self calibration of NI PCI-6251 (M series) using the DDK

[rolfheinrich]

Hi,

I am new in this froum, and I started with the DDK two weeks ago, and so far I got most of the things that I need to work with the NI PCI-6251 (M series) DAQ board plugged into an Intel Dual Atom board D945GCLF2 running (Pure)Darwin 9.6 (build 9G55 -- this is not Mac OS X).

BTW: NI-DAQmx for Mac OS X utilizes a hell a lot of Apple's closed source frameworks and therefore it does not work on a strictly Open Source Darwin installation. That's the reason why I have to live with the functionality provided by the DDK.

I read the articles on calibration and I searched the NI discussion forum, and my understanding so far is, that self-calibration is possible using the DDK, but it is not that easy as with MAX.

First question:

   Is it really basically possible to get self-calibration of a PCI-6251 somehow to work using the DDK?

If YES

I read the article http://zone.ni.com/devzone/cda/tut/p/id/3688

 about the sophisticated polynomial calibration scheme for the M series boards named NI-MCal, and the second question is:

   Is the NI-MCal routine implemented at the DAQ board and that does already most of the magic,

   or do I need to implement the whole NI-MCal sequence step-by-step by software by myself?

If YES (magic routine at the board)

   How can I utilize it using the DDK?

If NO (no magic routine at the DAQ board)

From said article I understood that the NI-MCal sequence works follows:

1. Linearity Correction

    a) measure the INL of the ADC by sweeping the on-board

        PWM based DAC through the entire range

    b) smoothing and curve fitting to a third order polynomial

    c) interpolation

2. Offset and Gain Calibration (to be repeated for each ADC range)

    a) measure the offset based on the corrected INL against

        the on-board ground reference.

    b) for the 10 V range use the on-board voltage reference as is,

        for the other ranges, use the onboard PWM DAC is used in conjunction

        with the calibrated 10 V range to generate precise voltage references

    c) measure the slope based on the corrected INL against

        the on-board voltage reference.

3. Store the calibration coefficients for later use

Final questions:

for 1.a) Is there somewhere an example on how to measure the INL using the PWM-DAC?

for 2.a) Is there somewhere an example on how to direct the ground reference into AI0?

for 2.b) Is there somewhere an example on how to combine the PWM-DAC with the

         calibrated 10 V range to generate precise voltage references for the other ranges?

for 2.c) Is there somewhere an example on how to direct the on-board voltage reference into AI0?

I won't have difficulties to get 1.b-c and 3. straight.

Many thanks in advance for any helpful response.

Best regards

Dr. Rolf Jansen

Message Edited by rolfheinrich on 03-12-2009 05:15 PM

Message Edited by rolfheinrich on 03-12-2009 05:16 PM

[rolfheinrich · Accepted solution]

I found out everything by myself.

- Yes, self-calibration of NI PCI-6251 is possible with the DDK,

   and I have a working implementation for Darwin on said

   Intel Dual Core Atom board D945GCLF2

- No, there is no magic routine at the board. Everything has

  to be done step by step by software.

1a) use AI_Config_FIFO_Bypass to select the calibration source 5

      board->AI_Config_FIFO_Bypass.setAI_Bypass_Config_FIFO(1);

      board->AI_Config_FIFO_Bypass.setAI_Bypass_Gain(1);

      board->AI_Config_FIFO_Bypass.setAI_Bypass_Cal_Sel_Pos(5);

      board->AI_Config_FIFO_Bypass.setAI_Bypass_Cal_Sel_Neg(0);

      and pass tMSeries::tAI_Config_FIFO_Data::kAI_Config_Channel_TypeCalibration to aiConfigureChannel()

      and set the PWM High/Low_Time (below code sets the PWM-CalDAC to +7.5 V)

      board->Cal_PWM.setLow_Time(896);

      board->Cal_PWM.setHigh_Time(128);

 2a)  use AI_Config_FIFO_Bypass to select the calibration source 1 (see code above)

 2b)  for calibrating the +/- 5 V range, set(Low/High)_Time to 128 / 64 (ca. 3.3 V),

        and measure the exact value with the calibrated 10 V range,

        then use this exact value as reference for the 5V range

 2c)  use AI_Config_FIFO_Bypass to select the calibration source 2 (see code above)

BTW: A calibration sweep of the PWM-CalDACs from - 10 to +10 V showed that the characteristics of the ADC at my NI PCI-6251 is almost linear, this was already suggested by the 2nd and 3rd order calibration constants (3.97e-11, 5.89e-15), that I found in the EEPROM of my board. As a matter of fact, the simulated deviation from the ideal straight line is only ±0.6 LSB at a reading of ±FS. Furthermore, a simulation with the calibration constants that I found in the EEPROM do not fit well to the results of said calibration sweep (see the 2 curves that are attached to this message).

Best regards

Dr. Rolf Jansen

[rolfheinrich]

I want to make my self calibration routine for the NI PCI-6251 more sophisticated and I made some more experiments with the settings of the AI_Config_FIFO_Bypass register. From this I already can unambiguously relate some values for AI_Bypass_Cal_Sel_Neg and AI_Bypass_Cal_Sel_Pos to physical signal sources.

Here comes, what I found out so far for the assignments for AI_Bypass_Cal_Sel_Pos/Neg to the signal sources

AI_Bypass_Cal_Sel_Pos

0 ground ? AI or AO

1 ground ? AI or AO

2 internal reference ca. 7 V

3 calibration source low ca. 0.5 V

4 calibration source mid ca. 2 V

5 calibration source high ca. 10 V

6 ground ? AI or AO

7 AOx where x depends on the value of AO_Bypass_AO_Cal_Sel

The voltages of the signals 3, 4, and 5 can be modified by the relation of High_Time to Low_Time in the Cal_PWM register.

AI_Bypass_Cal_Sel_Neg

0 ground ? AI or AO

1 ground ? AI or AO

2 temperature sensor ? ca. 1 V

3 AI0 ?

4 AI8 ?

5 ground ? AI or AO

6 ground ? AI or AO

7 calibration source high ca. 10 V

As you can see from the question marks, I have still some doubts, and perhaps somebody can enlighten me.

Many thanks in advance for any helpful response.

Best regards

Dr. Rolf Jansen

[JoeFriedchicken]

Hi Rolf,

After a lot of digging and a lucky stumble, I found some more information for you.

AI_Bypass_Cal_Sel_Pos

0    (SCXI)
1    (SCXI)
2    correct
3    correct
4    correct
5    correct
6    AI Ground
7    correct

AI_Bypass_Cal_Sel_Neg

0    (SCXI)
1    (SCXI)
2    correct
3    AI Sense
4    AI Sense2
5    AO Ground
6    AI Ground
7    correct

[rolfheinrich]

Hi Joe,

many thanks for the helpful information. With that I corrected already the calibration module of my program, because I assumed wrong selectors for AI Ground and AO Ground. Although the calibration did not show any significant change, I feel better now by using confirmed selectors.

By the way, I found out the relationship between the voltage reading of the internal Tsensor and the actual temperature in °C. For this I used MAX at a PC and during the warm-up phase of the DAQ-board, I noted down the voltage reading of the Tsensor from the test panel and correlated this to the displayed temperature.

A linear regression resulted almost exactly in:

     T(°C) = 100*abs(TsensorVoltage) - 50

Best regards

Rolf

[sk_a]

i am new to the labVIEW can any one help me how to generate a sequence using PCI6251. for example sequence should be the line should be 1 for 10us, 0 for 20us and 1 for 5us then it should go to zero.