IEC 60601-1-8 clause 6.4.1 (statistics of the distribution of the sum of the maximum ALARM CONDITION + SIGNAL GENERATION DELAY)

[manlongcheung]

The clause states:
"6.4.1 * ALARM SYSTEM delays
If the sum of the maximum ALARM CONDITION DELAY plus the maximum ALARM SIGNAL
GENERATION DELAY is greater than 10 s, then the statistics of each distribution or statistics of
the distribution of the sum shall be disclosed in the instructions for use."

What sort of statistics is expected to be disclosed in the IFU?

Currently, we display our total alarm delay time by saying it's less than a certain time. e.g. < 30 s

Alarm condition	Priority	Delay
Leak Detected	Medium	< 30s

Would this capture the intent? or would you need a mean and standard deviation for each alarm times, which may be difficult to find out given the complexity of environment the device can experience.

Thank you very much!

 

[Steve Prevette]

The clause states:
"6.4.1 * ALARM SYSTEM delays
If the sum of the maximum ALARM CONDITION DELAY plus the maximum ALARM SIGNAL
GENERATION DELAY is greater than 10 s, then the statistics of each distribution or statistics of
the distribution of the sum shall be disclosed in the instructions for use."

What sort of statistics is expected to be disclosed in the IFU?

Currently, we display our total alarm delay time by saying it's less than a certain time. e.g. < 30 s

Alarm condition	Priority	Delay
Leak Detected	Medium	< 30s

Would this capture the intent? or would you need a mean and standard deviation for each alarm times, which may be difficult to find out given the complexity of environment the device can experience.

Thank you very much!

If I was given the task as stated, (no I am not familiar with the source document) I'd collect data from the system on alarm condition delay and the alarm signal generation delay (given that the alarm condition has triggered) and do a monte carlo simulation to get the response curve - you have a 1% chance detection with a second, 3% within 2 seconds, etc. Though a bug in this - for example if I think about an over-pressure situation, is the potential rate of change of the pressure in the system. I'd need to know given what creates the pressure, how fast could it increase? And if it was, for example, a positive displacement pump, it could increase quite rapidly. And would I have to account for say the existence of a relief valve?

 

[Tidge]

Currently, we display our total alarm delay time by saying it's less than a certain time. e.g. < 30 s

Alarm condition	Priority	Delay
Leak Detected	Medium	< 30s

Would this capture the intent?

It has been a few years since I engaged with a NRTL on 60601-1-8, but I believe that this meets the intent of the standard.

Per the informative annex, I believe that the intent of disclosing the delays (especially those with times that can exceed 10 seconds, per the collateral) is to bring a level of clinical understanding to the presence of an alarm condition. i.e. "how long has the leak been present?"

As a practical matter: I would want to know the maximum possible delay in the alarm signal (for a specific alarm condition) rather than try to understand a mean and standard deviation of times