Security fence with barbed wire.

Pain or Injury

Suppose you have a machine. Suppose the machine could do someone harm. Obviously, you should take measures to prevent the machine from killing someone or injuring them. But what about causing them pain?

Over the weekend I had to fix a farm gate. Several times I got a shock off the electric fence (see link to an interesting video below) near the gate. Not very pleasant but short lived and it didn’t do me any harm. Does that mean that causing pain is okay on the factory floor, as long as we do no harm?

Let’s investigate. A good place to start is the basic safety standard for machinery, ISO 12100 (safety of machinery – General principles for design – Risk assessment and risk reduction).

 Figure 1 Definitions from ISO 12100

Figure 1 Definitions from ISO 12100

You will note the definition of harm does not mention “pain”. Thereafter the goal of ISO 12100 is to “eliminate the hazard or reduce the risk associated with the hazard”.

If we look at the machinery safety standards ISO 13849 or IEC 62061, they also only talk about the injury. The minimum requirement to get on the table below is “reversible first aid” and worst case that gets you to SIL 1.

 Figure 2 - risk assessment and required safety integrity from IEC 62061

Figure 2 - risk assessment and required safety integrity from IEC 62061

The lowest rung for ISO 13849 is “slight (normally reversible injury)” and with the worst case for frequency, probability, and avoidance you can get to PL c (approx. SIL 1).

However, when we look at the industrial robot safety standards ISO 10218 and ISO TS 15066, we find a discussion of “pain onset” in reference to the PFL (power and force limited) collaborative robot applications where some contact between the robot and a human is allowed. So at least for collaborative robot PFL applications pain is the threshold and not harm.

 Figure 3 - when PFL operation is possible according to ISO 10218

Figure 3 - when PFL operation is possible according to ISO 10218

The above graph is misleading in that the threshold for pain onset depends on a number of factors such as where in the body the contact takes place. Contact with the head should be prevented.

I was surprised to see lots of relevant information on this in an audio-video standard IEC 62368-1.

 Figure 4 - IEC 62368-1

Figure 4 - IEC 62368-1

The standard gives as its objectives “This part of IEC 62368 is a product safety standard that classifies energy sources, prescribes safeguards against those energy sources, and provides guidance on the

application of, and requirements for, those safeguards. The prescribed safeguards are intended to reduce the likelihood of pain, injury, and, in the case of fire, property damage.”. For me, energy sources could include movement i.e., kinetic energy but I would need to back and read the standard to be sure.

I must say I like their approach. For instance, the standard goes on to define three kinds of people.

      * Ordinary person – under normal operating conditions to be protected against pain and injury.

      * Instructed person – under normal operating conditions to be protected against injury.

       * Skilled person – under normal operating conditions to be protected against injury.

It then goes on to describe energy source classes.

 Figure 5 Energy source classes from IEC 62368-1

Figure 5 Energy source classes from IEC 62368-1

Whether it is pain or injury then depends on how familiar someone is with the system. To me, it makes sense that skilled people have to take more personal responsibility than an ordinary person. I believe a SIL determination according to IEC 61508-6 applies a 10x factor in such cases.

Personally, I wouldn’t want to work with a machine that caused me pain so therefore I am very happy that ISO TS 15066 and ISO 10218 series both set the limits for PFL based on pain onset (see here for a previous blog which includes details on a machine for measuring pain threshold). I guess it depends on how often I would be subjected to the pain and how easy it was to avoid it. You could also, I imagine, claim that repeated painful episodes could cause physiological harm.

In fact, this is proposed in the latest draft of ISO 13482 where if the frequency of exposure is high the limits are based on pain onset vs. the slight injury if the frequency of exposure is low.

I imagine this blog is not relevant to any avionics or automotive readers. Hard to say with medical functional safety.

This is only a short blog and doesn’t go into topics such as ergonomics and repetitive strain injuries. Hopefully, there was enough to interest you and introduce the difference between pain and injury.

For an interesting IFA paper on pain thresholds, see here.

For the full set of blogs in this series, see here.

               One of particular interest might be this one on SIL determination

Since I opened talking about my recent experiences with an electric fence, there is an interesting video of people trying to take a bicycle off an electric fence available here

Parents

  • Curves 1-6 are psychoacoustic data that map out the threshold of

    hearing, the softest sound-pressure levels that can be detected,

    using the left axis.  Curves 7-12 show reports by individual experimenters

    on the SPL threshold that induces some sort of painful percept, using

    the right axis.  This data is from the classic Licklider chapter in, 

    Handbook of Experimental Psychology, S.S. Stevens, ed., 1951, p. 995.

    So, for at least one type of stimulus relevant to certain Analog Devices

    customers, pain is one of the parameters that have been mapped out.

    However, it's important to note that long-term hearing damage can 

    happen due to exposure to sounds well below the threshold of pain

    shown in this data -- and so, injury happens before pain for sound.

Comment

  • Curves 1-6 are psychoacoustic data that map out the threshold of

    hearing, the softest sound-pressure levels that can be detected,

    using the left axis.  Curves 7-12 show reports by individual experimenters

    on the SPL threshold that induces some sort of painful percept, using

    the right axis.  This data is from the classic Licklider chapter in, 

    Handbook of Experimental Psychology, S.S. Stevens, ed., 1951, p. 995.

    So, for at least one type of stimulus relevant to certain Analog Devices

    customers, pain is one of the parameters that have been mapped out.

    However, it's important to note that long-term hearing damage can 

    happen due to exposure to sounds well below the threshold of pain

    shown in this data -- and so, injury happens before pain for sound.

Children
  • Hi John, thanks for the information. I think laser eye safety is another area where injury can occur without pain as in some cases the laser emits at frequencies we don't notice until its too late. I have never heard of functional safety being used to prevent either noise or eye safety issues. In the case of laser eye safety according to IEC 60825-1 there are some parallels and they do reference some of the techniques from functional safety such as a qualitative FMEDA. My blog mostly concentrated on machine safety and robots but I guess the audio visual equipment standard I quoted should consider ear damage. 

  • Historically, hearing aids and headphones were both purely

    analog ... later, when they became digital (and headphones

    became in-ear in form factor), the digital so closely modeled

    the analog antecedent that there was probably no reason

    to consider the functional safety toolkit ... but now, given

    how these in-ear devices are complex microprocessor systems,

    I wonder if its time has come ... consider what could happen

    if a failure mode leads to delivering white noise above the

    threshold of hearing damage to the listener. By the time

    the listener pulls the devices out of ear real damage could

    be done.