Regardless of configuration, obtaining proper results from a PC-based DAQ system depends on a number of system elements. These include the PC itself, the transducers you choose, your signal conditioning approaches, the DAQ hardware, and, of course, the software you run.

Affecting Speed
Back to Top

Let's look at the role of the PC. The computer you use for data acquisition can drastically affect the maximum speeds at which you're able to continuously acquire data. Today's technology boasts Pentium- and PowerPC-class processors, coupled with the higher performance PCI bus architecture and things like USB, as well as the more traditional ISA bus. Also, with the advent of PCMCIA plug-in PC Cards, portable data acquisition is rapidly becoming a more flexible alternative to desktop-based data acquisition.

The data transfer capabilities of the computer you use can significantly affect the performance of your DAQ system. For remote data acquisition applications that use RS-232 or RS-485 serial communication, your data throughput will usually be limited by the serial communication rates.

All PCs are capable of programmed I/O and interrupt transfers. DMA (direct memory access) transfers, not available on some computers, can increase system throughput by using dedicated hardware to transfer data directly into system memory. Using DMA, your processor isn't burdened with moving data and is free to engage in more complex processing tasks.

Boards Must Keep Pace
Back to Top

To reap the benefits of DMA or interrupt transfers, the DAQ board you choose must also be capable of making these types of transfers. The limiting factor for acquiring large amounts of data is often the hard drive.

Disk access time and hard drive fragmentation can significantly reduce the maximum rate at which data can be acquired and streamed to disk. For systems that need to acquire high-frequency signals, you should select a high-speed hard drive for your PC--and make sure that there's enough contiguous (unfragmented) free disk space to hold the data.

Applications requiring realtime processing of high-frequency signals typically need a high-speed 32-bit processor with an accompanying coprocessor, or a dedicated plug-in processor such as a DSP (digital signal processing) board. If your application only acquires and scales a reading once or twice a second, however, a low-end PC can be satisfactory.

You must also look ahead to determine which operating system and computer platform will yield the greatest long term return-on-investment, and still be able to meet your short-term goals. Factors that can influence your choice may include the experience and needs of your developers and end users, other uses for the PC--both now and in the future. There are also cost constraints, and the availability of different computers with respect to your implementation timeframe.

Conventional platforms include the Macintosh operating system (Mac OS), which is known for its simple graphical user interface, and Microsoft Windows 3.x, Windows 9x, and Windows 2000. The latter two iterations boast much improved user interfaces over Windows 3.x, and also offer the option of plug-and-play hardware configuration. In addition, Windows NT 4.0 shapes up as a robust 32-bit OS with the look-and-feel of Windows 9x.

Consider Transducers
Back to Top

Transducers sense physical phenomena and provide electrical signals that the DAQ system can measure. For example, thermocouples, RTDs (resistance temperature detectors), thermistors, and IC sensors all convert temperature into analog signals that an analog-to-digital converter (ADC) can measure.

Other examples include strain gauges, flow transducers, and pressure transducers. These measure force, rate of flow, and pressure, respectively. In each case, the electrical signals produced are proportional to the physical parameters they're monitoring.

Once these signals are generated, they must be conditioned. That's the role of signal conditioning hardware. The electrical signals generated by your transducers must be optimized for the input range of the DAQ board. Signal conditioning accessories amplify low-level signals, and then isolate and filter them for more accurate measurements. In addition, some transducers require voltage or current excitation to generate a voltage output. That's often part of signal conditioning circuitry.

The photo above depicts a typical DAQ system using National Instruments SCXI signal conditioning components.

Amplification Is Familiar
Back to Top

The most common type of signal conditioning is amplification. Low-level thermocouple signals, for example, are often amplified to increase resolution and reduce noise. For the highest possible accuracy, signals should be amplified so that the maximum voltage range of the conditioned signal equals the maximum input range of the ADC.

For example, SCXI has several signal conditioning modules that amplify input signals. The gain is applied to the low-level signals within the SCXI chassis--located close to the transducers--sending only high-level signals to the PC, minimizing the effects of noise on the readings.

Feeling Isolated?
Back to Top

Another common reason for signal conditioning is safety. The goal is to isolate transducer signals from the host computer. One reason is that a system being monitored may contain high-voltage transients that could damage the computer. An additional reason for needing isolation is to make sure that the readings from a plug-in DAQ board aren't affected by differences in ground potentials or common-mode voltages.

When the DAQ board input and the signal being acquired are each referenced to ground, problems can occur if there's a potential difference in the two grounds. This difference can lead to what is known as a ground loop, which may cause inaccurate representation of the acquired signal, or if too large, may damage the measurement system!

Using isolated signal conditioning modules can eliminate ground loops and ensure that signals are accurately acquired. For example, National Instruments' (NI) Model SCXI-1120 and SCXI-1121 modules provide isolation up to 250 V_rms of common-mode voltage. SCXI-1122 modules provide up to 450 V_rms of isolation.

Multiplexing
Back to Top

A common technique for measuring several signals with a single measuring device is multiplexing or MUXing. Analog signal conditioning devices and subsystems often use multiplexing when looking at slowly changing signals such as those from temperature sensors. This is in addition to any built-in multiplexing on a DAQ board.

In a multiplexed system, an ADC samples one channel, switches to the next channel, samples it, switches to the next channel, and so on. Because the same ADC is sampling many channels instead of one, the effective sampling rate of each individual channel is inversely proportional to the number of channels sampled.

NI's SCXI modules for analog signals use multiplexing so that as many as 3,072 signals can be measured with one DAQ board. With the NI Model AMUX-64T analog multiplexer, for example, you can measure up to 256 signals with a single board.

Filtering is also often required. The purpose of a filter is to remove unwanted signals from a signal you're trying to measure. A noise filter is typically used on dc-class signals, such as temperature signals, to attenuate higher frequencies that can reduce the accuracy of your measurement.

For example, many SCXI modules have 4 Hz and 10 kHz lowpass filters to eliminate noise before the DAQ board digitizes signals. AC-class signals, such as vibration sensor signals, often require a different type of filter known as an anti-aliasing filter.

Like the noise filter, the anti-aliasing filter is also a lowpass filter, however has a very steep cutoff rate, so that it almost completely removes all frequencies of the signal higher than the input bandwidth of the board.

If these signals weren't removed, they would erroneously appear as signals within the input bandwidth of the board. Boards designed specifically for ac-class signal measurement (such as NI's Models PCI-4451, PCI-4452, NI_4551, and NI_4552 dynamic signal acquisition boards, and the SCXI-1141 module) have anti-aliasing filters built into them.

Exciting Transducers
Back to Top

We've already mentioned excitation, so let's explore that a bit more. Some signal conditioning systems generate excitation for transducers. Strain gauges, thermistors, and RTDs, for example, require external voltage or current excitation signals. Signal conditioning modules for these transducers usually provide these signals.

RTD measurements are usually made with a current source that converts the RTD's variation in resistance to a measurable voltage. Strain gauges, which are very low-resistance devices, typically are used in a Wheatstone bridge configuration with a voltage excitation source. The NI Model SCXI-1121 and Model SCXI-1122 products, for example, have on-board excitation sources, configurable for current or voltage, which you can use for strain gauges, thermistors, or RTDs.

This brings us to the subject of linearization. It's a common signal conditioning function. Some transducers, such as thermocouples, have a nonlinear response to changes in the phenomena being measured. Both NI's NI-DAQ driver software and the popular LabView, LabWindows/CVI, ComponentWorks, and VirtualBench application software include linearization routines for thermocouples, strain gauges, and RTDs.

Is Conditioning Needed?
Back to Top

It's important to understand the nature of your signal, the configuration that's being used to measure the signal, and the affects of the surrounding environment. Based on this information, you can determine whether signal conditioning will be a necessary part of your DAQ system.

Consider analog inputs. Analog input specifications can give you information on both the capabilities and the accuracy of a DAQ product. Basic specs, available for most DAQ products, tell you things such as the number of channels, the sampling rate, resolution, and input range. The number of analog channel inputs will be specified for both single-ended and differential inputs on boards that have both types of inputs.

Single-ended inputs are all referenced to a common ground point. These inputs are typically used when the input signals are high level (greater than 1 V), the leads from the signal source to the analog input hardware are short (less than 15 ft.), and all input signals share a common ground reference.

If the signals don't meet these criteria, you should use differential inputs. With differential inputs, each input has its own ground reference. Noise errors are reduced because the common-mode noise picked up by the leads is canceled out.

How Fast To Sample
Back to Top

Sampling rate then needs to be examined. This parameter determines how often conversions can take place. A faster sampling rate acquires more points in a given time and can therefore often form a better representation of the original signal.

For example, audio signals converted to electrical signals by a microphone commonly have frequency components up to 20 kHz. To properly digitize this signal for analysis, the Nyquist sampling theorem tells us that we must sample at more than twice the rate of the maximum frequency component we want to detect. So, a board with a sampling rate greater than 40 ksamples/second is needed to properly acquire this signal.

Let's return to the role of multiplexing for a moment. As stated earlier, it's a common technique for measuring several signals with a single ADC. The ADC samples one channel, switches to the next channel, samples it, switches to the next channel, and so on.

Because the same ADC is sampling many channels instead of one, the effective rate of each individual channel is inversely proportional to the number of channels sampled. For example, a sampling at 1 Msample/second on ten channels will effectively sample each individual channel at:

At this point, let's discuss resolution. The number of bits that the ADC uses to represent the analog signal is its resolution. The higher the resolution, the higher the number of divisions the range is broken into, and therefore, the smaller the detectable voltage change.

The figure below shows a sinewave and its corresponding digital image as obtained by an ideal 3-bit ADC. A 3-bit converter (which is actually seldom used, but is a convenient example) divides the analog range into 23, or 8 divisions.

Each division is represented by a binary code between 000 and 111, as shown. Clearly, the digital representation isn't a good representation of the original analog signal because information is lost in the conversion. By increasing the resolution to 16 bits, however, the number of codes from the ADC increases from eight to 65,536, and you can therefore obtain an extremely accurate digital representation of the analog signal if the rest of the analog input circuitry is designed properly.

The term range refers to the minimum and maximum voltage levels that the ADC can quantize. Some multi-function DAQ boards offer selectable ranges so that the board is configurable to handle a variety of different voltage levels. With this flexibility, you can match the signal range to that of the ADC in order to take best advantage of the resolution available to accurately measure your signal.

The range, resolution, and gain available on a DAQ board determine the smallest detectable change in voltage. This change in voltage represents 1 LSB (least significant bit) of the digital value, and is often called the code width.

Ideal Code Width
Back to Top

The ideal code width is found by dividing the voltage range by the gain times two, raised to the order of bits in the resolution. Say a 16-bit multifunction DAQ board has a selectable range of 0 to 10 V, or perhaps -10 V to +10 V and selectable gain of 1 (unity), 2, 5, 10, 20, 50, or 100. With a voltage range of 0 to 10 V, and a gain of 100, the ideal code width is:

therefore, the theoretical resolution of one bit in the digitized value is 1.5 microvolts.

Critical Considerations
Back to Top

Although the basic specs described may show that a DAQ board has a 16-bit resolution ADC and a 100 ksample/second sampling rate, this doesn't mean you can sample at full speed on all 16 channels and get full 16-bit accuracy. You can purchase products on the market today with 16-bit ADCs and get less than 12 bits of useful data.

So, how do you tell if the board that you're considering will give you the desired results? That will be the focus of our discussion in the next installment in this multi-part eChips series.

Read Data Acquisition (DAQ) Fundamentals: Part II of a Multi-Part Series

		Click here to get your listing up.