EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: chofroa on February 11, 2021, 05:59:50 pm
-
Hello and many thanks to all writers and testers here for helping me to buy a good dmm :)
as 'winter's coming', I decided to control a basic interior thermometer and was surprised to read a great difference : about 2°C less on the bm829s with its K probe, correctly inserted (+/-).
Comparing with other indicators, it seems that the thermocouple K is outside the average.
is there a way to correct this ?
is it a manufacturing issue ?
thanks in advance !
(http://pix.toile-libre.org/upload/original/1613067322.jpg)
ps: hope it's the right place to post :)
-
A thermocouple is not very precise and always measures relative temperature. There is a sensor inside each thermocouple meter, if you short the input terminals you will see the temperature from it (You may have to wait a couple of minutes before it is correct).
-
The meter's spec is +/- 1.5 degrees Celcius.
Add in the error from the probe and that's the way it goes.
You could try another thermocouple and see if it helps. hook up the probe from the yellow meter (or just poke it with the Brymen's probes) and see what happens.
-
Thermocouple which is used does not matter in this case. As everything stays at the same room temperature, thermal EMF generated by it is basically 0. But you need to ensure that meter stayed in the same place for a few tens of minutes to ensure its temperature is equal to room temperature to measure accurately.
-
thanks all :)
@HKJ
trying to short and the result, after some minutes: not better :(
(http://pix.toile-libre.org/upload/original/1613084990.jpg)
@Fungus
with the 'yellow' probe : not better again :(
(http://pix.toile-libre.org/upload/original/1613084836.jpg)
@wraper
I did wait some minutes before taking the photos
-
@HKJ
trying to short and the result, after some minutes: not better :(
(http://pix.toile-libre.org/upload/original/1613084990.jpg)
This shows the problem: The internal sensor is 2 to 3C off in calibration.
-
Hello HKJ,
This shows the problem: The internal sensor is 2 to 3C off in calibration.
. as it's a new one, does that mean the manufacturer doesn't calibrate at the end of producing ?
. any way to solve this ? by software ?
. worth contacting Brymen ?
-
Hello HKJ,
This shows the problem: The internal sensor is 2 to 3C off in calibration.
. as it's a new one, does that mean the manufacturer doesn't calibrate at the end of producing ?
. any way to solve this ? by software ?
. worth contacting Brymen ?
First: The 2-3C is an estimate and not very precise, use a single piece of copper wire to connect the two terminals for a better estimate. Place the meter and a few other thermometers in a place without any air movement, let them rest for an hour or two (They need not be on). Then turn on, wait a bit more and check without moving them from the place.
Next: Check the datasheet, how precise is the meter supposed to be and how precise is your other thermometers.
It may be out of calibration.
Meters can be (re)calibrate with the right equipment (I believe Brymen (mostly) uses electronic calibration) and knowledge, but I doubt there are many dealers that can calibrate a Brymen DMM, the other solution is that you dealer replaces the meter with a better calibrated one.
Note: The internal temperature sensor is often inside the DMM chip, i.e. it is not an option to replace it.
-
. as it's a new one, does that mean the manufacturer doesn't calibrate at the end of producing ?
If the written specification is 1.5 degrees of error then it probably doesn't use a very calibrateable part.
eg. The reference might just be the horrible onboard temperature sensor of the MCU or something, not a separate sensor chip.
-
KKJ, Fungus,
with a short cupper :
(http://pix.toile-libre.org/upload/original/1613146291.jpg)
and the caracteristics :
the yellow :
(http://pix.toile-libre.org/upload/original/1613146424.jpg)
the red :
(http://pix.toile-libre.org/upload/original/1613146512.jpg)
(http://pix.toile-libre.org/upload/original/1613146551.jpg)
so I guess I've to do with it :(
I'll search what kind of thermometer is trustable :)
thanks again !
-
Another test you could do is to place the thermocouple probe into a pot of boiling water and see how close it gets to 100°C? Or in a bowl of melting ice for 0°C?
A thermocouple probe is calibrated over a range and it may be more accurate at other temperatures further away from room temperature.
-
Another test you could do is to place the thermocouple probe into a pot of boiling water and see how close it gets to 100°C? Or in a bowl of melting ice for 0°C?
A thermocouple probe is calibrated over a range and it may be more accurate at other temperatures further away from room temperature.
It is not the scale that is discussed, but the compensation (cold junction) sensor. It has to be fairly accurate for the meter to work around ambient temperature.
-
with a short cupper :
It do definitely look like it is out of calibration, try sending the picture to your dealer or maybe to Brymen and ask about it.
You could also look for some sort of thermosensor inside the meter (I doubt you will find any, but I have not checked), there is tear down photos here: https://lygte-info.dk/review/DMMBrymen%20BM829s%20UK.html (https://lygte-info.dk/review/DMMBrymen%20BM829s%20UK.html). I there is a NTC resistor, placing a regular resistor parallel with it may mostly fix your problem (At least around 20C).
-
It is not the scale that is discussed, but the compensation (cold junction) sensor. It has to be fairly accurate for the meter to work around ambient temperature.
I don't disagree about that, but I have a bunch of meters with thermocouple inputs and it is quite hard to get them all to agree. I get at least 1 or 2 degrees variation between instruments at ambient conditions. There is also some variation between probes in the same instrument.
-
It is not the scale that is discussed, but the compensation (cold junction) sensor. It has to be fairly accurate for the meter to work around ambient temperature.
I don't disagree about that, but I have a bunch of meters with thermocouple inputs and it is quite hard to get them all to agree. I get at least 1 or 2 degrees variation between instruments at ambient conditions. There is also some variation between probes in the same instrument.
Why do you think I mentioned the specifications, +/-2C is not a unusual tolerance for a thermocouple input and this meter has that, so it is not that much out of tolerance and the simple test method is not that precise either, but it looks like it is outside specified tolerance.
The general problem with (testing) thermocouple and inputs is that everything is a thermocouple and you have to be very sure that the temperature is equalized everywhere to avoid problems with that. I did write an article about DMMs and thermocouples here: https://lygte-info.dk/info/DMMThermocouplers%20UK.html (https://lygte-info.dk/info/DMMThermocouplers%20UK.html) it mentions some of the pitfalls with them.
I had a bit of fun making a thermocouple input here: https://lygte-info.dk/project/ThermoSensor%20UK.html (https://lygte-info.dk/project/ThermoSensor%20UK.html)
-
T1 and T2 probably have the same cold junction circuitry, but it may be worthwhile testing T2 to see if you get the same error as T1.
-
T1 and T2 probably have the same cold junction circuitry, but it may be worthwhile testing T2 to see if you get the same error as T1.
I hope that they have the same cold junction compensation or the idea behind two sensors in one meter is serious flawed.
Note: The idea is that the difference is fairly precise and not affected by ambient temperature.
-
I don't disagree about that, but I have a bunch of meters with thermocouple inputs and it is quite hard to get them all to agree.
I've got cheapo meters that are bang on and other meters that are 2 degrees off.
It's the luck of the draw, but not "out of spec".
I hope that they have the same cold junction compensation or the idea behind two sensors in one meter is serious flawed.
Yep. I imagine the difference is quite accurate on a BM869.
(All you need is a very accurate little oven to put one of the probes in :) )
-
(All you need is a very accurate little oven to put one of the probes in :) )
No owen, you want to get the exact temperature of the two connection. This do not what you see on a DMM, but you get as close as possible with the design.
A sensor at the plugs are a very good idea, but in praxis it is no very common.
-
hello and thanks again !
changing T1 > T2
(http://pix.toile-libre.org/upload/original/1613235287.jpg)
(http://pix.toile-libre.org/upload/original/1613235346.jpg)
and testing boiling water
(http://pix.toile-libre.org/upload/original/1613235394.jpg)
'gobal warming' effect ? ???