Topic: How accurate is a 5n1

[nincehelser]

Are you able to use one of the acurite tower style thermo/hygro sensors? IF so, I would go this route if you feel the all in one is not delivering good data. I have done this myself... However, I do not condone putting the sensor in a tree like Acurite, or over walkways etc like Nincehelser. I suggest a proper shield from Ambient weather, Davus; or make one from wood painted white, or some plastic large bowls inverted and painted black inside, white outside.

Good grief, Dr. Know.  I don't have any sensors over walkways.

[nincehelser]

I have done that.  The tower sensor always read lower by about 1.5 degrees than the 5 in 1.  My tower sensor didn't have humidity on it and I didn't want to use it's temperature and the 5 in 1 humidity for reasons we talked about before.

So you've been measuring 1.5F between two sensors in the same location?  That's certainly better than 2 to 2.5 you say you're seeing between random stations.

You could try calibrating them to each other and see if they hold well together over time.

[Inverno]

Yes, the problem is the 5 in 1 is higher in sun and it costs $140 more than a little tower sensor? To me, that's an accuracy problem.

[nincehelser]

Yes, the problem is the 5 in 1 is higher in sun and it costs $140 more than a little tower sensor? To me, that's an accuracy problem.

If I remember correctly, the SHT21 has about +/-1F accuracy.   Comparing to an uncalibrated tower (which is likely less accurate), you're at 1.5F difference.  You're dabbling in the range of uncertainty.  Calibrating the sensor and recomparing should give you a clearer picture of what's going.

You could try replacing or upgrading the Sensirion chip in the 5n1.

A 5n1 sensor is more like $90.  You can pick up a 5n1 sensor and new display at Walmart for $99.

[Inverno]

How much are the chips? I was thinking maybe a new single panel solar panel would help. Mine is almost 4 years old.

[nincehelser]

How much are the chips? I was thinking maybe a new single panel solar panel would help. Mine is almost 4 years old.

If it's that old, you might have to open it up and check to make sure what you've got.  I believe they started using the SHT21 in 2012.  Before that was something incompatible, but folks have replaced them, too.

The SHT21 isn't terribly expensive.  You can find them for under $10 without much of a problem.  Soldering can be a challenge, though.

The SHT25 costs a bit more, but I hear it can get you +/- .5F accuracy.  It's said to be a drop-in upgrade.

There are also compatible chips from folks like Sparkfun, but I've never looked into their specs.

Factory-fresh models have made the temp sensor assembly user-replaceable without soldering, but that probably doesn't help much in your case.

[haroldashe]

As George said, my 5in1 and tower usually run no more than 2 degrees apart, usually within 1 to 1.5.  The tower is in a large shade tree with an inverted plastic bowl over it, mainly to keep some of the rain off.

 [ You are not allowed to view attachments ]

[Inverno]

I had the same setup Harold, exact same. My tower was always cooler in the sun but about 1.5 degrees and within 0.3-0.5 at night. The issue I have is why is my $150 device less accurate with temperature than my $10 device.

[haroldashe]

Dunno...my 5in1 right now (8:13 AM) is exactly 1 deg. cooler than the tower (72.0 vs 73.0).  I suspect it's because the shade for the tower is not perfect and I expect some minor temp fluctuation.

[DoctorKnow]

My sensors are both in the sun this morning, and my 5 in 1 is 77.2, the tower is 77.9. - I see nothing unusual. The tower is sited to NOAA standard. The 5 in 1 is temporarily at 15 ft until I can get it back up on the top of the pole from it's previous roof experiment. Keep in mind the tower uses a thermistor, while the 5 in 1 uses a digital humidity chip to take it's temperature.

[haroldashe]

I guess I'm not "hung up" about a 1-2 deg. temp difference. And, besides, who knows which sensor is more "accurate" anyway? The 5in1 is close enough so that I'm never kicked off WUG or CWOP.  Below is a CWOP screen shot showing me (EW6192) compared with several other stations in the area.

 [ You are not allowed to view attachments ]

[nincehelser]

I had the same setup Harold, exact same. My tower was always cooler in the sun but about 1.5 degrees and within 0.3-0.5 at night. The issue I have is why is my $150 device less accurate with temperature than my $10 device.

Saying the $10 device is more accurate would not be correct.

You really won't know true accuracy until you calibrate to something known to be true (e.g. a NIST-traceable device).  Right now you just know you're getting a reasonable measurement given the uncertainties involved.

[nincehelser]

I guess I'm not "hung up" about a 1-2 deg. temp difference. And, besides, who knows which sensor is more "accurate" anyway? The 5in1 is close enough so that I'm never kicked off WUG or CWOP.  Below is a CWOP screen shot showing me (EW6192) compared with several other stations in the area.

That's exactly right.  Things like WUG and CWOP work on a "consensus"  basis.  As long as the stations involved seem to be agreeable within defined limits, they declare it "good enough".  It doesn't actually mean your reading is "correct".

It's a form of "groupthink", which you probably know isn't always a good thing. 

[Inverno]

The fact that I'm regularly higher than most stations that are not Acurite in my area, I'm going to assume that it is off by an amount of around 2 degrees. As for others being accurate, I've driven by them with my car thermometer, they are spot on, and usually my car reads lower than my 5 in 1 when parked very close.  Too many examples of the 5 in 1 being off by a couple degrees to ignore in my view. It baffles me that a chip cannot be more accurate or maybe the tunnel with the fan heats up too much. I know this subject has been discussed a lot, it's just frustrating for someone that is a perfectionist.

[DoctorKnow]

Inverno,

If you are willing to try an experiment, take a screw slightly larger than the vent holes for the intake at the bottom, and enlarge 3 or 4 of them just below the tiny black sensor on the chip on the inside that you can see by looking though the intake. I did this with mine, and I believe this is what finally fixed my spikes. I am sure you have read my complaints in the past as well. Everyone kept saying, fan fan fan fan, but it just didn't make sense that so many could all be a fan issue.

OUTDOOR (bottom line reading) Is the 5 in 1.... Here are my readings as of 2 PM full sun

OUTDOOR #7 is in the sunshield.

Code: [Select]

Outdoor #5 9 sec32.9°C 57% (dew 23.2°C, heat 38.2°C)91.2°F 57% (dew 73.8°F, heat 100.8°F)

Outdoor #7 11 sec33.5°C 57% (dew 23.8°C, heat 39.7°C)92.3°F 57% (dew 74.8°F, heat 103.4°F)

5 IN 1 Outdoor 6 sec 32.5°C 65% (dew 25.1°C, heat 40.1°C)90.5°F 65% (dew 77.2°F, heat 104.1°F)

[nincehelser]

The fact that I'm regularly higher than most stations that are not Acurite in my area, I'm going to assume that it is off by an amount of around 2 degrees. As for others being accurate, I've driven by them with my car thermometer, they are spot on, and usually my car reads lower than my 5 in 1 when parked very close.  Too many examples of the 5 in 1 being off by a couple degrees to ignore in my view. It baffles me that a chip cannot be more accurate or maybe the tunnel with the fan heats up too much. I know this subject has been discussed a lot, it's just frustrating for someone that is a perfectionist.

Sorry you feel that way.  All I can suggest is trying calibration, but a +/- 1F variance from true is a fairly common spec. 

You can, of course, buy chips tested to higher accuracy (e.g. SHT25).  But frankly, if all the stations near me started reporting the exact same temp on a hot day, I'd suspect something was awry.  The world just isn't that perfect.

[haroldashe]

My high temp today as reported by my 5in1 was 96.3.  Whether it was really 95 or 97 is not that important to me...still too hot to play golf and/or walk the dogs.  And, for what it's worth, the little tower sensor in the birch tree also reported a high of 96.

[nincehelser]

My high temp today as reported by my 5in1 was 96.3.  Whether it was really 95 or 97 is not that important to me...still too hot to play golf and/or walk the dogs.  And, for what it's worth, the little tower sensor in the birch tree also reported a high of 96.

Yep.  It's hot.

I'm on the road, but I got a heat index alert email today from my house in Texas.  I'd never seen one before, and at first I assumed it was from Nest, so I checked Nest, and it was reporting an outside temp of 101 from whatever internet source they use.  (I think it's based on zip code, but I'm not sure)

Turns out it was the alert was actually from my new 5n1 system, so I checked the outside temp it recorded at the time of the alert.... 101.   Can't complain about that.

[haroldashe]

We're supposed to be over 100 for 5 days beginning 7/7/16 and little chance of rain.  Much as I hate to do it, I guess it will be time to turn on the irrigation system.  As usual, I planted my "farm" in the bark beds along the back yard fence...tomatoes, squash, peppers & water melon.  It's amazing what you can grow in a small space!

[DoctorKnow]

One thing I am very pleased with is the humidity reading. The windows are drenched in water from how high the Dewpoint is outside, and the humidity reading is 100%. You can see the dewpoint running over the temperature reading very nicely. Only the airport is showing 100% besides mine...

[Inverno]

Doctor Know.... How did you ensure that the holes didn't crack? So you took a screw larger than the small holes and made them larger?

George/Harold... If you don't want exact accuracy, why even have a station? I want to know what the temp is at my house with the greatest accuracy possible. If I just cared that it was too hot for golf or lawn work, I'd just use the Accuweather app or watch my local news.

[DoctorKnow]

"Doctor Know.... How did you ensure that the holes didn't crack? So you took a screw larger than the small holes and made them larger?"

The plastic is not brittle. I didn't use a huge screw, just one slightly larger than the hole to allow more air, and I only did it to the four holes that are directly below the tiny sensor. I suppose you could enlarge more, but I don't see that it will be necessary. Just carefully turn the screw in and out.

[Inverno]

Mine is too brittle. It's 3.5 years old. I began drilling a small screw in and heard it splitting so I stopped.

[nincehelser]

Doctor Know.... How did you ensure that the holes didn't crack? So you took a screw larger than the small holes and made them larger?

George/Harold... If you don't want exact accuracy, why even have a station? I want to know what the temp is at my house with the greatest accuracy possible. If I just cared that it was too hot for golf or lawn work, I'd just use the Accuweather app or watch my local news.

You don't need "exact accuracy" to have useful data.  +/- 1F is usually good enough for most purposes. 

It's impossible to have "exact accuracy", anyway.  The best you can do is keep chipping away at the uncertainties involved.

I wouldn't bother with enlarging the holes.  You could probably remove the entire grill and not see a verifiable difference. 

[DoctorKnow]

Mine is too brittle. It's 3.5 years old. I began drilling a small screw in and heard it splitting so I stopped.

I wouldn't use a drill! I used a manual driver, and had no issue. I could have used a drill bit, but it wasn't necessary. It was soft plastic. You are just slightly enlarging the hole.