VOIP POE and Switch Power Supply Failure

[brazenhead]

Hi there,

here's an odd question that I've been asked to try to find an answer for: Has anybody noticed any correlation between the increased power demands of providing POE to VOIP phones and an increase in switch power supply failures? In other words, have you had any power supplies fail because they are being heavily loaded due to POE usage? I've only ever had a 5520 power supply fail on me once in past years so I don't think it is a concern, but then again, I don't have my switches power output maxed out like they will be shortly when we connect a bunch of Cisco power hungry VOIP phones.

Thanks in advance.

[Michael McNamara]

I haven't see any correlation or relationship between the two. Actually I've had very few issues with 5520s, probably because of where they are in their life cycle in relation to their MTBF (161379 – 210361hrs ~= 18 - 24 years!)

[Telair]

We run a lot of Avaya PoE switches where I am now (5520's, 5698's and some new 5650's on order).  So far the deciding factor about how long they last is if they are on the top floor of the building where the roof likes to leak ( twice in the last year ) and water flows down the bundled cables in to the switch.

My previous employer used Cisco 4500 chassis switches with 5 or 6 48-port PoE blades, and we did see a correlation between PoE loads and power supply failures.  They were cheap bastards and as long as they didn't exceed 100% utilization on one of the two power supplies, they would do it.  So we had a number of floors where the single power supply would be running at something like 98% maximum load 24x7.  It would blow within a few months and the redundant power supply would try to pick-up the load.  Then it was a race against the clock to see if we could get the RMA'ed power supply back and installed before the redundant unit would blow.  They did have a dust contamination problem as well in their closets.  But dust and a fully loaded power supply was guaranteed fireworks within a few months.  A few times we lost both power supplies as the backup exploded when it tried to take the full load in an instant.  On a few floors we had so many IP phones we had to put both power supplies in combined mode to power all the phones at once or else we would start having "rolling" blackouts as seemingly random phones would power down when someone else picked their handset up.  It was very disheartening to work like that when you could see the problems, but they wouldn't give you any resources to fix it with.

[brazenhead]

Yeah thanks for the responses. It occured to me after I posted this question that I could have just cited the fact that I have several 5520's that have been running for years that are maxed out on POE power output, loaded up with wireless access points, and I have never seen a power supply fail on those switches. And phones would be no different as far as a switch power suppply is concerned.

Thanks again.

[Dominik]

Most power supplys will do good for their liftime with high load.
The power supply killers are voltage peaks. It is a good idear to use a UPS to protect
your switch power supplys bevore these peaks.

You can also take a look at the new ERS4800 that have dual hotplugable power supplies like the ERS5600 switches.

Cheers