What constitutes a Modification?

[David Rooney]

From our BAFE rules re issue of Modification Certs.

5.3.1 For fire detection and fire alarm systems where there is no existing SP203 BAFE Certificate of Compliance, an SP203 certificate for a modification can only be issued if the modification meets one or more of the following criteria:- .

i) replacement of the original control and indicating equipment (cie) with a different type of cie (e.g. an addressable cie replaces a non addressable cie) or a power supply or repeat indicator panel is added to the system

blah blah blah.....

So, if we are replacing a control panel like for like, this would not constitute a modification?

From what I've read in 5839 ... (thanks Allen!)

46.4.1 Commentary
Modifications to the system can arise for a number of reasons. Examples include:
— extension of the system to protect areas of the building previously unprotected or newly constructed;
— change of detector type as a result of changes in occupancy or the occurrence of false alarms;
— re-siting of, or increase in the number of, detectors and/or fire alarm devices to take account of changes in the layout of the building;
— reconfiguration of the system (in hardware, software or both) to change the cause and effect logic in order to facilitate filtering of false alarms.

.... it seems to agree ... we are not changing or reconfiguring, just replacing.

Therefore should we re-commission the complete system, only test a certain percentage or what exactly... and where is this guidance in the BS please.

[Wiz]

David, There may not be a specific BS recommendation for the precise scenario you are describing, but it is clearly not a 'modification' as per the BS recommendations. The work you describe is obviously just a 'repair' and would be treated just the same as a 'like for like' replacment of a faulty smoke detector. The end result of carryning out the work is that everything stays as it originally was, so it is not a modification.

Commonsense dictates that the appropriate tests after completion of the work is to create a fire condition per zone, ensure sounder outputs and other outputs are functioning and check that the appropriate fault conditions are indicated for every fault monitoring circuit. You surely do not have to re-commission the entire system.

The above is all on the basis that the system was originally commissioned correctly and that servicing as per BS has been carried out since commissioning.

[David Rooney]

But without re-commissioning how do you guarantee your new processor is doing it's job properly for every device on the system?

Who is to say the manufacturer hasn't updated all the boards/processor etc, just because the packaging looks the same it could be a completely different panel inside...

[Galeon]

Had This one before , you have to test the lot trying to do this and get paid for it is another story.

[Allen Higginson]

But without re-commissioning how do you guarantee your new processor is doing it's job properly for every device on the system?

Who is to say the manufacturer hasn't updated all the boards/processor etc, just because the packaging looks the same it could be a completely different panel inside...

Who's to say that the processor won't fail 5 minutes after you have tested it?
Confirm the C & E and keep your fingers crossed best option.

[Wiz]

But without re-commissioning how do you guarantee your new processor is doing it's job properly for every device on the system?

Who is to say the manufacturer hasn't updated all the boards/processor etc, just because the packaging looks the same it could be a completely different panel inside...

David, although microprocessors can be found in all types of panels I assume you are specifically talking about an addressable system. The processor in a new replacement panel is not going to have faults that affect just some specific devices without you knowing about it. If there was a major operational change betwwen the old panel and the new panel, the manufacturers would surely highlight it anyway but I'm also sure you would certainly quickly know about either automatically or when you carried out the basic testing of the new panel.

So if you load exactly the same configuration as used in the previous panel, I know that you understand that the panel talks to all the loop devices it knows about continuously (and also looks for ones that 'shouldn't be there!) and therefore you will also understand that any problems with communication between the panel and the devices will therefore be highlighted in just a few minutes, if not seconds. To put your mind further at rest, you will then only need to create a few fire and fault conditions on random devices, and I can assure you that if these all work ok, then you can be 99.9999999999999999999999999999999999% certain there will be no problems. You don't need to test every device on an existing system that has been previously commissioned, and subsequently correctly serviced, just because you have replaced the CIE. Obviously, some checks of any other inputs and/or outputs etc being used the panel should also be made. All you need basically need to do is carry out a check of each circuit you had disconnected from the old panel and re-connected to the new panel, but you only need to test one or two of each type of possible fire/fault condition relevant to that circuit. Fully recommissioning every mcp, detector, alarm warning device etc. etc. would surely be unnecessary and over the top. This is why the BS doesn't contain a recommendation that you need to do such.

[David Rooney]

But we have had jobs where the panel firmware has been up graded and suddenly has "issues" with certain devices on the loop that are not necessarily indicated on the panel.

We have had repeater panel problems and problems with older style CHQ-BS's not operating with later firmware....

[Benzerari]

Old Morley analogue addressable panels do not like Discovery but still OK with XP95.......    just an example....   

I would test some devices in each loop

[Benzerari]

There is a similar scenario;

Have extended a conventional system, an old conventional panel by new HAES panel, in addition to some new devices, by the way of testing all new devices OK, I tried to test the existing ones and found 4 existing MCPs not working at all, while some other existing ones did work.

I was in a position to replace the 4 existing non working MCPs, with no additional cost, otherwise if I leave them, the customer may say one day, this happened since you upgraded the system at some stage.... and we haven't been told previously that they are not working so why particularly now, it's better to cover your back some times with little loss.

This is to say it’s better to test most of the devices in modification if not all. I do some times trust experience more than BS and BAFE  texts

[Galeon]

We always give the choice of doing a complete one off test of the system , if they dont go for it anything we pick up whilst we are testing gives a bill to the client.
Your undertaking is more than honourable and fair which is reasonable, but why should you be picking up the pieces, of persons who don't give a monkeys before you got there.

[spanner]

IMO, if you are to replace a panel, (even if it is like for like) you should carry out a full system test. I know it can be a pain, but at least your arse is covered. Bare in mind that you will be signing the docket to say that your replacement panel works, This could mean that the blame falls back on yourself if the panel decides to go tit's up and not operate.
Also, what if your configuration was to become corrupt when you upload it, say a few devices or a few lines of C&E. the panel may show healthy, but might fail in an emergency.

[Allen Higginson]

Good practice but not necessary if like for like.