Thermometers: What's the Problem with Accuracy?

[rbowman]

From what I've seen as far as weather goes fractional values aren't used. It's -2 in Berlin at 0500 and will hit a high of 2 with -6 overnight on Tuesday

[CRNG]

On Sun, 04 Feb 2018 11:51:08 -0500, "(PeteCresswell)" wrote in

For cheap thermometers, that's not bad. Just take the average of the four and then calculate the difference between the average and each of the four and use that as a correction factor for each of the four. That will probably bring you to a 99.9% accuracy for each of the four.

[gregz]

Best way of testing bulb thermometers is insert in water beaker with stirring mechanism for rapid low. Add ice should be hear 32 degrees. Air currents can be tricky and scales can slip. +- 1 degree typical variance. My two outside units are at least 3 degrees off. Can't find easy way to cal. Some resistors probably need tweaked.

Greg

[Cindy Hamilton]

They would just say 10. At

the current temperature is -14 and today's high is expected to be -5. That's for Windsor, which is a short drive from where I am.

Cindy Hamilton

[Meanie]

You have a refrigerator/freezer reading in the high 60s to low 70s. You have bigger problems than an inaccurate thermometer.

[Art Todesco]

My new Samsung fridge has built-in digital readings. But it always seems to read the same as the setting, that is, 0 for the freezer and 37 for the fridge. I have seen the numbers change if, for instance, you leave the door open for a long time while cleaning. But, because I can't trust the microprocessor in the Samsung, not to lock up and shut down everything and thus spoil the contents, I bought a unit that magnet mounts outside the box and has a senor inside. Its reading vary all over the place. Say at dinner prep time, when the box gets opened and closes many times, I will see maybe a rise of 6 or 8 degrees. I think the Samsung sensor is more thermally loaded and takes longer to change. My sensor it under a shelf but taped tightly to the shelf ... but it still changes more quickly. Thermal mass seems to be the difference. BTW, the unit I bought, reports via wifi, once per hour (or whatever I set it to do) to the company's cloud server. If the temp goes outside the range I've set, it will fire me a text and email ... lot's of good if you're vacationing in Europe, but at least you can call someone, or change the email to them. The unit was only $120 including the service, which is certainly cheaper than all the wasted food.

[TimR]

So when it calls for heat it closes a contact?

Ingenious.

Um, what's the dead band set to?

[dpb]

On 2/5/2018 12:10 AM, Bod wrote: ...

It's never become inate for me;

343 C just doesn't correlate to Tsat@2200, sorry...yes, I know they cross at -40 and STP and all that, but it "just ain't natural". :)

[trader_4]

Same here. And like you my background is science and engineering too.

[dpb]

Yeah, being as it is designed to be a thermostat it has a builtin-contact onboard. I debated which way to go; could use a green LED and have it lit when OK but decided to go the red and be normally off and turn it on when indicated a problem. I wasn't sure at first but the LED is bright enough it's easily visible in the wellhouse window when on even during middle of the day. I haven't put it on a battery or used a UPS at the moment so if we're out of power I still have to check, but it covers the bulk of the time...

I've left it at default of 2 C; it's adjustable, too.

I was going to comment that had to compute where wanted setpoint to start for the user interface in C; at the price point you can't expect much documentation but I just presumed when ordering since it's all digital there would be a F/C option in setup--turns out there's not, at least in this one. I'm sure they must be out there...

A link

[trader_4]

Same here with a Kitchenaid. I think they have an algorithm where it will only change slowly, not in response to every door opening. If people saw it going from 37 to 42, varying, a lot of people would probably think it's broken. Also, staying on 37 makes the fridge look better, seem rock solid, etc.

[(PeteCresswell)]

Per Boris:

You are not alone.

[Mark Lloyd]

[snip]

I've seen digital thermometers that show an extra digit (.1 degree). However, this additional precision is useless because of the limited accuracy. They seem to be taking advantage of the fast that a lot of people don't know the difference.

[(PeteCresswell)]

Per dpb:

What type would you say is the least prone to inaccuracy?

Bi-metallic spring?

Digital?

I have played around with digital sensors and a Raspberry PI and the accuracy was pretty bad.... but maybe that's just el-cheapo sensors and/or my programming approach...

[gfretwell]

Where you live anyway. I speak fluent celsius but people just c*ck their head here when I say the water was 22.8 this morning.

[gfretwell]

Why do you have a queen?

[Sam E]

[snip]

Or 11, which is closer to 10.6. I've been rounding numbers so long it's almost automatic.

[dpb]

On 2/5/2018 12:41 PM, Bod wrote: ...

Have a nephew who was European Sales rep for Cessna who lived within about that same radius...he's since returned to the States but we were there a couple times while I was doing the coal flow testing at Kingsnorth Station over in the Rochester/Chatham area...

[(PeteCresswell)]

Per Art Todesco:

How about a link?

[Tekkie®]

[e-mail] posted for all of us...

What's his name, Bod?