Why "certification" need not be a cost factor
"Certification" evokes all sorts of sentiments in those affected by it. Depending on whether one is dealing with the person who ordered the certification or the end user of the certificate, perceptions vary. One person might complain about the costs entailed by acquiring the certificate, another might automatically demand certificates without actually knowing what he can achieve with them.
The purpose of a certificate often seems to have been forgotten, which is why I shall try to illuminate that in the following.
First, one asks oneself
- What use is a certificate to me?
- What use is it to others?
- How much will it cost me?
A certificate is confirmation of a property by an independent authority. Anything and everything can be "certified" e.g. products, people and companies. Unfortunately, there is no lack of "independent" organisations that assiduously print out certificates whose value is questionable. What can one do against such practices?
The value of a certificate for the client depends on the acceptance and reputation of the organisation issuing the certificate. A certificate is only worth something to the client if the end user of the certificate is convinced that the certifying organisation has actually checked the properties warranted by the certificate and that the certified object really does have these properties.
Naturally the confirmed properties must also offer some benefit for the end customer. Sadly, one often sees certificates that only cover partial aspects so that their benefits are debatable. One can easily find examples of this such as electronic dictionaries for which a renowned organisation confirmed the availability of a certain number of words. Such a certificate says little about the rest of the quality of the product and is, therefore, of dubious merit.
The end customer or "user" of a certificate can spare himself the time-consuming work of testing the product features out himself. In the example given above, he need not count whether the manufacture was exaggerating the number of words given.
This brings us to the heart of the matter:
- A certificate guarantees the properties of an object (product). This is best effected on the basis of well-defined evaluations (assessments) and instructions given in recognised standards.
- The certificate is issued by a (hopefully) independent and trustworthy organisation.
- A trustworthy and generally accepted assessment procedure and certificate eliminates the need for the end customer to check the required properties himself (e.g. acceptance test, EMC test, operational reliability).
- The client demanding an assessment on the basis of an acknowledged standard and a certificate by an independent "third party" can use this for marketing purposes to build up a trustworthy reputation on the market.
It should also be noted that with relatively complex products, such as the PROFIsafe field devices considered here, numerous assessments and checks to prove their specified functional safety have to be undertaken. The question, however, is who can logically perform such tasks. In order to maintain and expand the manufacturer's know-how, it is advisable to carry out the relevant activities within the manufacturing company. These include, for example, design, calculations, fault analyses, FMEA, EMC. Independent "third party" evaluation or assessment is then limited to ensuring the completeness and correctness of the activities and that they are sufficiently documented.
Let us now turn to an example related to our practice.
Which properties of a functionally safe component connected to a functional safety fieldbus are of interest?
First of all there are:
- the functions of the components as such (e.g. distance measurement and output on Profibus and Profinet)
- the safety function and
- the parameters for both
A functionally safe absolute rotary encoder gives the absolute value of the position. This is its function.
Its safety function could be displaying the position correctly within a defined area of tolerance.
If an internal error occurs that leads to the issuing of an incorrect position on the bus, then the sensor has to take up its pre-defined safe state. This can mean issuing an error message or "halting" communication. The following picture illustrates this.
The parameters of the safety function would, in this case, be the specified permitted deviation, the time it takes before the error is detected and the safe state is adopted. In addition, there is the "SIL" or the Performance Level (PL), which indicate what risk reduction can be achieved with the device.
Besides these properties, the fact whether the sensor can fulfil its function and the safety function given the environmental conditions of its application is also of interest. The following are thus also confirmed:
- It can comply with the required EMC conditions.
- It can withstand the specified environmental conditions.
- It complies with basic product standards.
A certificate for such a product is only fully beneficial if all the relevant properties have been tested and documented. If anything is missing, such as confirmation that the product can bear up against the specified environmental conditions, then this property has to be checked by the user of the product himself using the manufacturer's details as a basis. In such a case, he has to compensate for the missing assessments in the certificate by carrying out his own tests or analyses. In general, this leads to more work and added expense.
Often, and particularly with functionally safe devices, there is, however, a lot of leeway when it comes to applying and interpreting norms and standards. That is why the spectrum ranges from "scanty" certificates that only confirm the EN 954-1 or parts of the IEC 61508 right through to all-encompassing certificates that cover all aspects of functional safety, application-specific standardisation and environmental influences.
This leeway also arises, amongst other reasons, from demands that have not been clearly defined or from deliberately liberal interpretations - after all, the testing and certification centres also feel the pressure of competition and not only the product manufacturers.
Finally, a "cheap certificate" only pays off if the product has been designed in an equally "cheap" way, since with a reputable product the costs of extensive testing and certification of the product features are of little consequence compared to the development costs. The following diagram illustrates this.
With a cheap product, the "pure" development costs are, on the whole, lower since expense is spared with the development processes that should ensure quality. If the same standards were applied to its certification as with the high-quality product, then the cost of the tests necessary for certification and redevelopment is even higher ("we always have time to do it twice, but not to do it right"). In addition, a cheap product usually comes across as expensive to an end customer since he has to take into consideration maintenance costs and down times and these are, as a rule, higher than with reputable products.
One can therefore conclude the following:
- Manufacturers of reputable products have no problems with the quality, functional safety of a product or the costs of an independent test and certification process.
- Manufacturers of products of moderate quality, on the other hand, always have a problem in compensating the "additional costs" of testing and certification with higher product prices.
- Usually for the end customer a "cheap" product should appear "expensive".
With best regards,
Please send any comments on this expert opinion to: expert(at)profisafe(dot)net