Stand Alone or computer based
The modern hardware data logger is typically a portable battery-operated device with a large memory, powered by the latest microprocessor technology and capable of acquiring, processing, storing and analysing electrical signals at various speeds from a wide range of sensors - at regular intervals, or in response to an event such as a threshold being crossed or a switch being activated.

The sensors may communicate with the logger through a cable or wireless link and may sense temperature, humidity, pressure, flow, wind speed, current, voltage, resistance and a host of other physical parameters that are important in monitoring and controlling processes or conducting research.

Data stored by stand-alone data loggers is typically downloaded into a computer for more detailed analysis and reporting, though some data loggers have sophisticated on-board processing and analysis capability and can carry out some control functions such as activating an alarm or a switch.

Stand alone loggers due to restricted performance have become a lot more specialised in their operation for example car or truck loggers remote field loggers etc. Because of battery life and storage space stand alone loggers tend to have smaller channel counts. One of the main features of portable loggers is that they use less power and are packaged in a robust way.

Hardware data-loggers are normally used in portable applications where it is necessary to monitor a small number of parameters for a relatively short period of time. For more permanent applications computer based systems are normally used.

Data Logging Systems
Computer based data loggers or data logging systems use a computer, usually a PC, to collect data through sensors, record or log the data in order to analyse and display the results. Data logging systems can also provide additional features such as real time calculation monitoring alarm processing and control. SCADA (Supervisory Control And Data Acquisition) is a further evolution of computer based logging systems, where data is presented in graphical forms so that operators can supervise experiments or processes.
Because of the versatility of computer based loggers they are now used in a wide range of applications and industries. Computer based systems are very scalable and can start from as few as 8 inputs or channels to many thousand.
The basic equipment for computer-based measurement consists of sensors, a scanner or measurement unit a computer and some application software designed for data logging applications. Normally, the sensors are plugged into the signal input-output device which in turn is connected to the computer using a standard port such as RS232, Ethernet or USB. Or plugged directly into the computer bus. Optionally, a printer is also useful for creating hard copy. The software is normally used for managing the data collection, display, storage and analysis and transmission of data. Computer based systems can be sub categorised to be either centralised or distributed

In a centralised system sensors are brought back to a centralised point. In this case I/O plug in boards are sometimes used which connect directly with the standard computer PCI bus. In a distributed system the measurement units are taken to the sensor sources, from there they transmit back to the computer where the data is then processed.

The benefit of distributed systems is that they reduce cabling cost and improve the quality and integrity of the measured data. Distributed systems are normally used in areas where sensors or group of sensors are spread out over large distances, such as a power station or steel processing plant.

Within these categories loggers can be sub categorised by speed. Logging speed is a main characteristic of loggers. Slow moving parameters such as temperature or strain would not normally be scanned at more than once a second whereas transient logging in certain physics experiments may require mega-sample scanning rates. With a large number of inputs or very high scanning rates, logging systems can acquire vast amounts of data which requires storage or processing. For example a high resolution temperature channel logging at one sample a second will require 500K bytes of data a day. High speed systems running at 500M bytes /second would fill a 100 G byte drive in 200 seconds.

To minimise data storage, data loggers are normally provided with various modes of operation. These include periodic logging, logging at pre-determined logging frequencies or event logging, logging if a specific event occurs. Sometimes data loggers can combine a number of these modes simultaneously.
There is sometimes confusion between scanning speeds and logging speeds.

The scanning speed of a data logger is the speed by which data can be acquired and digitised for logging or analysis. The logging speed is the speed at which the data can be recorded. Logging speeds should always be either equal or less than scanning speeds.

In scientific experiments the scanning speed should be at least twice the frequency of the signal which is required to be monitored or logged. This is called the Nyquist frequency. The theory being that in order to obtain information about a signal then it would be impossible to replicate or determine characteristic of this signal without the minimum of two samples. Practically 4 samples or more are preferred.

Speed resolution and accuracy.
There is a physical trade off between resolution and accuracy. The higher the sampling speed the lower the resolution that can be practically obtained. This is due to a number of factors namely the type of analog to digital converter used to digitise the signal the size of the signal and the ambient noise conditions. To ensure a good signal to noise ratio the signals will need appropriate amplification and have signal extraction methods applied to them. Mains bourne noise is present in most electrical environments which can be superimposed on low level signals produced by strain or temperature sensors.For analog signals various techniques can be employed to improve the integrity of signals.
Various types of analog digital conversion techniques can be used that integrate measurements over mains periods and thereby significantly reduce the effects of mains interference.In addition filtering methods can be applied. This technique although effective can prove to be quite expensive particularly if there are a large number of inputs.

The measurement units which provide the interface between the computer and the sensors normally provide the appropriate signal conditioning for the sensors being used. Signals from sensors such as strain gauges and thermocouples are normally of a low level from a few microvolts to 10's of millivolts. In addition these signals can be non linear as is the case with thermocouples. In addition as in the case of strain gauges the sensor requires energisation to perform the measurement. Thermocouples are relative measuring devices and as such require compensation. Measurement units such as the Datascan range provide full signal conditioning linearisation and and energisation for these types of sensors. In addition the device uses an ADC ( analog to digital converter) that integrates over mains cycles and thereby rejects significant amount of mains bourne noise. These features make the connection of analog sensors simple and straight forward.

One of the other features of Datascan is that it can be used remotely via a twisted pair RS485 network. In situations where the signals to be monitored are at a distance from the computer this feature radically reduces cabling cost and significantly improves signal integrity. Using this form of network measurement units can be distributed up to 1Km away from the host computer.

In a lot of experimental or monitoring applications data is required by more than the one operator. Client / Server operation provides the means by which additional users can connect to the host computer via an Ethernet network and view data, graphics or acquire data for further analysis or report writing.

Distributed and remote operation.
Many applications now require data to monitored or logged from remote locations. Previously continuous monitoring of remote sites used to be expensive due to the communication cost. With the advent of the internet it is now possible to connect remote computer based systems to a router and an ADSL line and continuously monitor log and download data from any global location.
Products like the DataWeb 4000 series have built in web servers which provide a real time display of measured parameters over the internet displayed on a standard browser.
These type of systems enable the regular download of data in Excel format. In addition email and sms messaging can be used to send messages in the event of an alarm or malfunction. The measured results from a large number of these units can be consolidated using products like the Control Centre Software which seamlessly consolidates a large number of remote DataWeb modules into one system providing alarm monitoring, trends and mimic displays.

Integrated and embedded systems.
Embedded system loggers are a new generation of data loggers which provide the benefits of hardware loggers with the flexibility and expandability of computer based loggers.

Embedded system loggers are dedicated data acquisition computers with embedded SCADA or data logging software. These loggers provide the means by which physical I/O can be connected either through an integrated bus system or through communication networks to an integrated computer running a standard operating system such as Windows CE or Linux.

The Coral 8000 series is a system of this type. The hardware is equipped with a PXA270 CPU (520MHz) microprocessor running the Windows CE.NET 5.0 operating system. Interfaces are provided for VGA, USB, Ethernet, RS-232/485. The unit provides either and 4 or 8 slots for high performance parallel I/O modules and serial-type I/O modules.

The operating system,Windows CE 5.0, has many advantages, including hard real-time capability, small core size, fast boot speed, interrupt handling at a deeper level, achievable deterministic control and low cost. Using Windows CE.Net 5.0, CORAL 8000 gives has the ability to run PC-based SCADA or Data Logging software. Embedded systems normally use a configuration software to configure the specific application and then this configuration is downloaded onto a runtime unit or target system for operation. These target systems can then be dedicated to specific applications.
The target systems can be connected together using various communication standard such as Ethernet or RS485, in addition they can be interrogated remotely using the internet.These systems provide a low cost means of creating low cost dedicated systems for a wide range of applications.

Measurement systems provide a wide range of products to serve this spectrum of applications. From the basic building blocks of hardware and software to hardware stand alone loggers and computer based or embedded systems.