Disturance in temperature reading

[Dmitri Eskine]

Hi everybody,

I have a NI PCI-6023E I/O board with CB-68LP connector block and standard
68-pin cable. The system is used and configured in the differential mode.
And I am using LabView5.0.1 to acquire data simultaneously from three
sensors: LVDT (volts -5 to +5), pressure sensor (volts 0-10), and
thermocouple (K, RT-550 deg.C, thus millivolts). In LabView I am using "AI
single channel" for each of my channels in a while loop. I am doing this
because I need to process each channel before recording (transform
millivolts to temperature using subVI and CJC reading, transform voltage to
distance and pressure using different coefficients) and in addition I need
rapid acquisition (better than 50 points per sec per channel) wi
th time for
each point. The use of standard data-loggers and "acquire waveforms" is very
complicated for me because of problems concerning time recording for each
point and data transformation.
My problem is that temperature reading is inaccurate. I have oscillations of
temperature of about 6 deg. C (at any given temperature) and, starting from
some values coming to voltage channels (e.g. LVDT), there is direct
influence of this voltage on the thermocouple reading (the higher the
voltage from LVDT, the higher the mV from thermocouple).

I will really appreciate any suggestions and examples of VI to solve this
problem.

With many thanks in advance,

Dmitri

--
Dmitri Eskine, PhD
Laboratory of Materials Science
NIMR, Rotterdamseweg 137,
2628 AL Delft, NL
tel: +31 ( 0)15 278 4463
fax: +31 (0)15 278 6730
e-mail: D.Eskine@tnw.tudelft.nl

[Sergey Krasnishov]

Dmitri Eskine wrote:

> Hi everybody,
>
> I have a NI PCI-6023E I/O board with CB-68LP connector block and standard
> 68-pin cable. The system is used and configured in the differential mode.
> And I am using LabView5.0.1 to acquire data simultaneously from three
> sensors: LVDT (volts -5 to +5), pressure sensor (volts 0-10), and
> thermocouple (K, RT-550 deg.C, thus millivolts). In LabView I am using "AI
> single channel" for each of my channels in a while loop. I am doing this
> because I need to process each channel before recording (transform
> millivolts to temperature using subVI and CJC reading, transform voltage to
> distance and pressure using different coefficients) and in addition I need
> rapid acquisition (better than 50 points per sec per channel) with time for
> each point. The use of standard data-loggers and "acquire waveforms" is very
> complicated for me because of problems concerning time recording for each
> point and data transformation.
> My problem is that temperature reading is inaccurate. I have oscillations of
> temperature of about 6 deg. C (at any given temperature) and, starting from
> some values coming to voltage channels (e.g. LVDT), there is direct
> influence of this voltage on the thermocouple reading (the higher the
> voltage from LVDT, the higher the mV from thermocouple).
>
> I will really appreciate any suggestions and examples of VI to solve this
> problem.
>
> With many thanks in advance,
>
> Dmitri
>
> --
> Dmitri Eskine, PhD
> Laboratory of Materials Science
> NIMR, Rotterdamseweg 137,
> 2628 AL Delft, NL
> tel: +31 ( 0)15 278 4463
> fax: +31 (0)15 278 6730
> e-mail: D.Eskine@tnw.tudelft.nl

Dmitri,

If I'm understanding right, it seems to be more electrical/wiring
problem, not LabVIEW related. I'd recommend to check your wiring and
ground referencing of unused channels. Search NI support/developer zone
for "channels crosstalking" and read application note 025.

Best Regards,
--
Sergey Krasnishov
____________________________________
Automated Control Systems
National Instruments Alliance Member
Moscow, Russia
sergey_acs@levsha.ru
http://acs.levsha.ru

[Ben]

I do not have all the specs in front of me but here's my thought.

6023 is 12 bit board. Range is 0 to 10V.

Max value of 12 bits =4096.
Divide 10V by 4096 = 0.0024V = 2.4mV

Compare this with voltages from K type thermocouple.

This may explain what you are seeing.

If so,
you need board with more (smaller)ranges,
or
16-bit board.

Just my thought,

Ben

[PdB]

Just bring the TC signal up to a similar level as the other channels using electronics (and you could add an extra low-pass filter). This is common practice and is called signal conditioning.