Multifunction DAQ

cancel
Showing results for 
Search instead for 
Did you mean: 

PCI-6120 noise performance is higher than in the spec sheet.

Solved!
Go to solution

I am using the PCI-6120 board in warp mode, at 1 MHz. Everything is working fine but there seems to be a significant discrepancy between the measured noise performance (with 50 ohm or short) and quoted performance in the 6115/6120 spec sheet. Below is my comparison.

 

AI Range          Quoted N+Q (uV)          Measured N (uVrms)

+/-42  V          550                               1410

+/-20  V          220                               600

+/-10  V          110                               280

+/-5    V          69                                 140

+/-2    V          31                                 60

+/-1    V          18                                 30

+/-0.5 V          10                                 20

+/-0.2 V          5.1                                15

 

It seems to me that the N+Q data is shifted by one analog input range as the numbers agree well just not on the right ranges. For example the 42 V range quoted N+Q agrees with the 20 V range measured noise. Has anyone else witnessed this?

0 Kudos
Message 1 of 10
(3,811 Views)

The specifications for the PCI-6115 and PCI-6120 are not applicable when using Warp Mode.  Warp Mode causes the board to sample faster than the specified ADC settling time, as a result your accuracy will suffer.  The specifications give are valid in non-warp mode sampling up to 800 kS/s maximum.  I apologize for any inconvenience this may cause.

 

Regards,

Seth B.
Principal Test Engineer | National Instruments
Certified LabVIEW Architect
Certified TestStand Architect
0 Kudos
Message 2 of 10
(3,809 Views)

Seth, thanks for the reply but it is not a warp mode issue. I checked the device under normal sample rates using the onboard sample clock from 10 kHz - 800 kHz and the noise levels change less than 10 %. I also tested the device using the onboard counter as the timing source just like in warp mode but at sample rates from 10 kHz - 800 kHz and got similar results. The latter was the lowest of the two configurations. ~ Mike

0 Kudos
Message 3 of 10
(3,805 Views)
I accidently checked solution solved.
0 Kudos
Message 4 of 10
(3,797 Views)
You can uncheck it by using the options button on his reply.
Now Using LabVIEW 2019SP1 and TestStand 2019
0 Kudos
Message 5 of 10
(3,794 Views)

Hello Mike,

 

Sorry for the delayed response.  Your measured N values, are they averaged or single point values?  Also, has the card been self-calibrated recently?

 

Thanks,

Seth B.
Principal Test Engineer | National Instruments
Certified LabVIEW Architect
Certified TestStand Architect
0 Kudos
Message 6 of 10
(3,734 Views)
Seth... The values are averaged and no the board has not been self-calibrated recently. It's actually a system that has only been running for a month so it's hard to believe the calibration has changed. I can try calibrating it though. ~ Mike
0 Kudos
Message 7 of 10
(3,724 Views)

I tired the self-calibration. The noise levels did not change. Here is some additional data I took at a 500 kHz sample rate.

 

Range        N(single point)        N(averaged)

42             11.3 mV                 1365 uVrms

20             4.82 mV                 589 uVrms

10             2.22 mV                 272 uVrms

5               1.16 mV                 142 uVrms

2               0.44 mV                 55 uVrms

1               0.25 mV                 30 uVrms

0.5            0.16 mV                 20 uVrms

0.2            0.13 mV                 15 uVrms

0 Kudos
Message 8 of 10
(3,712 Views)
Solution
Accepted by topic author fastball09

So what you've measured here is actually system noise, which we spec on page 4 of the specifications, not noise uncertainty, which we spec on page 3.  The system noise for the PCI-6120 is shown below.  To calculate system noise from these specs, you using the following formula:

 

(System Noise in LSBrms) * (Input Range)/(2^Resolution) = System Noise in Vrms

 

So, for the 42V range, the specified noise in LSBrms is 1.2 LSBrms.  The input range is 84 (±42V is 84 volt total).  The resolution is 16 bits.  Thus:

 

1.2*((84)/(2^16)) = 0.001538 V = 1,538 uV

 

For each of the ranges, the system noise is:

 

Range LSBs System Noise (µV)
42 1.2 1538.09
20 1.2 732.42
10 1.2 366.21
5 1.5 228.88
2 1.7 103.76
1 2 61.04
0.5 2.2 33.57
0.2 2.8 17.09

 

Hope this addresses your question.

 

Seth B.
Principal Test Engineer | National Instruments
Certified LabVIEW Architect
Certified TestStand Architect
0 Kudos
Message 9 of 10
(3,687 Views)
Thanks Seth, that explains it. I missed the caption on the tables. It is performing as expected then.
0 Kudos
Message 10 of 10
(3,685 Views)