high Noise on Analog Output PCI 6289

[chiplearner]

Hi guys

I am using  a PCI 6289 M series device for A/D characterization , I am using the device both for the analog output generation and also for data acquistion of the digital lines , The dynamic tests for the A/D were showing a very poor SNR ratio of the ADC, it is a  12 bit adc so technically in the ideal case i should atleast see  90 db but the resulting figure was way lower around 40 db .

I connected the A0 from the PCI 6289 brought out to a scb-68 block to a RHODE and SCHWARZ spectrum analyzer , the above spectrum analyzer is a decent one with good analysis capability and I was surprised to see a SNR of 50 db , the data sheet says that PCI 628X devices have  a internal 16 bit DAC so technically the SNR should be atleast 90 db even if 6 db were removed for the harsh environment of the computer .

please advise , have anybody else faced this problem , the data sheet does not give a figure for the SNR ratio of the card , since i need to test a 12 bit adc , thought that 6289 would be up to the task since it has a 16 bit DAC.I am using the Noise shielding cables and also SCB-68 block , the block is far removed from other power supplies .

any ideas why the noise is occuring , one of my colleagues says maybe the harsh environment inside the pc enclosure causes it , is it necessary to use a separate PCI box for the card and then connect the card to the computer , is this standard procedure ?????.

and also does the PCI -6289 have a filtering option for the analog output , all i could find was the filtering for input.

[Tom_W_[DE]]

Hi chiplearner-

I would expect the ideal SNR for a 12-bit ADC to be about 74dB, not 90dB.  However, the 40dB figure you are seeing definitely seems very low.

The error you're seeing in your output may be related to the operating conditions of the PC and are likely related to the absolute accuracy that is achievable given the temperature conditions within your PC case.  The M Series Help contains helpful tables for the 628X devices to calculate the absolute accuracy you should expect at full range.  You can view the last calibration conditions and the current operating temperature for your device in MAX on the calibration tab under NI-DAQmx Devices>>[your PCI-6289].  You will need to make absolutely sure that the conditions are similar (the help file indicates that a change of +-10C is the operational limit for properly calibrated output values).

The PCI-62XX M Series cards do not feature onboard analog output filtering; the only filtering functionality available is the analog input LP filtering on the PCI-628X devices.

Hopefully this helps-

[chiplearner]

Hi Tom

Thanks a lot for responding to the thread , did not expect a response in the holiday season:) . The SNR for the 12 bit one was a typo , was referring to the 16 bit onboard dac in the M series card , which technically should ensure a clean input signal having atleast 90 db .

The info you have given helps a lot , i suspect the conditions inside the pc might be causing it , also i am driving the analog output from the card at almost its limits of 2.5 MS/s , the sample rate specification for the card on the data sheet says 2.8 MS/s , also i notice the settling time requirement for the outputs is 3 us which i suspect will put the onboard DAC under quite a load.

If the conditions inside the PC cannot be rectified , where can i find a external PCI box which can be connected to the computer or does such a box exist only for the PXI cards ? will that necessarily rectify the problem and also since filtering is not possible on the card itself , I remember reading the scb -68 connector block has  single RC output filter option , i would have to use discrete R,C components to make it but notice on the right hand corner some pcb holes are provided , would this help.

Would appreciate your Response

Thank you

Chip learner

[Tom_W_[DE]]

Hi chiplearner-

I would expect the combination of the PC ambient conditions and the high update rate you're using to both contribute to increased noise on the output.  NI does not offer PCI expansion chassis, but you may be able to find this type of hardware from a third party vendor.  While I'm sure that improving the operating conditions of the card would help, I'm not sure if it would completely remove the problem. 

I would suggest testing in a variety of situations- for example, try to test with the PC case open to allow for better cooling that will hopefully bring the card's temperature closer to its calibrated temperature.  I would also suggest that you generate at a lesser rate to see if the problem is improved.  The 3uS spec you mentioned would be a good starting point to try. 

You might also consider a supplemental cooling solution within the case to cool the components.  Finally, the filtering and breadboard options on your SCB-68 block may be very helpful in filtering the output.  Some suggestions on how to use these options are available in this KB.

Please keep us posted on your progress-