Netonix Forums

The fan in one of my switches gave up :(

This was in a fairly small enclosed room with minimal equipment inside - we had a hot summer, at times it was running at 6000RPM 8 hours/day, for weeks. Fun fact: according to my monitoring, it rotated approximately 658 million times before croaking.

This switch is outside its warranty period, am going to replace the fan myself (with the handy replacement part sold by Netonix :)

Of the WS-12-250-AC models we have in production we've had 60% of them suffer a fan failure. Yes, we've tried loosening the housing screws but that only slightly delays the inevitable. I'm curious what others have experienced.

As I know, only one was lost and we use mainly WS-12-250-AC.

Most all of my switches in service at my WISP are the original WS-24-400A units and in almost 4 years I have replaced I think 2 or 3 fans

9 switches with 3 fans each = 27 fans

3 / 27 = 0.1111 so 11% but that is in +/- years

My boxes get pretty dusty and warm so fans run a lot.

I use SUNON Vorpal magnetic bearing fans which are not cheap at $4.50 each our cost.

Normal no name fans from China used in most devices cost about $1.50 each

So we are buying good fans the environment of WISPs is just pretty harsh.

Now I have another 2 WS-24-400A switches in my office and they have never had a fan failure so I think it is all in the harsh environments?

sirhc wrote:Most all of my switches in service at my WISP are the original WS-24-400A units and in almost 4 years I have replaced I think 2 or 3 fans...

The WS-24 I got around that time is starting to rumble just a little. Loosening the screws just a tad made them quieter but loosening them more made them rumble again. They have a Goldilocks sweet spot. In the 3 decades I worked in IT, fans have always been the Achilles' heel and I kept an inventory of spares whenever I could. In some cases, I'd have to peel off the label to get to the bearing to lubricate them.

For sure where possible, I setup SNMP or syslog monitoring to alert me of fan failures and where I couldn't, I would tour periodically to look and listen. I always dreaded the phone calls where some critical piece of kit died. Often it included them saying, "it was making a real racket for a long time but them the noise stopped so I didn't bother to call". When the noise stops it usually meant the fan stopped.

We do sell spare fans on our web store, most vendors do not and make you hunt for something to work.

https://www.netonix.com/replacemet-parts.html

Having over 200 of these switches in use, I've only had one unit that had fan problems because of how tight the screw is.

I've had outright fan failures at two sites. That was environment related. Red dog gets into everything!

sirhc wrote:
So we are buying good fans the environment of WISPs is just pretty harsh.

Now I have another 2 WS-24-400A switches in my office and they have never had a fan failure so I think it is all in the harsh environments?

That certainly seems to be a contributor. One of our recent failures was in a switch located in a clean, temperature controlled indoor environment, but the rest have been inside outdoor enclosures at remote sites. Being enclosed they don't get real dusty but they do get very hot in the summertime and the fans run full speed for many months. The biggest concern of course is a fan failure in the summertime can quickly spell doom for the whole unit. However, when they have usually failed is in the Fall when the ambient temperatures cool down and the fans begin to modulate their speed. With the reduced speed input, the fan will seize and go to zero. The switch temps then soar. Sometimes the fan will then restart and run intermittently as temperatures rise and fall but sometimes they don't restart at all without mechanical help.

I am not sure what to tell you Rebus?

Where as some fans fail the percentage of fan failures is low or so it would appear based on RMA's for fan failures and the number of replacement fans we sell.

We do sell replacement fans which users can keep some in stock in the event of failure.

The switch will send you alerts if the fan fails if you set up SMTP alerts or monitor via SNMP.

Total replacement fans sales since we started in business:
Long cable used in old WS-24-400A/B and WS-12-250A/B = 63
Short cable used in all other models that use fans = 140

RMA fan replacements = <100

Total replacement fans = 303

Considering 40,000+ switches sold that would put failure rate at 303/40,000 = 0.7%

I am sorry you seem to be experiencing a higher than normal fan failure, not sure why. We do buy GOOD quality fans not cheap OEM fans.

SUNON magnetic bearing or maglev fans are not cheap and SUNON is recognized as a quality fan manufacturer.

Sirhc - can you post the Sunon part number for the fan used in WS-8-150-DC? And bonus points if you also have MTBF data for the fan.