It has long been the case that test and measurement equipment is designed to withstand severe input overloads. With the recent changes in relevant safety standards, these design requirements are more important than ever—particularly since the interpretation of the standards seems open to question.
Applications
Usually, you are protecting measurement terminals against the AC power supply. In the United States, this is 110 V; in Europe and other places, it is 230 V. Since 230 V is a more stringent requirement, multinational corporations design for the worst-case input overload. Even thermocouple inputs, designed to take only a few millivolts of genuine input, often can withstand a 240-V + 10% overload. This is because users invariably connect the thermocouple inputs to the AC line.
Typically, good DMMs withstand 240-VAC overloads when measuring resistance. And for many years, it has been the norm to protect oscilloscope inputs from 240-V power, even on the sensitive ranges below 5 mV/div. Usually, this is specified as <400-V peak overload capability.
Standards
The key international safety standard for test and measurement equipment is IEC 1010. Despite the fact that this is an internationally recognized standard, it is not acceptable for use in Europe. The so-called harmonized standard, EN 61010, is the relevant document.
Many people fail to realize that the law in Europe is specifically worded to demand compliance with the local variant of the standard, if such exists. This means that EN 61010 takes precedence over IEC 61010. The IEC has renumbered many standards by adding 60,000 to the old number, making them line up better with the harmonized European standards.
EN 61010-1 now requires that inputs be marked with their Installation Category.1 This is getting tricky now. It is very easy to mishear this as Insulation Category in verbal communications.
Installation and Insulation
Installation Category is a technical term worth getting to know. The subject relates to voltage transients on a power line and is derived from IEC 60664. There are many causes of voltage transients on the power lines, and lightning is far from the only one.
Other causes of transients include flashovers, network faults, switching on/off heavy plant, and current transients from older equipment such as welders.2 As a result, the AC line must be considered as having these transients as an expected and statistically predictable phenomenon.
Data relating to line transients has been enshrined in the safety standards for items such as switches, filters, transformers, and switched-mode power supplies. The idea is not only to protect against electric shock, but also to maintain functionality.
If the equipment is physically, and therefore electrically, close to the source of the transient, then the transient will be larger. The transient will dissipate as it travels down the line, not only because of the attenuation with distance, but also because it will expend energy destroying other equipment on the way. For example, if the electricity is supplied by an overhead feed from a substation, then the houses closest to the transformer will be hit the hardest should the transformer or overhead lines take a lightning strike.
The possible proximity to one of the many types of transients is defined as one of four possible types of location. These are listed in Appendix J of EN 61010-1 as well as in IEC 60664.
Ordinary wall sockets, such as found in a factory or home, are classified as Category II. The least severe is Category I, which is a filtered supply like the one you would find inside equipment after going through a chassis-mounting filter and a transformer.
Let’s take the case of a transformer whose output is entirely remote from interaction with the user. If the insulation of the transformer breaks down due to a transient overvoltage, then no shock damage can occur to the user. Nevertheless, there is a requirement for the transformer to maintain its basic operation.
The safety standard requires a basic level of insulation, specified in tables as basic insulation. For example, for 230-V operation, the basic insulation requirement at Category II calls for a test of 1,350 V rms at 50/60 Hz, 1,900 VDC, or 2,500-V peak impulse.3 The manufacturer is free to choose any one of the three test voltages.
If the output of the transformer is accessible to the user, then a higher standard of insulation is necessary. Imagine the case of a model railway transformer. The likelihood of someone touching the tracks is very high so the standard imposes a higher level of protection to guard against injury. In this case, the higher test voltage from the reinforced insulation group is needed.
The other way to achieve better insulation is double insulation. Do-it-yourself enthusiasts are familiar with this on power tools, where a plastic button presses on an already insulated switch. There are two separate physical barriers, double insulation, between the user and the dangerous voltages. To give an idea of the increased requirements, the tests for this reinforced/double insulation are 2,300 V rms at 50/60 Hz, 3,250 VDC, or 4,250-V peak impulse.4
Inputs
Since the new requirements dictate that signal inputs must be marked with an Installation Category, the manufacturer has to make a technical judgment.5 Does the input have to survive the overvoltage event, or does it merely have to not cause direct injury to the user?
Let’s take an input that has traditionally been marked as <400-V peak. Is it safe to mark this as <250-V Category II? We can be very confident that the input will withstand a short-term overload when connected to the AC power. This is what it was designed to handle. However, if the overload remains on for any appreciable time, there will be the additional risk that a transient will come along which will certainly blow the input.
The technical judgment here includes the consideration that this transient is almost like a double-fault condition. The safety standards specifically require you to deal with a single-fault condition unless this fault causes other parts to fail.
The other technical judgment is whether the instrument now is unsafe. Let’s assume that the insides of the instrument are not accessible to the user. Consequently, there is no shock hazard to the user. In this case, we only need to meet basic insulation requirements. The question then arises about whether the nonfunctioning of the device is unsafe.
By analogy with the transformer, the input must survive a transient at the stated basic insulation level for that category of input. If a <250-V Category II rating is to be claimed, then the input now must withstand a test voltage of 1,350 V rms at 50/60 Hz, 1,900 VDC, or 2,500-V peak impulse. And yet, you do find equipment with inputs rated as <400-V peak, <240-V Category II. This contradictory statement obviously is of some concern.
Within Europe, interpretation of these safety standards is left up to the manufacturer. This means that one manufacturer can interpret the standards in a completely different way than another, and there is nobody to say that this is incorrect.
Even if an accredited test house certifies the product to a particular safety standard, the house does not have the authority of law. In fact, the only authority of law is a court, which would decide the fate of an individual who signed a safety certificate.
The test house does not sign this certificate. It is signed by a European resident who is assuming responsibility—and liability—for the product. This signatory is not required to have any training in electrical engineering. It is assumed that the person who signs such a certificate would get the best available technical advice. To do otherwise would invite a charge of negligence should anything problematic occur with the product.
One large company has opted for a <400-V peak, <100-V Category I rating on some of its equipment. Now at this level, the overvoltage tests are 500-V peak, 350-V rms at 50/60 Hz, or 500 VDC. This is a much more plausible scenario, and inputs can realistically survive this reduced onslaught.
There is an additional and somewhat compelling argument regarding the safety aspect of a broken input on a measuring device. If this measuring device fails when connected to a live supply, then it may indicate that it is safe to touch this live circuit. That is clearly unsafe.
References
- EN 61010-1: 1993/A2: 1995, Paragraph 5.5.5.
- Capacitors for RFI Suppression on the AC Line: Basic Facts, Fourth Edition, Evox-Rifa.
- EN 61010-1: 1993, Tables D3 and D4.
- EN 61010-1: 1993, Tables D9 and D10.
- EN 61010-1/A2: 1995, Clause 5-1-5.
About the Author
Leslie Green, CEng MIEE, is a senior principal engineer at Gould-Nicolet Technologies. He has 20 years of experience in developing test and measurement equipment. For the last 14 years, Mr. Green has been developing oscilloscopes for Gould-Nicolet. Gould-Nicolet Technologies, Roebuck Rd., Hainault, Ilford, Essex IG6 3UE, U.K., 011 44 181 500 1000.
Published by EE-Evaluation Engineering
All contents © 2000 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.
October 2000