We own a CR3000 Micrologger and we have the following problem.
We connect an instrument (e.g. a thermometer (0-5V)) in a analog input of the datalogger. When we use the single-ended channel (instruction VoltSE) it is possible to get good data. When we use the differential channel (instruction VoltDiff) we don't get good data. The datalogger records a positive number or NAN for battery voltage and regarding the thermometer a value that fluctuates from a negative number to a positive number, very close to the actual value, is recorded. With a multimeter we check the voltage differences in the analog channel we used and we find a non-zero voltage (e.g. 2.5 Volt for thermometer), very close to the actual voltage.
I would like to underline that we cheched all the differential channels. All the diff channels produce the same problem. Futhermore we cheched that our ground cable is working.
The only way to get good data is to add a wire link between the L and the adjacent ground.
Maybe we have never collected data from the diff channels since the purchase of the CR3000, but we have previous experience in handling a datalogger with a CR10X datalogger.
What do you think? There is a problem in the harware or the software? Upgrading the operating system would help? The existing OS in our datalogger is "11".
Thank you.
Christiana
This will be a common mode voltage issue. With the CR3000 both the high and low inputs must be within the common mode range of the datalogger which is +/-5V of the datalogger ground. (See section 4.2.1 of the latest manual)
IF your sensor is not connected to a supply that has a common ground reference with the logger the problem will be that the signal is outside the common mode range. This will be by nature of the signal referencing, or by leakage to a power supply or even build up of static electricity which makes the inputs drifting outside of the common mode range. This problem is partly due to the high input impedance of the dataloggers (a good factor in most circumstances).
Tying the L input to ground or making a single ended measurement will cure these issues, although you need to be careful to check this does not cause an issue for the sensor if there is a common power supply, as a few sensors may not like their low outputs being tied to power supply ground.
Following you will find the program we send via LoggerNet for a thermometer, as well as a short part of the data we get.
Public BattV
Public Temp_C
Units BattV=Volts
Units Temp_C=deg C
DataTable(Table1,True,-1)
DataInterval(0,1,Sec,10)
Sample(1,BattV,FP2)
Sample(1,Temp_C,IEEE4)
EndTable
BeginProg
Scan(1,Sec,1,0)
Battery(BattV)
VoltDiff (Temp_C,1,mV5000,1,True,0,250,0.02,-30)
CallTable(Table1)
NextScan
EndProg
"TOA5","CR3000","CR3000","4773","CR3000.Std.11","CPU:Thermometer Thies.CR3","11289","Table1"
"TIMESTAMP","RECORD","BattV","Temp_C"
"TS","RN","Volts","deg C"
"","","Smp","Smp"
"2012-11-20 16:57:18",20,13.3,25.40073
"2012-11-20 16:57:19",21,13.31,25.39414
"2012-11-20 16:57:20",22,13.31,25.39909
"2012-11-20 16:57:21",23,13.3,25.39987
"2012-11-20 16:57:22",24,13.36,25.40152
"2012-11-20 16:57:23",25,13.52,25.40811
"2012-11-20 16:57:24",26,13.77,25.45095
"2012-11-20 16:57:25",27,14.01,25.41305
"2012-11-20 16:57:26",28,14.35,25.37186
"2012-11-20 16:57:27",29,15.4,25.26476
"2012-11-20 16:57:28",30,17.03,25.17578
"2012-11-20 16:57:29",31,19.25,24.90146
"2012-11-20 16:57:30",32,"NAN",23.98036
"2012-11-20 16:57:31",33,"NAN",21.94704
"2012-11-20 16:57:32",34,"NAN",18.79983
"2012-11-20 16:57:33",35,"NAN",14.78755
"2012-11-20 16:57:34",36,"NAN",9.08633
"2012-11-20 16:57:35",37,"NAN",2.200375
"2012-11-20 16:57:36",38,"NAN",-3.879831
"2012-11-20 16:57:37",39,"NAN",-7.738317
"2012-11-20 16:57:38",40,"NAN",-8.613295
"2012-11-20 16:57:39",41,"NAN",-9.941416
"2012-11-20 16:57:40",42,"NAN",-11.93854
"2012-11-20 16:57:41",43,13.29,-13.61599
"2012-11-20 16:57:42",44,13.28,-5.04747
"2012-11-20 16:57:43",45,13.28,4.35154
"2012-11-20 16:57:44",46,13.29,21.35017
"2012-11-20 16:57:45",47,13.27,24.61014
"2012-11-20 16:57:46",48,13.26,25.34177
"2012-11-20 16:57:47",49,13.27,25.34835
"2012-11-20 16:57:48",50,13.27,25.34341
"2012-11-20 16:57:49",51,13.27,25.34341
"2012-11-20 16:57:50",52,13.26,25.34671
"2012-11-20 16:57:51",53,13.27,25.34506
"2012-11-20 16:57:52",54,13.27,25.34671
"2012-11-20 16:57:53",55,13.27,25.3583
"2012-11-20 16:57:54",56,13.27,25.35994
"2012-11-20 16:57:55",57,13.27,25.36488
"2012-11-20 16:57:56",58,13.27,25.32698
"2012-11-20 16:57:57",59,13.26,25.37312
"2012-11-20 16:57:58",60,13.27,25.38631
"2012-11-20 16:57:59",61,13.27,25.40773
"2012-11-20 16:58:00",62,13.26,20.74936
"2012-11-20 16:58:01",63,13.27,11.66352
"2012-11-20 16:58:02",64,13.27,9.68119
"2012-11-20 16:58:03",65,13.28,7.361057
"2012-11-20 16:58:04",66,13.27,5.251846
"2012-11-20 16:58:05",67,13.27,2.521404
"2012-11-20 16:58:06",68,13.27,0.3462811
"2012-11-20 16:58:07",69,13.27,-1.220797
"2012-11-20 16:58:08",70,13.27,-2.543999
* Last updated by: Christian on 11/21/2012 @ 5:53 AM *
The erroneous battery voltage is almost certainly caused by the out of common-mode issue.
If you force any of the analogue inputs outside of +/-8V relative to ground, the accuracy of other measurements including the battery voltage and internal temperature will start to go bad after a while because of current leakage that happens between channels (including internal channels used for battery voltage etc).
One way to check the problem is to connect your sensor in the differential configuration and use a high impedance voltmeter to measure the voltage of the logger H and L inputs individually relative to the logger ground. If the voltages are more than 5 V relative to ground you will start to get bad measurements. If they are more than 8V relative to ground you will start to cause problems for other inputs.
I checked the following:
1) connection of a thermometer Thies Clima with CR3000 (0-5V output signal, supply voltage 24V - not common ground reference with logger)
-- single-ended channel (VoltSE) --
battery voltage: 13.31V
voltage between H - ground: 2.68V (23.60 degrees C)
-- differential channel (Voltdiff) --
voltage between H - ground: 1.319
voltage between L - ground: -1.188
voltage between H - L: 2.534
the datalogger does not record good values for temperature (I have the problem I mentioned in the first message)
2) connection of a Setra barometers CS100 with CR3000 (0-2.5V output signal, supply voltage 12V directly from datalogger and 5V in order to stay all the time open)
-- single-ended channel (VoltSE) --
battery voltage: 13.30V
voltage between H - ground: 2.076V (1015 mbar)
-- differential channel (Voltdiff) --
voltage between H - ground: 2.078
voltage between L - ground: 0.005
voltage between H - L: 2.073
Good data with pressure about 1014 mbar.
3) connection of a Young 05103 anemometer with CR3000 (0-5V output signal, supply voltage 12V directly from datalogger)
-- single-ended channel (VoltSE) --
Wind Speed with 0 m/s
voltage between H - ground: 0.03V
Wind direction steady at 270 degress
voltage between H - ground: 3.74V
-- differential channel (Voltdiff) --
Wind Speed with 0 m/s
voltage between H - ground: 11.52
voltage between L - ground: 10.45
voltage between H - L: 1.060
Wind direction steady at 270 degress
voltage between H - ground: 11.49
voltage between L - ground: 10.58
voltage between H - L: 1.013
Datalogger records zero values for both wind speed and wind direction.
I would expect not to have problem when get power supply for the instruments directly from datalogger. Even in this case I see high and low inputs exceed the values of common mode range +/-5V.
So, which could be the cause of our problem?
Thank you.
Christiana
Some of the problems will be the wind sensor when running in differential mode. You must have the 05103V version of the sensor if it has 0-5V outputs (we normally support the standard 05103 sensor which is lower power).
That said you should not have problems with the V version of the sensor if you have the ground wires connected correctly. Make sure you have the earth wire and power ref connected to the logger power ground and in differential mode you are referencing both L channels to the REF output from the sensor.
The voltages you are measuring in differential mode would imply there is not a proper ground return for the sensor. (The sensor itself, unless faulty, cannot output more than +8V relative to its own ground as there is an 8V protective clamp in its own circuit.)
Please try to resolve that problem first. IT may be affecting the Thies measurement.
Having said that the Thies sensor should work OK if those voltages relative to the logger are stable. They may drift outside the common mode range when you disconnect the meter though as the input resistance of the meter may be low enough to pull the grounds together. The most reliable solution if it affects nothing else would be to tie the L input to the logger ground.