I'm having an issue with invalid temperature readings coming from a vibrating-wire piezometer datalogger. My configuration is a CR800 with an AVW200 and an AM16/32B mux. I have 14 Geokon 4500 piezometers on this logger but programmed all ports (16) on the mux in case I needed to easily add another probe or two. The AVW200 connects over its serial port to the CR800 C1 and C2. There is usually a modem on the serial port of the CR800.
What is happening is that I get invalid temperature readings from the thermistors on ports 8, 9, 11, 12, and 14. To verify that I have a Geokon handheld single-channel vibe-wire reader that will show a reading in the expected 20-24 degC range. The logger is recording various values
between -14 to +10 which I know are incorrect. I have several loggers like this and most of them (but not all) are affected this way.
I know that with an ohmmeter I can check the pins for the thermistor in circuit (but not powered) at the mux, and will get 3.0 to 3.3 Kohms. I do get values in that range for these problematic thermistors. But Loggernet port monitor for VW(8,6) yields a value of 20,506. An ohmmeter on port8 at the connected thermistor pins gives 3.3kohms. If I then move the probe up to a spare port (15 or 16), the problems follow? So it looks obvious to me that I have a cable issue, or that the problem is within the thermistor in the probe - but why does a handheld reader have no problem, and also why do I seem to measure a valid resistance value which disagrees with the logged value?
I have tried replacing the modules (cr800, avw200, am16/32) twice now with the same results.
Any ideas that might help me fix this? My script:
'Declare Variables and Units
Dim Count
Public AVWRC
Public VW(16,6)
Public Freq(16)
Public TR(16)
Public TT(16)
Public Digits(16)
Units Freq=Hz
Units TR=ohms
Units TT=deg C
Units Digits=B_units
'Define Data Tables
DataTable(Table1,True,-1)
DataInterval(0,15,Min,10)
Sample(1,Digits(1),IEEE4)
Sample(1,TT(1),IEEE4)
Sample(1,Digits(2),IEEE4)
Sample(1,TT(2),IEEE4)
Sample(1,Digits(3),IEEE4)
Sample(1,TT(3),IEEE4)
Sample(1,Digits(4),IEEE4)
Sample(1,TT(4),IEEE4)
Sample(1,Digits(5),IEEE4)
Sample(1,TT(5),IEEE4)
Sample(1,Digits(6),IEEE4)
Sample(1,TT(6),IEEE4)
Sample(1,Digits(7),IEEE4)
Sample(1,TT(7),IEEE4)
Sample(1,Digits(8),IEEE4)
Sample(1,TT(8),IEEE4)
Sample(1,Digits(9),IEEE4)
Sample(1,TT(9),IEEE4)
Sample(1,Digits(10),IEEE4)
Sample(1,TT(10),IEEE4)
Sample(1,Digits(11),IEEE4)
Sample(1,TT(11),IEEE4)
Sample(1,Digits(12),IEEE4)
Sample(1,TT(12),IEEE4)
Sample(1,Digits(13),IEEE4)
Sample(1,TT(13),IEEE4)
Sample(1,Digits(14),IEEE4)
Sample(1,TT(14),IEEE4)
Sample(1,Digits(15),IEEE4)
Sample(1,TT(15),IEEE4)
Sample(1,Digits(16),IEEE4)
Sample(1,TT(16),IEEE4)
EndTable
'Main Program
BeginProg
SerialOpen(Com1,38400,4,0,0)
'Main Scan
Scan(15,Min,1,0)
'Geokon 4500 Vibrating Wire Piezometer measurement 'Freq()'
AVW200(AVWRC,Com1,0,200,VW(1,1),1,1,16,450,6000,2,_60Hz,1,0)
For Count=1 To 16
Freq(Count)=VW(Count,1)
TR(Count)=VW(Count,6)
'Calculate thermistor temperature 'TT()'
TT(Count)=1/(1.4051E-3+2.369E-4*LN(TR(Count))+1.019E-7*LN(TR(Count))^3)-273.15
'Calculate digits
Digits(Count)=Freq(Count)^2/1000
Next
CallTable(Table1)
NextScan
EndProg
Possible explanations are:
1) that there is a lot of noise pickup in your wiring that upsets the loggers relatively fast measurement of the resistance of the thermistor compared to the slow integrations used in most handheld meters. If this seems likely check the wiring and screening connections. Also if there is pickup and the local AC circuit has 50Hz frequency you could try the _50Hz instead of the _60Hz integration time in the AVW instruction.
2) there is some leakage to ground in your wiring or the sensor itself, perhaps caused by water ingress. The logger makes a measurement of the resistance which is ground referenced, so it will be upset if there are parallel connections of the thermistor to ground in the wiring/sensor. This would not affect a handheld meter which would be floating relative to ground. You could check this with a handheld resistance meter and with the sensor deselected measure the resistance between one of the thermistor wires to ground. This connection should be open circuit, if not there is your problem.
3) the relays in the AM16/32B are worn out. This is highly unlikely across so many channels and multiplexers especially as the switching frequency for AVW sensors is so low which means the relays should normally last decades in this type of application.
For this case it is critical to double-check the wiring of the sensors into the mux and also the wiring connections between the mux and the AVW200. Also, if the mux isn't switching right, you would get the wrong resistance on the line at measurement time.
I'm curious what the resistance reading on the ohmmeter is on
the "common" side of the mux when the channel is switched in.
However, we should also consider that the AVW200 may be a part of the problem.
You should take a laptop to the field site and connect it to the AVW200 and run DevConfig and use the Troubleshooting tab to examine how the thermistor measurements are doing from the perspective of the AVW200. You will also confirm that the vibrating wire coil circuit (frequency) measurement is working properly (graphically/visually). Those readings may also be suspect.
Please look at the AVW200 manual here:
http://s.campbellsci.com/documents/us/manuals/avw200.pdf
see section 5.5, p33 for details about the Troubleshoot screen.
When you use one of the recent versions of DevConfig, (free download from campbellsci.com) then it shows the thermistor reading with units of Ohms in the Troubleshoot Tab's screen. The Troubleshoot screen has the ability to specify a channel on the Mux and switch into that channel and make a diagnostic reading of the sensor.
You can save the results of the diagnostic reading into a file (Save Last Results button in the Troubleshoot tab). You are probably at the point where you want to contact your local Campbell Scientific support representative and provide them with that information, etc. and work towards a solution.
One other thing that can be done is to bypass the mux completely. Temporarily wire one of the sensors directly into the AVW200 and see what kind of thermistor reading the Troubleshoot Tab shows.
This will help identify whether the mux or the AVW200 is the source of the problem.
Thanks for all the excellent suggestions. I finally had the chance to get out to this remote station last week and look at it. I had assumed it was a grounding issue.
Apparently the problem is somewhere within the instrument(s) itself. I verified that the problem probes in question have minor shorts (~50kOhm)between random wires. I then tested all probes and found that all had no shorts other that the problematic 3 probes. I never had thought to test this way because I carry a small portable probe reader and it always records good numbers.
After seeing the shorts and realizing that what was happening was that the thermistor reading gets corrupted by the coil energizing - I then moved just the thermistor lines over to a free channel on the mux. This seems to work fine. I'm guessing the handheld reader must have some kind of isolation between the coil energize lines and the thermistor lines.
Roy
I didn't work much on this instrument but i have some idea that there may be the noise in your wiring that can interfere with the logger.
Also there are the chances of the shortening of your wires with the ground which also can produce disturbance to the logger.
Also check the AM16/32B relays the problem may be there because of low frequency across multiplexers.
* Last updated by: LaurenMeek on 10/20/2015 @ 11:18 AM *