EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: dkuhn on August 27, 2018, 04:50:56 pm
-
I work in Enterprise IT and we have recently been hit with a rash of component failures on a specific line of network switches. The manufacturer is proactively sending out replacements for them, which is swell, but I'd love to figure out the failure mode as I find this stuff interesting.
The failure is specifically related to the cooling fan in the back of the switch failing over time. The manufacturer claimed the issue was fixed about a year ago, and how have sent out another batch of proactive replacements that also covers the devices that were already replaced. They claim that increasing the minimum fan speed prevents the issue from presenting itself. This seems odd to me. If it's a bearing failure, wouldn't spinning faster reduce lifetime? Trying to make some sense of this with my colleagues and we haven't really gotten anywhere. I'd potentially be willing to do a teardown of one of these devices if that helps.
https://documentation.meraki.com/zGeneral_Administration/Support/MS225%2F%2FMS210-48LP%2F%2FFP_Proactive_Replacement
Thanks in advance to anyone that can hopefully lend some light on this! (Total newbie when it comes to electrical things, apologies if I didn't include enough info)
-
I am assuming that if it is mounted sideways (vertically), not having a fast enough rotation may allow gravity to pull down and minimize the space between the center coil and magnets in the fan, causing wear and tear on the top part of the central hub. Increasing rotation creates an "air cushion" that maintains the space better and reduces wear?
Perhaps in a related issue, I've had a bunch of fans for my Dell CPU cases in my entire office (several machines) go after only a few years. They would start making these funny noises (like chirping sounds) and then eventually seize up altogether. And these fans were in some ridiculously customized "blower" type cases that clipped exactly into the chassis so you could not find an easy replacement. They looked kind of like this:
(http://www.censuspc.com/images/D/d_16722.jpg)
Anyways, when I opened them up, the central hub was full of rusty metallic dust and bone dry. The coil was in the central hub, and surrounding it (part of the rotating plastic fan) were the permanent magnets. The plastic "blower case" also seemed slightly warped and because of the tight tolerances it seemed that the fan was scraping the inside edge of the opening.
Perhaps the plastic drying out and warping also caused premature failure, maybe due to heat in the case?
-
Many fan bearings nowadays have "hydrodynamic bearings" which means that the bearing surfaces are separated by a thin layer of grease/oil. This needs certain speed range to work (comparable to car hydroplaning/skidding on water)
If the speed is too slow there is going to be direct metal-to-metal contact and associated wear.
Ball bearings also have recommended speed range that would depend on lubricant viscocity etc.
-
I am assuming that if it is mounted sideways (vertically), not having a fast enough rotation may allow gravity to pull down and minimize the space between the center coil and magnets in the fan, causing wear and tear on the top part of the central hub. Increasing rotation creates an "air cushion" that maintains the space better and reduces wear?
Perhaps in a related issue, I've had a bunch of fans for my Dell CPU cases in my entire office (several machines) go after only a few years. They would start making these funny noises (like chirping sounds) and then eventually seize up altogether. And these fans were in some ridiculously customized "blower" type cases that clipped exactly into the chassis so you could not find an easy replacement. They looked kind of like this:
(http://www.censuspc.com/images/D/d_16722.jpg)
Anyways, when I opened them up, the central hub was full of rusty metallic dust and bone dry. The coil was in the central hub, and surrounding it (part of the rotating plastic fan) were the permanent magnets. The plastic "blower case" also seemed slightly warped and because of the tight tolerances it seemed that the fan was scraping the inside edge of the opening.
Perhaps the plastic drying out and warping also caused premature failure, maybe due to heat in the case?
Oh, now this definitely seems like a potential winner here. The switches have the same style of fan blower if I remember correctly. I'm going to go find one and tear it apart a little later to see if that is indeed the case.
Many fan bearings nowadays have "hydrodynamic bearings" which means that the bearing surfaces are separated by a thin layer of grease/oil. This needs certain speed range to work (comparable to car hydroplaning/skidding on water)
If the speed is too slow there is going to be direct metal-to-metal contact and associated wear.
Ball bearings also have recommended speed range that would depend on lubricant viscocity etc.
This also makes a lot of sense. I also reached out to a contact I have at their engineering department to get their input.
-
Thanks in advance to anyone that can hopefully lend some light on this! (Total newbie when it comes to electrical things, apologies if I didn't include enough info)
Rule #1: do not 100% trust manufacturer excuses explaining failures. They will try to save their reputation by all means, putting blame on anything but themselves. Provided reason of failure can be complete BS or maybe not. Only manufacturer knows, but they will not tell you anyway :)
They say:
We have redesigned and tested the switch with new fan hardware that is both more powerful and quieter than the previous component that caused issues.
It is obvious that fan they uses fails prematurely and they had to pick another one. End of story.
-
Everything has a life expectancy and it looks like the fans that are failing have reached theirs. I have had to replace many fans when the lubricant dried out and there was steel to steel contact in the bearings producing the reddish iron oxide that you see. If the fan stops then the stalled motor is just acting as a heater and if the temps get too hot the plastic housing/blades can soften and distort.
The best plan would be to have a maintenance plan so the fans are replaced on some schedule to try to prevent this from happening at the most inconvenient time. Check when the units were bought and figure how long it took for the fans to start failing then start a maintenance plan to start replacing the fans at, say, 75% of that time when you can schedule a planned downtime.
-
Thanks in advance to anyone that can hopefully lend some light on this! (Total newbie when it comes to electrical things, apologies if I didn't include enough info)
Rule #1: do not 100% trust manufacturer excuses explaining failures. They will try to save their reputation by all means, putting blame on anything but themselves. Provided reason of failure can be complete BS or maybe not. Only manufacturer knows, but they will not tell you anyway :)
They say:
We have redesigned and tested the switch with new fan hardware that is both more powerful and quieter than the previous component that caused issues.
It is obvious that fan they uses fails prematurely and they had to pick another one. End of story.
Totally get this. I'm more curious about the actual failure mode. When I pull one from production, I'd like to tear it down and send the part to someone who knows more about it to figure out what actually caused it to fail so we can get some unbaised information. This effects tens of thousands of businesses and will incur significant costs to remedy (travel, etc) and I'd like to have a specific failure mode to justify it other than "it failed, again".
Everything has a life expectancy and it looks like the fans that are failing have reached theirs. I have had to replace many fans when the lubricant dried out and there was steel to steel contact in the bearings producing the reddish iron oxide that you see. If the fan stops then the stalled motor is just acting as a heater and if the temps get too hot the plastic housing/blades can soften and distort.
The best plan would be to have a maintenance plan so the fans are replaced on some schedule to try to prevent this from happening at the most inconvenient time. Check when the units were bought and figure how long it took for the fans to start failing then start a maintenance plan to start replacing the fans at, say, 75% of that time when you can schedule a planned downtime.
These devices are less than a year old across the board. This is a new model they released that have premature fan failures across the entire line.
-
"These devices are less than a year old across the board. "
That isn't old age, it is a manufacturing defect. It sounds like the fans were not properly lubricated. I thought you were going to say the fans were running 24/7 for several years. That failure is not acceptable.
-
"These devices are less than a year old across the board. "
That isn't old age, it is a manufacturing defect. It sounds like the fans were not properly lubricated. I thought you were going to say the fans were running 24/7 for several years. That failure is not acceptable.
Yeah, they're doing the right thing and replacing every device, but I'm still curious as to what the specific failure mode is which I imagine we won't know until a teardown of a defective unit is done.
-
I'm still curious as to what the specific failure mode is which I imagine we won't know until a teardown of a defective unit is done.
Exactly. You can leave one unit and wait till it's fan fails, then tell us all what exactly failed :-DD
You see? - We, sitting where we are, cannot tell whats wrong with those fans, unless we buy faulty equipment and inspect. Obvious failure of fans - shitty bearings.
[edit] In short about actual most probable failure mode: bearings.