened to pick a tasty piece of flamebait instead of anything with any particular relevance, and so stood out like a sore thumb -- your reaction to "greatseo"'s perceived rudeness is ample demonstration of that!  Come to think of it, that's a bloody daft choice of name for a spambot, rather gives the game away...  I would imagine if there was a human in the feedback loop they might have chosen a less obvious section of text to copy.

Back to thermocouples, huh?

[SgtRock]

Dear Baljemmet:

--Thank for the informative post. I guess I did get the hook caught in my lip on that one. The depth of my ignorance is very great in this area. I was unable find a meaning for SMF except for the obvious crude one, which is not what you meant, I do not think. I promise to keep my next post, strictly on topic.

"Three weeks in the lab will save you a day in the library every time" Stanley Williams, HP Labs

Best Regards
Clear Ether

[Kiriakos-GR]

Unfortunately in this case the bot happened to pick a tasty piece of flamebait instead of anything with any particular relevance,........
...........Back to thermocouples, huh?

I had no idea that it was a bot, I had just think that some one was attacking to me. LOL 
Now about the flame-bait, I love this forum so much, and I do not use this tactics,
but some times the small misunderstandings its not avoidable.   

Yes back to the thermocouples.

The U1272A  has 1% +1C accuracy.
If this thread helped me in a way, was that from now and on I will just add this damn 1C at every temperature measurement of my.

[SgtRock]

Dear Smile:

--Please disregard the erroneous comments that were previously in this post.  See below for the links to the specifications of your meter and thermocouple.

--Agilent U1252B specifications:

http://cp.literature.agilent.com/litweb/pdf/5989-5509EN.pdf

--U1186A Thermocouple (K-Type) and Temperature Probe Adapter specifications:

http://www.home.agilent.com/agilent/product.jspx?cc=US&lc=eng&nid=-536902467.714195&pageMode=OV

--Let us assume for the moment that your mercury thermometer is completely accurate and measures 24.0 C. The temperature accuracy specifications for the U1252B are +- (0.3% + 3) degrees C. The temperature accuracy specifications for the U1186A  K-type Thermocouple are +- 1.1%.

--The worst case senario analysis of the above specifications, assuming a true temperature of 24.0 C, gives the following range of possible readings that are within specifications: 19.8 to 28.1. So you meter would appear to be within specification with regard to temperature.

--Many thanks to ejeffery for pointing out the errors that were previously in this post.

"Three weeks in the lab will save you a day in the library every time" Stanley Williams, HP Labs

Best Regards
Clear Ether

[Kiriakos-GR]

Hey give a hand to the poor electrician .. 

My DMM  specifications :  Accuracy: ±1% +1°C
And the same goes for the thermocouple ±1% +1°C


Is it logical to assume that I had to add 1C ( DMM side) + 1C ( K-probe) = total compensation +2C  ?   

[ejeffrey]

thermocouple are off by too much, but that the reading is off in the wrong direction. I will explain. See below for the links to the specifications of your meter and thermocouple.

No, you are misreading it.  The error specified is the magnitude, it can be off in either direction.  The "+" means that you have to add the fractional uncertainty (1%) with the offset error (3C) to get the maximum magnitude of the error.

[SgtRock]

Dear Ejeffery:

I see upon review that you are nearly completely correct. I will be correcting my erroneous post immediatly. Agilent should have stated +-(0.3% +3C). My Fluke 87 manual is even harder to read and refers to the "lease" significant digit. Thank for catching my egregious error so soon.

Best Regards
Clear Ether

[saturation]

I took a quick ice water photo, I presume you know what errors can occur if its not ice shavings; the temp is not homogenous as it equilibrates accross the cubes and the container, so spots are warmer than others.  However, this shows at least 2 meters including a separate IR reading independent of the thermocouple, read ~ 1C of each other.

Likewise, the ambient reading of the 2 Agilent's are within 1C of each other.

[Kiriakos-GR]

Agilent is very clear about temperature measurements in the user manual of the U1272A .

It forbids the immersion in water of the Teflon yellow K probe.

For temperature measurements of Air / liquids / metals, there is special made probes.
The mini probe in the device, needs a minimum of 60 min, so to adjust at the environmental temperature.. 
More info page 66 .