BS5839-1 and radio systems

[kurnal]

Looking at BS5839-1 it appears that where a network of radio linked panels is installed, failures of the radio communication between the nodes are treated exactly the same as a failure to communicate with an individual device, ie  if after trying to communicate for for two hours without success a fault condition must be recorded within 120 seconds (2 hours and 2 minutes after communications were lost.)

I would view a failure to communicate with an entire node on the network a much more significant fault than a failure to communicate with an individual device. Or am I missing something?

[Wiz]

Prof., I think the recommendations were based around what the existing equipment could achieve at the time rather than as some specifcally calculated important criteria.

It may also been thought that each 'node' was probably a self-contained unit where everything in and around that node would carry on working OK even if signalling between nodes had failed and therefore possibly be less than a total catastrophe.

I also believe the stricter BS recommendations are all concerned about failures that could happen during a fire rather than potential faults that may appear at other times i.e. it is thought that faults are generally unlikely to happen, and those that do, invariably happen during non-critical circumstances.

Finally, a two hour delay in ascertaining that there might be a fault is far better than a system with no monitoring whatsoever, and which could mean that the first indication of a problem is when the system didn't work in an emergency that occurs days, weeks, months or years after the fault actually first occured!

[colin todd]

Wizard, As always so sorry to correct you, but the fault monitoring recommendations of BS 5839-1 are not primarily related to what happens in a fire. That is why many types of fault warning can actually be suppressed during a fire, though not all of them of course.

With regard to radio, you transmit too frequently at your peril. That is one reason we are careful not to go overboard on the time to detect a fault.  It is not, in the current version, about what is on the market, as available systems can beat the recs.

[kurnal]

It may also been thought that each 'node' was probably a self-contained unit where everything in and around that node would carry on working OK even if signalling between nodes had failed and therefore possibly be less than a total catastrophe.

Thanks Wiz that is a good point and actually confirms what raised my concerns originally.

My initial concerns were raised during a fire risk assessment when I simply did a couple of tests on a networked system of EMS panels.

The  maintenance guy said he often had problems with his weekly tests and that one node was regularly in fault due poor radio coverage.

I found that the whole thing is compounded by a real mess in configuration and system design. The outcome of which was that we finally discovered that 2 break glass call points in one building were actually linked to a different panel in an adjoining building. The "engineers" are being called back in on Jan 3 to sort it all out and see what can be done to  improve radio coverage across the site.

[colin todd]

Kurnal, these issues do not relate to frequency of monitoring, which I thought was your initial point.

[kurnal]

Yes Colin you are right it was initially  a general point over the monitoring of comms failures and Wiz correctly put it into context by pointing out that just like any networked system each node should stand alone and provide protection within its own area (though global C&E instructions may be lost.)

The query was prompted by a particular installation and it was only later on doing more digging I found out  that it is a complete and utterly non compliant shambles thus confirming Wiz's point.

[Wiz]

Prof., it is difficult to write a C&E configuration for a networked system that won't be quite heavily affected by a failure of communications between nodes.
The best systems operate by the C&E configuration being saved at each node of the network so, in theory, the C&E could still function even with the loss of communication between nodes. What I mean by this is that the C&E programme isn't necessarily only held in a 'main' panel and that communication paths all need to be working for any C&E to happen. Therefore if the C&E configuration is 'that all alarm warning devices operate on any alarm input', then even with a communication path down between two nodes, an alarm input associated with, say, node 'A' of the network will still cause the sounders associated with node 'A' to operate.
However, if the alarm input is associated to node 'A' then the alarm outputs on node 'B' won't operate if the communications link between them is faulty, even if the C&E in both nodes says that it should.
The point Mr C.T. makes about the problems caused by having too many radio transmissions is very valid but was possibly more of a problem with the type of wireless systems available years ago. I'm sure that the timing of transmissions has been tightened up much more on newer systems and maybe it would be possible to monitor the communication links more frequently now.
Probably networks wired as a loop (for wired systems) and wireless systems where any node is in range of at least two other nodes is the first step to improve netwrok communication and monitoring.

[kurnal]

Thanks Wiz.