6008 accuracy

[TimRsandiego]

The voltage range of my inputs are only 0 - 5V.  I was thinking that the 6008 ADC's accuracy would then be 5V / 2^12 = 1.22V for differential, but I guess the numerator of this equation is always 20V, regardless of what input range you specify in MAX?

[SnowMule]

The RSE range is fixed at 10V at 11 bits.  There's 8 ranges for differential, ranging from 1 to 20 volts, at 12 bit resolution. 

It is a ±V range, so there's really a 10-volt span for those 12 bits at the ±5V range.

[johnsold]

Are you asking about accuracy or resolution?

The calculation you showed is the resolution for a 5 V range.  The 6008 specifications show the only available single-ended range as +/-10 V or a 20 V span, giving a resolution of 4.88 mV.  In differential mode you can have a +/-5 V range or 10 V span for 2.44 mV resolution.

The absolute accuracy specification for the single-ended range is 14.7 mV at 25 degrees C.  Accuracy is not the same as resolution.

Lynn

[TimRsandiego]

Okay, I was referring to resolution, but now I am investigating accuracy.  I'm now in differential mode, and I have a differential voltage on an AI input of 500 mV.  I have a DMM hooked up here to monitor this input voltage, which is very stable (from a power supply).  I also have MAX running, and have set the input range to 0 - 1V on that channel.  In MAX, I'm running continuous samples and am seeing this input swing from 325 mV to 700 mV, even though I have a very stable input coming directly from an adjustable power supply!  Why is MAX seeing such a huge swing???  I need to be able to read voltages at this level (~0.5V) with good accuracy, so wild swings like this are unacceptable.  Help please?  Thx.

[johnsold]

I suspect that you have a grounding problem.  The variations you see may be power line frequency signals which are getting into the system.  A floating DMM may not see tehm or may average over a longe enough time to filter out most of the variation.  Try setting the samplng rate to 500 or 1000 Hz and collect about 100 ms of data.  See if you observe periodic variation with 5 or 6 cycles, depending on the power line frequency.

Lynn

[TimRsandiego]

I had noise on my input that the DMM wasn't seeing.  A 0.1 pF and 1.0 pF cap in parallel fixed it.  Now I'm seeing +/- 6 mV swings with the input signal range set to 0 - 1V.  Thanks.