GPIB Interface Bus
From the same era as the RS232 interface is the GPIB interface bus. This bus was originally developed by Hewlett Packard in the late sixties. While most people currently know Hewlett Packard through their printers and computers, the company was many years ago mainly involved in measurement devices and laboratory equipment.
The idea was to not only use this equipment in a stand-alone fashion but to also be able to interconnect these devices and use remote control techniques to operate them. In an era where computers were still mainly mainframes, this was quite a progressive idea. Serial and parallel communication interfaces were available at that time, but they weren’t really useful for multidrop equipment networks as the designers had in mind. Hewlett Packard, therefore, developed a multi-drop interface bus which they named HP-IB for Hewlett Packard Interface Bus.
The specifications were quite amazing for that time. The design of the interface bus was based on an 8-bit bi-directional parallel bus, with a data rate of one megabyte per second. The connector looked like the 36 pin Centronics connector, but smaller with only 24 pins. Cables were often used with double-sided connectors which allowed multiple cables to be used on one device. In that way, devices could be connected in a network in a chain, star, or other topology. Practically up to four cables could be connected to a device at once, but this was mainly because of mechanical issues and not because of electrical limitations.
The maximum cable length per segment was 20 meters, with 15 devices on a cable. Devices are addressed with unique addresses in the range 0..30 on the bus and with extenders, it was possible to use the full 31 devices on one logical network segment.
HP-IB was licensed to other hardware manufacturers under the name GPIB General Purpose Interface Bus and in 1975 even became an IEEE standard with the name IEEE 488.
You might think that a fifty years old interface bus is not used anymore on modern equipment, but the opposite is the case. A GW-Instek GPM-8213 power meter which I bought a few months ago, for example, is still equipped with a GPIB port, as is many other laboratory equipment from other manufacturers.
It is strange to think about how such an old interface bus can survive more than 50 years. Parallel printer cables have been replaced by USB, RS232 on general-purpose computers is almost not used anymore, but on laboratory equipment, GPIB is often still the primary communication port. To understand this, we have to dive a little bit deeper into the use of this type of equipment.
First of all, laboratory equipment has a much longer lifespan than general computer equipment. My 6,5 digit HP 34401A digital multimeter was manufactured somewhere in the previous century but is still working fine. Hewlett Packard developed the model in 1992 and the meter survived the transition to Agilent and later to Keysight Technologies. The sale of new 34401A DMMs was discontinued on December 1st, 2016. That is a commercial life span of 25 years!
Secondly, The connector chosen to connect the equipment was robust and versatile. If you are using measurement equipment in a laboratory setting, you want to connect equipment in different ways depending on the experiments and measurements to perform. The connector allowed for many mating cycles and the stacking of connectors allows for fast adding or deleting equipment to the interface bus.
Last but not least, The IEEE 488 standard not only defined the physical characteristics of the interface bus but also the command structure to be used to communicate from central controllers to the devices on the bus.
I have to admit that the latter was not always the case. At the beginning of the existence of HP-IB and GPIB, manufacturers defined their own proprietary command set to communicate with equipment. Some used ASCII based protocols, other binary-based and this made it sometimes difficult to use devices from multiple manufacturers on one interface bus.
Anytime things appear to be going better, you have overlooked something.