WXforum.net
Weather Station Hardware => Davis Instruments Weather Stations => Topic started by: stromb0li on June 11, 2018, 10:13:08 AM
-
Hey WXForum,
My new Vantage Pro 2 arrived in the mail late week! :) I've set it up indoors while I'm working on testing a few other items and I've noticed 2F+ differences between what the console and the ISS sensor is reporting (both sitting about 3 feet from eachother in the same room).
Should I let this sit indoors for another week or so and see if the sensors get closer or is it common that an offset is needed for one sensor over the other?
Thank you!
-
Is the console back light on?
-
Which reported value is higher, and do you have a USB logger installed and plugged in?
-
Console backlight is/has been off. Originally I thought it was because the ISS was sitting in the sunlight in the room, but I've put the shades down, today it's cloudy and rainy, and the temp is still 2F off.
The console is what is reporting the higher temp over the ISS.
Yes, I do have the data logger plugged in; I'm using weewx to log the data.
-
Console backlight is/has been off. Originally I thought it was because the ISS was sitting in the sunlight in the room, but I've put the shades down, today it's cloudy and rainy, and the temp is still 2F off.
The console is what is reporting the higher temp over the ISS.
Yes, I do have the data logger plugged in; I'm using weewx to log the data.
There is a fair amount of correspondence here about the heat produced by the USB data logger distorting console temperature readings. see particularly my post http://www.wxforum.net/index.php?topic=18758.msg216491#msg216491 and the surrounding thread for a detailed description of the problem. The bottom line is that the offset you are seeing may reflect this problem.
-
Hmm, that's a pretty interesting/poor design decision if the sensor is right by the data logger/charging port on the console and is from the slight heat emission of the two...
DualTemps2.pdf in this thread is almost exactly what I see, except mine are consistently closer to 2F offset: http://www.wxforum.net/index.php?topic=18758.msg216610#msg216610
I actually am not placing batteries in the console as well; I have a larger UPS the console is plugged into to prevent loss of data.
-
A fairly straightforward way of confirming this would be to disconnect your USB cable for a couple of hours and then plug it back in again and compare the two temperature records after weewx catches up. My guess is that you will see more or less what I did.
Additional note: the experimental setup I used should have provided close to 'best case' conditions for airflow around the Davis console. Most typical installations would likely have more restricted airflow and would likely see a greater heating effect, so your 2 degree offset would certainly not be surprising.
-
The console is enclosed and will never match the outdoor sensor unit. A strong fan blowing on console (back side air vents) and outside sensor may come close but the outdoor uses a more accurate sensor SHT31 vs what the console uses so even then there could be a difference.
-
You can set the offset using the Weatherlink software. I have two consoles, two Envoys and have to now and then reset the temperatures. the variation you are experiencing are within the accuracy level. My screen I am showing is for a mac which is different.
-
The console is enclosed and will never match the outdoor sensor unit. A strong fan blowing on console (back side air vents) and outside sensor may come close but the outdoor uses a more accurate sensor SHT31 vs what the console uses so even then there could be a difference.
In a temperature/humidity-stable indoor environment, the two sensors should come to read the same temperature and humidity (within their accuracy specs) after a reasonable equilibration period. The indoor sensor will certainly have a much slower response to changing conditions than the SHT31 - particularly if this is in a FARS, but their steady-state readings should be comparable.
-
The console is enclosed and will never match the outdoor sensor unit. A strong fan blowing on console (back side air vents) and outside sensor may come close but the outdoor uses a more accurate sensor SHT31 vs what the console uses so even then there could be a difference.
In a temperature/humidity-stable indoor environment, the two sensors should come to read the same temperature and humidity (within their accuracy specs) after a reasonable equilibration period. The indoor sensor will certainly have a much slower response to changing conditions than the SHT31 - particularly if this is in a FARS, but their steady-state readings should be comparable.
n I can see this discussion going for the record. So what should the person do. He /She can send it back and get a replacement under the warranty. Have Davis recalibrate both. Make the adjustments in the software. Or maybe they should dismantle and check voltages...
-
You can set the offset using the Weatherlink software. I have two consoles, two Envoys and have to now and then reset the temperatures. the variation you are experiencing are within the accuracy level. My screen I am showing is for a mac which is different.
Assuming that the readings of the outdoor sensor are correct within its specs (as I would tend to), a 2 degree variance of the indoor sensor, seeing the same conditions, is not within the latter's specs, particularly if it is not consistent, as LaBob found was the case for the offset due to the USB logger heating effect. There is no way of adjusting the offset on the console that will correct a variable error. This is particularly problematic if the computer to which the USB logger is connected is not on 24/7, since you would need to change the offset based on whether the computer is on or off.
-
The console is enclosed and will never match the outdoor sensor unit. A strong fan blowing on console (back side air vents) and outside sensor may come close but the outdoor uses a more accurate sensor SHT31 vs what the console uses so even then there could be a difference.
In a temperature/humidity-stable indoor environment, the two sensors should come to read the same temperature and humidity (within their accuracy specs) after a reasonable equilibration period. The indoor sensor will certainly have a much slower response to changing conditions than the SHT31 - particularly if this is in a FARS, but their steady-state readings should be comparable.
n I can see this discussion going for the record. So what should the person do. He /She can send it back and get a replacement under the warranty. Have Davis recalibrate both. Make the adjustments in the software. Or maybe they should dismantle and check voltages...
It is my belief, based on a fair amount of evidence, that this represents a serious design flaw in the Davis unit. A replacement console will (ex hypothese) show the same behavior, and there is no 'calibration' that Davis or anyone else can do that will fix it. If the console is plugged in to a live computer's USB port 24/7, a fixed offset will partially ameliorate the problem, but see my other latest post. If the OP wants/needs accurate indoor temperatures, I'm afraid that (s)he needs to use a different, separate, sensor to monitor them.
-
It's a design flaw if the associated computer is not on 24/7. And also in situations where the inside temperature is not reasonably constant (because the available offset is constant).
Reportedly, the serial and IP loggers are better in this respect.
Which leads to the question - what's the similar temperature situation with the new WiFiLogger?
-
It's a design flaw if the associated computer is not on 24/7. And also in situations where the inside temperature is not reasonably constant (because the available offset is constant).
It's a design flaw that becomes blindingly obvious if the associated computer is not on 24/7. It remains a design flaw even if you are lucky enough to be using a setup where its results happen to be less serious.
Reportedly, the serial and IP loggers are better in this respect.
The serial logger is, indeed, supposed not to suffer from this problem - since it has to be powered from the console, which may be running on batteries, Davis may have been much more careful to minimize power wastage (as heat). The normal 'use case' for the IP logger would have it continuously powered, so a fixed offset would probably ameliorate most likely problems.
Which leads to the question - what's the similar temperature situation with the new WiFiLogger?
Worth asking, but here again, the normal 'use case' would have it continuously powered.
-
After installing my USB years ago, that's the first thing I noticed was the temp increase. I even called Davis and they confirmed that the USB will cause the temp to run high. I put in a -1.9F in to correct and it matches the home thermostat year round. Personally, I care about the accuracy outdoors, not in, especially when it's easily corrected.
-
It remains a design flaw even if you are lucky enough to be using a setup where its results happen to be less serious.
So, it's a design flaw if it causes heating which affects the temperature sensor. OK.
]
Which leads to the question - what's the similar temperature situation with the new WiFiLogger?
Worth asking, but here again, the normal 'use case' would have it continuously powered.
...and located within the console. Please note that I'm not being critical of the WiFiLogger - which is obviously a very beneficial and technically-excellent product. I'm just asking if it affects the temperature sensor.
-
After installing my USB years ago, that's the first thing I noticed was the temp increase. I even called Davis and they confirmed that the USB will cause the temp to run high. I put in a -1.9F in to correct and it matches the home thermostat year round. Personally, I care about the accuracy outdoors, not in, especially when it's easily corrected.
I assume that you do run your connected computer 24/7. Your report is particularly fascinating in view of the facts that both LaBob and I were (almost certainly subsequently) successively told by various people at Davis that the issue 'had never been reported before' and that I was subsequently told that my contact had checked one of their consoles (method unspecified) and confirmed that there was no such effect! The latter individual did not respond to subsequent messages asking how this determination had been made.
-
I'm running through the test now to see the temps without the USB data logger. Is there a way to get the console to show decimal points for the temperatures?
Thank you.
-
I'm running through the test now to see the temps without the USB data logger. Is there a way to get the console to show decimal points for the temperatures?
Thank you.
you can have in tenths as i remember you hit temp then 2nd key then units and it will show tenths.
-
I'm running through the test now to see the temps without the USB data logger. Is there a way to get the console to show decimal points for the temperatures?
Press and release <Temp>, then repeatedly press <2nd> and <GRAPH> alternately until the temperatures are displayed in the format you want.
-
After installing my USB years ago, that's the first thing I noticed was the temp increase. I even called Davis and they confirmed that the USB will cause the temp to run high. I put in a -1.9F in to correct and it matches the home thermostat year round. Personally, I care about the accuracy outdoors, not in, especially when it's easily corrected.
I assume that you do run your connected computer 24/7.
I do.
As far as Davis's obliviousness to the issue, don't know what to tell ya. I bought the USB ~7 years ago and noticed the issue immediately. There have been thread(s) on this exact thing in this forum, but that was years ago.
-
That worked on getting the tenths of a degree to show up, thank you! :)
And the results are in after letting things sit for about an hour... It's a notable change with the USB Data Logger unplugged (and I'm even using the overly expensive/official Davis cable)! I'm seeing the difference between the two 0.8F now vs 2.1F. In this case, the USB logger is adding a margin of error of 1.1-1.4F!
I do see, the humidity sensor is about 2% off as well, but didn't pay attention to those before/after the data logger was added.
-
That worked on getting the tenths of a degree to show up, thank you! :)
And the results are in after letting things sit for about an hour... It's a notable change with the USB Data Logger unplugged (and I'm even using the overly expensive/official Davis cable)! I'm seeing the difference between the two 0.8F now vs 2.1F. In this case, the USB logger is adding a margin of error of 1.1-1.4F!
I do see, the humidity sensor is about 2% off as well, but didn't pay attention to those before/after the data logger was added.
That sounds about right. Relative humidity readings are dependent on the temperature the sensor 'thinks' it is, so a temperature offset will necessarily affect the accuracy of the humidity reading.
-
I'm seeing the difference between the two 0.8F now vs 2.1F.
And 0.8F is within the accuracy specifications for the devices.
-
<blockquote>Which leads to the question - what's the similar temperature situation with the new WiFiLogger?
</blockquote>
Worth asking, but here again, the normal 'use case' would have it continuously powered.
I have Envoy/USB with VVP running Weatherlink and Cumulus and also a Vue/WiFiLogger (backlight not on) both connected to power outlet and also with batteries installed.
They are located in the identical location and after several minutes of comparing indoor temp the
Envoy/USB had temps of 24.7°C - 24.9°C - 24.9°C - 25.1°C - 25.2°C
WiFiLogger had temps of 27.1°C - 26.9°C - 27.0°C - 27.1°C - 26.6°C
so the difference between the two is from 1.6°C to 2.4°C higher for the Vue/WiFiLogger.
The outdoor temp was the same for all as it should coming from the same ISS.
Enjoy,
Paul
-
I have Envoy/USB with VVP running Weatherlink and Cumulus and also a Vue/WiFiLogger (backlight not on) both connected to power outlet and also with batteries installed.
They are located in the identical location and after several minutes of comparing indoor temp the
Envoy/USB had temps of 24.7°C - 24.9°C - 24.9°C - 25.1°C - 25.2°C
WiFiLogger had temps of 27.1°C - 26.9°C - 27.0°C - 27.1°C - 26.6°C
so the difference between the two is from 1.6°C to 2.4°C higher for the Vue/WiFiLogger.
Very useful information! It would also be nice to know whether the offset changes over time depending on the 'activity' of the logger and/or on the actual ambient temperature.
-
I have Envoy/USB with VVP running Weatherlink and Cumulus and also a Vue/WiFiLogger (backlight not on) both connected to power outlet and also with batteries installed.
They are located in the identical location and after several minutes of comparing indoor temp the
Envoy/USB had temps of 24.7°C - 24.9°C - 24.9°C - 25.1°C - 25.2°C
WiFiLogger had temps of 27.1°C - 26.9°C - 27.0°C - 27.1°C - 26.6°C
so the difference between the two is from 1.6°C to 2.4°C higher for the Vue/WiFiLogger.
Very useful information! It would also be nice to know whether the offset changes over time depending on the 'activity' of the logger and/or on the actual ambient temperature.
By your definition, this is a "serious design flaw" in the WiFiLogger. However, I can't think of an in-console design with those capabilities that would not have this "serious design flaw".
-
Very useful information! It would also be nice to know whether the offset changes over time depending on the 'activity' of the logger and/or on the actual ambient temperature.
I've done a further more precise check and
From 12:50:00 to 13:07:01 the Envoy/USB has been 26.3°C and the WiFiLogger mostly 26.6°C and twice 26.7°C so very close to each other this time.
I presume the Envoy/USB is being polled by VVP every few seconds and I can see the time and temp in Cumulus/Weatherlink.
The WiFiLogger realtime data screen updates every 5 seconds, except at the 5 min the realtime data screen does not update for about 2 minutes. At this 5 min upload interval WiFiLogger begins to FTP the 5 day logger data to my website (about 170,000 file size), and uploading WL, WU, PWS, CWOP, AWEKAS, WOW and one other. During this 2 min delay the indoor temp display does not change nor is there any change when the realtime data screen begins its new 5 sec updates so presume the indoor temp has not changed during that delay period.
Not sure why earlier there was about a 2°C difference and now nearly identical. Nothing has changed other than a couple of hours later.
I have been on this computer for the whole time doing other stuff.
I hope I have been clear enough.Paul
-
I have Envoy/USB with VVP running Weatherlink and Cumulus and also a Vue/WiFiLogger (backlight not on) both connected to power outlet and also with batteries installed.
They are located in the identical location and after several minutes of comparing indoor temp the
Envoy/USB had temps of 24.7°C - 24.9°C - 24.9°C - 25.1°C - 25.2°C
WiFiLogger had temps of 27.1°C - 26.9°C - 27.0°C - 27.1°C - 26.6°C
so the difference between the two is from 1.6°C to 2.4°C higher for the Vue/WiFiLogger.
Very useful information! It would also be nice to know whether the offset changes over time depending on the 'activity' of the logger and/or on the actual ambient temperature.
By your definition, this is a "serious design flaw" in the WiFiLogger. However, I can't think of an in-console design with those capabilities that would not have this "serious design flaw".
I think that it is a significant problem, and certainly one that one needs to know about, and I find it worrying that Paul is now not seeing the previous discrepancy, since this might suggest that the effect is variable, and so could not be fixed by a simple offset. I regard the USB logger's problem as much more serious because that device is much more likely to be used in a way that makes it impossible to ameliorate the issue by any simple action (such as setting a fixed offset) - i.e. in an application such as mine where the attached computer is switched off overnight. As I have implied, devices with a direct IP connection are likely to be used specifically because a 24/7 connection is needed, making potential amelioration of the issue easier.
As I have said elsewhere, IMHO, the right answer is to have (at least as an option) the indoor TH sensor outside the enclosure of the console, or at least better isolated from potentially heat-producing accessories. The design flaw is at least as much in the console layout as in the loggers.
-
Is there documentation on how to use an external indoor temperature sensor with the console? I have a HTU21D-F Temp/Humidity sensor, which seems to be with +-1F of the outdoor sensor that might be a good alternative for the console.