Accuracy specification (or a good thermometer for calibrating DS18b20's)

[foufee]

Hi,
  I'm looking at getting a calibrated thermometer to provide a reference whilst calibrating DS18B20's, and was looking at http://nz.rs-online.com/web/p/digital-thermometers/4109066 (manual: http://www.chauvin-arnoux.com/sites/default/files/D00AZW86_24.PDF ) but the line about accuracy is throwing me off a bit ( ±0.1% of reading +1°C from -40°C to 1350°C) , is this saying that there is a 1°C offset error and a 0.1% measurement error, or is it more likely 0.1% + 1 least significant digit (so 0.1°C when at the highest accuracy setting).
 
  Or does any have any suggestions of a good (and reasonable priced) reference thermometer?
 
  Cheers

[DTJ]

It's a thermocouple meter so unless you calibrate the meter plus the thermocouple as a pair you will have an error of >1 C.

Look for a unit with an integrated thermistor or RTD probe. You mentioned the 18B20’s so I assume you are working with temps say 0 to 100C? A possible option is a certified thermometer (they used to be around $100). Alternatively build something using one of the more expensive calibrated digital sensors or buy one of their dev kits.

[uncle_bob]

Hi,
  I'm looking at getting a calibrated thermometer to provide a reference whilst calibrating DS18B20's, and was looking at http://nz.rs-online.com/web/p/digital-thermometers/4109066 (manual: http://www.chauvin-arnoux.com/sites/default/files/D00AZW86_24.PDF ) but the line about accuracy is throwing me off a bit ( ±0.1% of reading +1°C from -40°C to 1350°C) , is this saying that there is a 1°C offset error and a 0.1% measurement error, or is it more likely 0.1% + 1 least significant digit (so 0.1°C when at the highest accuracy setting).
 
  Or does any have any suggestions of a good (and reasonable priced) reference thermometer?
 
  Cheers

Hi

How close are you trying to get?

Do you simply want all the product to be the same or is there a real accuracy spec?

What temperature range are you trying to cover?

Does your budget cover $100K, $10K, $1K or $100? There are very real reasons why you would spend at each of those levels depending on what your real needs were.

Temperature is one area where you *can* afford to buy a primary standard. It's good at one temperature (0C) but it's a start. It's also the basis for a lot of systems. You can also *make* a fairly good triple point cell without a massive lot of effort. It will not be cal lab rated, but it will be a lot more accurate than you probably need.

Bob

[Galaxyrise]

You can also *make* a fairly good triple point cell without a massive lot of effort. It will not be cal lab rated, but it will be a lot more accurate than you probably need.

Do you have more on that? I tried to figure out what would be involved in making one, but I couldn't invent anything that was practical for me to do on my own.

[foufee]

We're trying to get to less than 0.5°C, and prepared to spend a few hundred dollars on kit to help achieve it. The temperature range we need to cover is -10°C to about 40°C

Cheers.
 

Hi,
  I'm looking at getting a calibrated thermometer to provide a reference whilst calibrating DS18B20's, and was looking at http://nz.rs-online.com/web/p/digital-thermometers/4109066 (manual: http://www.chauvin-arnoux.com/sites/default/files/D00AZW86_24.PDF ) but the line about accuracy is throwing me off a bit ( ±0.1% of reading +1°C from -40°C to 1350°C) , is this saying that there is a 1°C offset error and a 0.1% measurement error, or is it more likely 0.1% + 1 least significant digit (so 0.1°C when at the highest accuracy setting).
 
  Or does any have any suggestions of a good (and reasonable priced) reference thermometer?
 
  Cheers

Hi

How close are you trying to get?

Do you simply want all the product to be the same or is there a real accuracy spec?

What temperature range are you trying to cover?

Does your budget cover $100K, $10K, $1K or $100? There are very real reasons why you would spend at each of those levels depending on what your real needs were.

Temperature is one area where you *can* afford to buy a primary standard. It's good at one temperature (0C) but it's a start. It's also the basis for a lot of systems. You can also *make* a fairly good triple point cell without a massive lot of effort. It will not be cal lab rated, but it will be a lot more accurate than you probably need.

Bob

[DTJ]

Perhaps look at Sensirions development kits for temp or humidity (does temp as well).

https://www.sensirion.com/products/humidity-sensors/evaluation-kits/

I have a EK-H5 kit which I think I paid $40 and is good for +/-0.3°C
It has the sensor, USB adapter & software to display temp / RH.

They have other sensors @ +/-0.2°C.

I got mine from RS components.

[uncle_bob]

You can also *make* a fairly good triple point cell without a massive lot of effort. It will not be cal lab rated, but it will be a lot more accurate than you probably need.

Do you have more on that? I tried to figure out what would be involved in making one, but I couldn't invent anything that was practical for me to do on my own.

Hi

Google will turn up a number of pages on doing triple point cells. The typically approach is to use test tubes as part of the structure and a bit of glass blowing to fuse things. You fill the empty cell with as good distilled water as you can get your hands on. Heat it up to boiling for a while and then melt the pinch off tube. As it cools you get a vacuum since the headspace was filled with steam rather than air. You can also do the enclosure with a variety of materials. Whatever you use needs to be "boil compatible" unless you have a good vacuum pump.

Bob

[uncle_bob]

We're trying to get to less than 0.5°C, and prepared to spend a few hundred dollars on kit to help achieve it. The temperature range we need to cover is -10°C to about 40°C

Cheers.
 

Hi

There are a lot of ways to go for something like that. I'd look at Fluke.

Bob

[Paul Moir]

Something to keep in the back of your head is that you can do this with a higher accuracy RTD and a good multimeter which can do 4 wire measurements.  Some bench multimeters will read in degrees if you tell them you have an RTD plugged in.  So if you need or have a good multimeter anyway, this might be a solution.