Why Do My Analog Inputs Seem To Float?

The most common data acquisition problem is unusual or incorrect signal input values. This document briefly outlines the top 10 reasons analog inputs produce unexpected values.

When determining the condition that caused the unexpected values, try to isolate the problem and component involved before you continue. This is especially important in multicomponent systems that involve several E Series data acquisition (DAQ) devices, cables, signal conditioning modules, and terminal blocks. Troubleshooting becomes more complex when you are using multiple components in a system, which is often the case with an SCXI or PXI system.

Note: It is useful when troubleshooting to measure the shorted voltage on a specific channel, as well as a known voltage from a battery or voltage source other than the transducer, to verify that the problem is not within the transducer.

The following list provides troubleshooting information regarding analog input measurement. It begins with simple and obvious issues, and progresses to the more complex issues.

Faulty E Series DAQ Device Connections, or Damaged SCXI Pins and Connections

The E Series DAQ device can be incorrectly connected to the computer bus, or the cable assembly and connector block can be incorrectly connected to the E Series DAQ device. Another source of erroneous analog input readings can come from bent pins and bad connections on the SCXI system. Any given SCXI system has at least five different connections and in most cases more.

SOLUTION
Power off the computer and check the connection of the E Series DAQ device to the computer bus. Also check the connections of the circuit or device under test (DUT) to the E Series DAQ device. In this case the connection may be loose, or the pins may be bent or broken. Typical places to check for bent pins or bad connections on a SCXI system are on the cable connector to the E Series DAQ device and to the SCXI chassis. The chassis usually has an SCXI-1349 adapter that connects the cable assembly to the module in the system. Check the SCXI-1349 adapter for a bad interface to the SCXI module. Remove the SCXI modules and check the pins at the back of the chassis. Also check the connections between the SCXI module and the SCXI/TBX terminal block. Check all the interfaces and connections for continuity and consistency before continuing with the following troubleshooting processes. Doing this first makes the other debugging possibilities less tedious and frustrating.

Unstable E Series DAQ Device State

In certain circumstances the E Series DAQ device can enter a state of instability. This usually occurs after a program crash, which leaves the E Series DAQ device busy with a previous task and unable to respond correctly to the current request.

SOLUTION
Perform a device reset by using the Device Reset.vi programmatically in LabVIEW. Restarting the computer also clears and resets the E Series DAQ device. You can also delete the device from Measurement & Automation Explorer (MAX), and restart the computer to reset and reconfigure the E Series DAQ device.

Multiplexer Settling Time

Another reason you for getting unusual analog input readings is that the settling time required by the multiplexer is not being met. When the multiplexer switches from one signal input to another a capacitive spike appears on the input to the E Series DAQ device amplifier, which looks like a voltage spike. If the input is sampled at this point, you can get erroneous data. This case is also dependent on the settling time of the amplifier. Refer to Amplifier Settling Time.

SOLUTION
Increasing the interchannel delay between samples gives more time for the multiplexer and amplifier to settle, which allows the inputs to the amplifier to stabilize before sampling. In legacy devices, using round-robin sampling provides the maximum amount of time for the sample to settle at a given specific scan rate.

Amplifier Settling Time

One of the most common reasons for floating analog input measurements is that the settling time needed to stabilize the E Series DAQ device amplifier has not been reached. When an input is applied, it takes a period of time to align the input to the amplifier with the output applied by the multiplexer. The amplifier includes a capacitor across the inputs of the amplifier; therefore, the charge built-up on the capacitor from the previous sample must converge or dissipate to align with the new input. Typically, this charge converges or dissipates through the resistance of the source, which creates a resistor-capacitor network (RC network). This RC network has an RC time constant associated with it that results in the input of the amplifier converging at a rate proportional to the product of the source resistance and the amplifier input capacitor. Therefore, the larger the source impedance, the longer it takes to align the inputs. Refer to Figure 1.

Figure 1. RC Network of the Source Resistance and the Amplifier Input Capacitor

SOLUTION
Increasing the interchannel delay between samples gives more time for the multiplexer and amplifier to settle, which. means the inputs to the amplifier can stabilize before sampling. In legacy devices, using round-robin sampling provides the maximum amount of time for the sample to settle at a given specific scan rate.

Damaged Amplifier or Analog-to-Digital Converter

If the amplifier or analog-to-digital converter (ADC) is damaged, the measured analog input values will be incorrect. Often, when large voltages or currents are applied to the amplifier, semiconductor structures within the amplifier are likely to fuse or blow. This causes the amplifier to no longer output expected values. Similarly, it is also possible, yet less likely, to damage the internal structures to the ADC.

SOLUTION
Make sure the E Series DAQ device is properly connected and that the E Series DAQ device is not in an unstable state as mentioned in previous sections. Short several of the analog inputs to ground and measure their signal levels. The measured signals should all read ground (GND). Similarly, apply a known voltage signal to the inputs and measure the signals. You can use MAX to verify signal values. If these tests reveal that the applied signal does not correspond to the measured signal, it is likely that the E Series DAQ device is damaged. To determine whether it is the amplifier or the ADC you can internally route signals to the ADC in MAX. To do this, run the Test Panel on the E Series DAQ device. Input the following strings into the Analog Input Channel Number input box:

_AIGnd	Internally routes analog input ground to the ADC
_Ref5V	Internally routes analog input 5 V reference to the ADC

If a problem results after any of these tests, the amplifier or ADC is damaged and the E Series DAQ device needs repair or replacement. Call NI to initiate a request for repair or replacement of the E Series DAQ device.

Large Source Impedance

A source impedance directly correlates to the time needed for the amplifier to settle (refer to Amplifier Settling Time). Therefore, a large source impedance can result in floating analog input signals due to an increase in settling time of the amplifier. This phenomenon occurs is because as a signal is applied, it takes a period of time to align the output of the multiplexer with the input to the amplifier because the amplifier includes an input capacitor. The charge built up on the capacitor from the previous sample needs time to converge or dissipate to align with the new input. This charge converges or dissipates using the resistance of the source, which creates an RC network. This RC network has an RC time constant associated with it that results in the input of the amplifier converging at a rate proportional to the product of the source resistance and the amplifier input capacitor. Therefore, the larger the source impedance, the longer it takes to align the inputs. Refer to Figure 2.

Figure 2. RC Network of the Source Resistance and the Amplifier Input Capacitor

SOLUTION
Increasing the interchannel delay between samples gives more time for the amplifier to settle because this case is directly related to the amplifier settling time. This means the inputs to the amplifier can stabilize before digitizing. In legacy devices, using round-robin sampling provides the maximum amount of time for the sample to settle given a specific scan rate. However, another solution is to insert a buffer (also called a voltage follower), which is characterized by high input impedance, a low output impedance, and unity gain. As the input voltage changes, the output and inverting input change by an equal amount. This effectively masks the high source impedance with the low output impedance of the buffer. Refer to Figure 3.

Figure 3. Unity Buffer

Damaged SCXI Chassis Fuses

When measuring analog input from SCXI modules, the analog input signals are routed through two analog input fuses. These fuses protect the system and the E Series DAQ device from large signals. One fuse is for the positive channel and the other is for the negative channel. If any of these fuses are blown, it causes the analog input readings to float or run out of range.

SOLUTION
Check the analog input fuses by unscrewing the SCXI system fan nearest the power supply. Both fuses are behind the fan. They look like resistors, not a typical fuse. The fuses are only socketed, not soldered to the board; therefore, you can remove the fuses. Check the continuity of the fuses to determine if they are blown. If either of the fueses are blown, replace them.

Large Value Bias Resistors

Typically in a high-performance data acquisition system it is desirable to include bias resistors on the input channels. These bias resistors are used to reduce noise on the input channels. Their value determines the amount of isolation from ground. This isolation is important in a referenced single ended (RSE) or nonreferenced single ended (NRSE) system. Bias resistors can however increase the settling time of the amplifier. This is because the charge remaining on the input capacitor of the amplifier cannot dissipate through the connection to ground as freely. This fact mirrors the case of the source having a large impedance. An RC circuit is formed between the input capacitor and the bias resistor of the amplifier. The larger the higher the value of the bias resistor, the slower the amplifier input can converge to the actual signal input on the line. This commonly occurs when using a connector block with bias resistors. The TBX-1303 terminal box, for instance, comes with 10 MW and 10 W resistors that you can configure as pull-up and bias resistors. Refer to Figure 4. Although 10 MW is reccommened as the resistor level of bias resistors for RSE, this resistor level increases the necessary amplifier settling time. Therefore, the system sees problems when scanning multiple channels with large voltage swings between channels or when sampling at a very fast rate. Refer to Figure 5.

Figure 4. Resistor Configuration of TBX-1303 SCXI Terminal Block

Figure 5. RC Circuit Path

SOLUTION
Several solutions exist for this situation. You can reduce the sampling rate or increase the interchannel delay, which allows more time for the amplifier to settle. Similarly, reducing the value of the bias resistor decreases the amplifier settling time. However, reducing the resistor value too much can cause ground loops in an RSE system. Doing this causes no problem in an NRSE.

Grounded Signal Sources

Often, when the circuit or device under test (DUT) is ground-referenced it is possible to get unwanted common-mode voltage or ground-loop currents when using a ground-referenced measurement system. The Figure 6 shows the problems with this setup. In this case the measured voltage, V_m, is the sum of the signal voltage, V_s, and the potential difference, V_g, which exists between the signal source ground and the measurement system ground. This potential difference is generally not a DC level. The result is a noisy measurement system often showing power-line frequency (60 Hz) components in the readings. Ground-loop introduced noise can have both AC and DC components that introduce offset errors and noise in the measurements. The potential difference between the two grounds causes a current to flow in the interconnection. This current is called ground-loop current. Refer to Figure 6.

Figure 6. Incorrect System Configuration of a Grounded Signal Source

SOLUTION
You can still use a ground-referenced system if the signal voltage levels are high and the interconnection wiring between the source and the measurement device has low impedance. In this case, the signal voltage measurement is degraded by the ground loop, but the degradation can be tolerable. You must observe the polarity of a grounded signal source before connecting it to a ground-referenced measurement system because the signal source can be shorted to ground, which can damage the signal source. Nonreferenced measurements are provided by both the differential (DIFF) and the NRSE input configurations on a typical E Series DAQ device. With either configuration, differences between references of the source and the measuring device appear as common-mode voltage to the measurement system, and are subtracted from the measured signal. Refer to Figure 7.

Figure 7. Correct System Configuration of a Grounded Signal Source

Incorrectly Configured Gain in Measurement & Automation Explorer

When you are dealing with SCXI modules with jumper-configurable gain, the gain setting in MAX often causes analog input signals to be out of range or smaller than expected. The can happen when MAX configures and reports gain. On modules with software configurable gains, setting the gain in MAX configure the gain on the SCXI module. However, on SCXI modules without a software configurable gain such as the jumper configurable gain SCXI modules, the gain setting of the module in MAX is strictly for reporting purposes. It allows MAX and the Test Panel to accurately represent a signal. Therefore, a 5 mV signal that is amplified 1000 times by an SCXI module only shows up as a 5 mV signal in MAX if MAX correctly reports the gain of the module as 1000 times.

Example: If you are measuring a ±2.5 mV signal with an SCXI-1121 jumper-configurable gain module, you correctly set the jumper gain as 2000 on the SCXI module to maximize bit precision on the E Series DAQ device. In MAX you set the signal range to ±2.5 mV. Assuming you do not change the gain settings in MAX, the software still has the gain set to 1, which is the default for most SCXI modules. Now MAX assumes it is getting a mV level signal and automatically adjusts the gain of the E Series DAQ device to maximize the bit resolution and applies a second gain. If this is the case, the signal is in the ±5 V range at the input to the E Series DAQ device and well beyond the range after the second stage of amplification by the E Series DAQ device. Any gain that MAX automatically configured for the E Series DAQ device pushes the signal out of range.

SOLUTION
Always make sure that the gain settings in MAX for the SCXI modules are correctly reported. Double check the gain settings or jumpers on the module and cross-reference them with the gain settings in the properties of the SCXI module in MAX.
Related Links:
Field Wiring and Noise Considerations for Analog Signals - Tutorial - Development Library ...
Why Do I See a Voltage Spike on My Analog Inputs When I Use the AI_READ_SCAN DAQ Function Call But Not When I Use the AI Single Scan VI in LabVIEW?
What Is the Maximum Achievable Scan Rate of the Simultaneous Sampling SCXI Modules?
What Is the Difference Between Interval Scanning and Round Robin Scanning?
How Can I Check the Interchannel Delay of My DAQ Task Using LabVIEW?
How Do I Create a Buffer to Decrease the Source Impedance of My Analog Input Signal?

Reader Comments | Submit a comment »