Noisy analog signal in DasyLab from USB-1616HS-4

[Great Lakes Cal]

I have a USB-1616HS-4 that i just got to replace a USB-2416. i switched modules to get the faster DA rate. i have the 1616 set to 10-khz. I have the signal wires of a load cell connected to the CH0 and CH8 terminals. i also tried connecting the ground from the 10-volt excitation voltage supplied by an external source to the S (signal ground) terminal but didnt see any difference. When i have Instacal open and start the analog test i can see a stable voltage from the load cell that changes when load is applied to the load cell. the voltage displayed would be what i would expect and it all looks good. but when i attach a digital meter to the analog input module in DL, i just see a bunch of jumpy numbers that dont look at all like what i see in Instacal. how do i get a nice stable voltage reading from my load cell in DL that matches what Instacal shows me? i have tried averaging modules and digital filters. the averaging modules just make the numbers jump around slower. the filter modules, one high pass and one low pass result in a 60-hx sine wave. what would be the best approach for connecting the load cell wires to the USB-1616 and what would be the best method in DL to get a solid cleans display of the load cell voltage? thank you

[DAQman]

First, let's discuss the actual voltage output by your gauge. Most gauges have a gauge factor (GF) of about two, which means that your quarter bridge output at 1000uE will only be 0.005 volts with an excitation of 10 volts. If you want a resolution of 100uE, the device must be capable of resolving 0.0005 volts. Next, by replacing a USB-2416-4AO with the USB-1616HS-4, you traded off a device that has a worst-case input noise spec of less than 20uVrms (78mV range) for one that is 427uVrms (100mV range).

If you use up to four strain gauge inputs, try the following. First, exit DASYLab and run InstaCal. Double-click the USB-1616HS-4 to bring up the Board Configuration page. Select the Temperature Input Settings tab and change No. Samples/Avg to 64. Click OK and exit InstaCal. In DASYLab, enable five channels on the Analog Input module. Set the Range limits to -0.100 and 0.100 for the first four channels. Change the Analog Type on the fifth channel to temperature. This one temperature channel will engage the InstaCal averaging. This averaging is different from what you can do in DASYLab. Instead, the A/D measures the channel 64 times and returns an average value. This helps reduce channel-to-channel interference. Measuring less than a millivolt will still be difficult, but it should improve.

Because the USB-1616HS-4 has no electrical isolation, I don't recommend connecting the excitation power supply ground to the device. 

Also, the value displayed in InstaCal is a 10-second average. If you wish to have a similar setup, set the input sample rate to 1,000 S/s block size 100 and use an Average module set to 100 values.

[Great Lakes Cal]

the load cell output is +/- 20 mv with a 10-volt excitation. 0-mv at 0-pounds and 20-mv at 10,000-lb. most of the testing they do is between 100 and 5,000-lb. so we would be measuring 0.2-mv to 10-mv. i had to make the switch to the USB-1616 because i needed the faster sample rate. the 1-khz sample rate of the USB-2416 was not enough. i'm starting the get the idea the there are no good options for both low voltage and fast sample rates

[DAQman]

For some odd reason, I thought you had quarter bridge strain gauges. Thanks for the update.

[Great Lakes Cal]

still doesnt seem like the USB1616 is a good choice for low voltage measurements. not sure why it was recommended to me. i guess just because it has the sample ate i needed and a 100-mv range. i think al of the other modules with a higher sample rate dont have less than a 1 volt measuring range. if the USB-1616 has an input noise spec of 427-micro volts, it's going to be very difficult to measure 200-microvolts or get a stable reading at 0-lb which would be 0-volts. are there any other options? not sure what to do with the USB-2416 and the USB-1616 i bought now. these two as well as the USB-2408-4AO were what was recommended i buy based on my 20-mv measuring range and at least 2,500 hertz sample rate.

[DAQman]

The reality is that you trade off speed to get accuracy and stability. To get a quality measurement with the USB-2408-4AO, you must set the (Data Rate) filter to a low value and low sample rate. It also has isolation that helps prevent ground loops. The USB-1616HS has no isolation and was designed to sample at high speeds.

If you require no more than four analog inputs, look at the DT9824. It's a 24-bit device with channel-to-channel isolation and to the host computer and can sample each channel at up to 4800 Hz. The accuracy is 32uV with 6.5uVrms of noise. 

[Great Lakes Cal]

i only need one channel for one load cell input to measure the voltage from the load cell. i need to sample at at least 2,500 hertz.