Temp/Humidity/Solar board questions

[dsc3507]

I purchased a hobbypbards T/H/S kit and although it is working I am not clear on calibration.

I sent a message to the owfs dev. group and Paul A. was kind enough to answer but on very broad terms and I am still not seeing the whole picture.

It is my understanding that the owfs dev. group has created data entries for the most common ow chips and has then created specific viewable or settable entries for those chips. These registers are taken from the Maxim or Dallas data sheets.

The Hobby board uses the DS18s20 for temperature readings. I find my reading to be off by several degrees. Under the OWFS in Linux there are registers for TRIM. TRIMBLANKET, and TRIM VALID. o where can I find any documentation on how to use these values. I do know that my display shows TRIM VALID to be yes (TRUE) and if I change the trim numbers it makes changes in the resultant temperature readings. But there does not seem to be any pattern in these changes. Changing TRIM by one (least significant) digit can make a 50 degree change in read temperature. The Maxim/Dallas datasheets do not seem to mention these TRIM vlaues????

So I am at a loss as to what to do here. How is this calibrated or, since the chip is suppose to be calibrated within a few tenths of a degree as is, how do I turn off any trim?

Likewise I find the humidity readings to be off. OWFS has a listing for a couple od sensors but NOt the one that comes with this board. So how does one calibrate it?

I suspect you have to do the math outside of OWFS but sonce the devices are suppose to have a certain accuracy then trying to just add or subtract degrees or %humidity may lead to a non linear result.

What are others seeing doing in this regard?

 

[Michael McSharry]

I do all calibrations outside of owfs for all sensors. The humidity sensor does degrade and most dont need much accuracy so do not service it or replace it every 6 months or so. My experience with temperature calibration is that it is done more to normalize installations than to account for errors in the sensor. For example, a sensor may be mounted at 2' from floor and a thermostat at 4'. There will be a temperature gradient over that distance so a bias is added to the account for this. xapmcs1wire is the generic application that I use locally.

 

[dsc3507]

I do all calibrations outside of owfs for all sensors. The humidity sensor does degrade and most dont need much accuracy so do not service it or replace it every 6 months or so. My experience with temperature calibration is that it is done more to normalize installations than to account for errors in the sensor. For example, a sensor may be mounted at 2' from floor and a thermostat at 4'. There will be a temperature gradient over that distance so a bias is added to the account for this. xapmcs1wire is the generic application that I use locally.

How do you calibrate outside of OWFS? I get about a 4 degree difference between the ds18s20 and 2 different stand alone thermometers placed within inches of each other - same height and no significant ambient air movement.

For the Humidiy, this is a brand new sensor. It is reading 10% lower than other sensors in the area that I believe to be right.

All of these readings are consistent just wrong.

So for the 18s20 in OWFS how do you set the TRIM values in the device directory or do you not have them?

 

[Michael McSharry]

I'm using the DallasSemi software with the MS J library of functions in the .dll that comes with the download. I just take the values provided by the Temperature or Humidity containers and add a bias for their calibration. For some sensors the calibration is different than a bias.

 

[pete_c]

Doug,

I currently utilize Michael's software with numerous 1-Wire sensors. Most recently installed two more combo sensors. Current count is two temp/humidity/solar sensors outside and one combo temp/humidity sensor inside. I have another three "legacy" AAG temp/humidity sensors inside.

The "legacy" AAG combo sensors are running fine after about 5 years or so with consistent readings both for temperature and humidity.

Originally had one combo sensor outside for about 3 years or so.

My newest weather device is a Pagoda in which I installed a combo HB board.

I am still testing on my two outside sensors and new inside combo sensor.

 

[geoff]

...
So I am at a loss as to what to do here. How is this calibrated or, since the chip is suppose to be calibrated within a few tenths of a degree as is, how do I turn off any trim?
.....
What are others seeing doing in this regard?

I was lucky enough to "inherit" a calibrated laboratory thermomenter and have tried to use it to calibrate my assorted 1-Wire temp sensors. At one time I connected 7 or 8 random sensors and stuck them together in a small box to try and keep them all at equal temperature. I found they varied by +/- about 2 degrees. Mostly within 1 degree (Centigrade) of the lab thermometer. I pull the readings from OWFS and adjust them with 2 factors along the lines of:

correct_temp = A * reported_temp + B

The corrections are minor, usually just leave A = 1.00 and B = a small add/subtract correction.

My caution is that sometimes the temp sensor seems to be affected by whatever else is on the same board. Maybe some of the other components are heating up the PCB. I've found the temp sensors in DS2438s used for things like the barometer board are next to useless. Similarly the temp sensor on the solar board is a bit pointless because by definition the solar board is sitting in the sun!