Use hard drive anti-vibration grommets or not?

[Fraoch]

Just got a shiny new Fractal Design Define Mini to downsize from my current gargantuan case.

The hard drive sleds come with anti-vibration silicone rubber grommets. I believe @cyberjock offhandedly mentioned in another thread about a month ago that these are not good on hard drives - the resulting undamped vibration is not good long-term.

Should I use these grommets?

I've had a WD Black in my desktop Fractal Design Define R4 using the grommets. It's been running 24/7 for almost 3 years according to SMART.

 

[GrumpyBear]

The whole purpose of the rubber grommets is to dampen the vibrations. If you mount the drives (somehow) without the grommets then the vibrations from the drives will the be undampened.

I believe one of the differences between the enterprise NAS drives and home use ones is that the enterprise drives are designed to either dampen or withstand the vibrations from many drives in a chassis.

The WD Red line also states that they are designed for 5 or less drives in a chassis with the original firmware and up to 8 with the newer NASworks3 which feature improved vibration tolerances.

 

[cyberjock]

I haven't mentioned the rubber grommets in quite a while. But, when I saw someone do some tests in a lab, it turns out that the rubber grommets do two things:

1. They prevent the transfer of energy (and therefore vibration) from the hard drive to the chassis. Less vibration translates to less noise.
2. Since you are preventing the transfer of energy, that energy has to go somewhere? Yeah, it does... the hard drive itself has to deal with those forces.

Yes, the rubber grommets do absorb "some" of the energy, but it is not a significant amount of energy. Their primary function is to insulate the energy from being transfered to the chassis.

What I saw when we did tests with an ATX chassis that was popular at the time (I don't remember the case as this was about 7-8 years ago) was that the hard drives have to deal with much larger amounts of vibrational energy *because* the energy isn't being transferred. Not entirely unexpected if you think about it. You are literally taking a hard drive and making it basically "float" on the rubber. The hard drive isn't really "'fastened" down, so the hard drive pretty much has to vibrate much more to transfer the energy to the chassis (and that energy has to go somewhere). Everyone knows the whole "energy cannot be created or destroyed" stuff.

I never recommend using cases that require the rubber grommets, I won't buy them myself, and I *always* screw hard drives into their bays or slots in a chassis using as many screws as possible. The more screws the more secure locations to transfer that energy.

But, that's just my opinion. YMMV, this is without warranty, blah blah blah.

 

[Fraoch]

@cyberjock, that makes a lot of sense. Thank you!

I think I'll remove the grommets in this install. The case is quite quiet anyway, a little vibration won't be catastrophic. I'd prefer a little noise and longer hard drive life.

 

[Tywin]

What I saw when we did tests with an ATX chassis that was popular at the time (I don't remember the case as this was about 7-8 years ago) was that the hard drives have to deal with much larger amounts of vibrational energy *because* the energy isn't being transferred. Not entirely unexpected if you think about it. You are literally taking a hard drive and making it basically "float" on the rubber. The hard drive isn't really "'fastened" down, so the hard drive pretty much has to vibrate much more to transfer the energy to the chassis (and that energy has to go somewhere). Everyone knows the whole "energy cannot be created or destroyed" stuff.

The energy is dissipated in the rubber grommets, they are dampers.

 

[cyberjock]

The energy is dissipated in the rubber grommets, they are dampers.

Yeah, that's what good quality rubber should do. But that's not the kind of quality rubber you get.

Then, you have to add in the fact that the properties of the rubber change as you compress the rubber. Most of the ones in the market out there have a setscrew that compresses to a given amount based on the threads. It's a very delicate balance, and the 5 systems we tested all were MUCH worse off with the rubber grommets.

Worked on a sub for 2 years as the vibration analysis guy, so I had the equipment and opportunity to test things "just because".

 

[Bidule0hm]

The energy is dissipated in the rubber grommets, they are dampers.

Yes, but they allow a greater drive's movements amplitude.

I'm not sure that this is a problem though. Exterior vibrations are bad (especially in the same axis that the spindle axis because they can cause head crash or high fly R/W), but the vibrations induced by the seeks? They aren't in the same axis than the spindle axis and they are obviously in sync with the head movements (and it's a closed loop system so the movement is corrected).

I'm very curious to know if it impact the performance and also if it decreases the drive life...

 

[joeschmuck]

With today's hard drives for NAS use you will find that they have anti-vibration capabilities built in and this is where you will see statements like 5 or 8 drives can be mounted in the same rigid drive bay. The drives (as I recall) will "slightly" speed up or slow down their rotation in order to combat excessive vibration. I have six WD Reds all nut-to-butt and they are ultra quiet, absolutely no vibration and they have over 20465 hours on them and their warranty expires in Oct 2015, I'm sure they will surpass that. You will not see me bad mouth the WD Reds (2TB version). Okay, I can feel vibration when I place my fingers on the drives but it's very minor and almost difficult to feel.

My opinion is you can use the grommets or not, today's drives are not like the ones we had 6+ years ago.

I thought I'd inject a thought for the equipment on the other side of the grommets... So lets say you do have a hard drive that vibrates. What is that vibration going to do to your MB, PS, Seated cards? Those vibrations could spell disaster for your computer electronics.

Head crashes occur all the time in spite of best engineering practices to eliminate them but due to the very small mass of a head, damage is also minimal. I have seen some very serious head crashes and been the cause of a few myself on the old 14" platters and the heads are about 3/4" in size (lots of mass). I think that is getting a bit off topic though.

 

[Bidule0hm]

Yeah, but this is intended to avoid resonance of several drives together due to the rotation vibrations, not the seek vibrations (and they are far more powerful). I'm not worried about the rotation vibrations.

So there is head crash even in a perfectly fine drive? interesting :)

 

[SmallGuy]

I think it's difficult to have an advertised advice on this without physical facts. The only way is to proceed with environment trials (accelerometers on three axis...)
Depend on the the physical arrangement of the drive (vertically, horizontally, mechanical assembly...), the number of drives, the rotation speed, the design of the grommets, the material, the wind direction, and eventually the presence of bees.
I wonder what can happened if some bees are landing on your NAS, and decide to fart on it... Are the grommets useful in that particular case?
:P
Sorry but this topic made me laugh.

 

[DrKK]

Every hard drive I have ever installed in any case or chassis has rubber grommets. I have never had a hard drive fail.

Just a data point.

 

[cyberjock]

I think it's difficult to have an advertised advice on this without physical facts. The only way is to proceed with environment trials (accelerometers on three axis...)
Depend on the the physical arrangement of the drive (vertically, horizontally, mechanical assembly...), the number of drives, the rotation speed, the design of the grommets, the material, the wind direction, and eventually the presence of bees.
I wonder what can happened if some bees are landing on your NAS, and decide to fart on it... Are the grommets useful in that particular case?
:p
Sorry but this topic made me laugh.

You should laugh. Because the answer is not as trivial as I may have mistakenly made it sound.

Various things to consider are:

• Amplitude
• Frequency
• Axis
• Brand/model/firmware of disk
• How well balanced are the platters when spinning (some models have the platters secured at the top and bottom for extra support from platter vibrations when spinning).
  
• Kind of rubber
• Age of the rubber
• Compression of the rubber
  
• Temperature of the rubber
• Has it been frozen before (some rubber breaks down if it gets too cold... such as if it sits on your porch in the winter when it arrives from Newegg)
• I'm sure there's more I haven't mentioned, but I think you get the point

This is NOT a trivial question, and the answer isn't trivial either. I'm sure someone could design some rubberized system that is amazing. None of the ones I tested were worth anything, and all of them had significantly higher amplitude across the board, on all axis. So the watered down answer is "almost certainly isn't better", but I'm not the one that should be giving the expert opinion.

I also take the stance that rubberized hard drive bays didn't exist in the 80s, when things were MUCH more sensitive to vibration. Back in the 80s the people that played with computers were extremely hardcore compared to today's people that call themselves hardcore. I've heard that people that were around in the 80s knew so much about what was in the hardware they were buying that they wish it went back to those days because the cold hard facts are distorted with "company secrets" and marketing bullshit.... err.. I mean hype.

So I tend to think it's more logical than illogical to think that the rubber is a terrible and stupid idea, but since it adds hype and makes it seem better people will eat that stuff right up. Lets face it, companies lie to us every day, even more than 30 years ago. So I'm much more willing to buy that rubber grommets are a bunch of crap versus someone coming up with "this great idea" after hard drives had been around for decades. How many electronics have billed themselves as better/faster/smaller/whatever, but weren't?

Back in the day I had all my drives mounting in rubber grommets. I also had relatively higher failure rates (3x higher?) than I've had in the last 5 years.

Ultimately, it's your data, your drives, and your choice. I just know what I'd do and shared my experience and opinion. ;)

If the rubber makes you feel better, use it.

 

[no_connection]

I would say that "it depends".
Having a firm base for the disk head to work from does mean less hunting as it compensates for the disks counter movement.
Vibrations induced from the outside does however directly transferred to the disk then.
So Ideally I would want a heavy firm plate to mount the drive to, but that would be suspended with damping to the outside.

Still I wonder how much modern disks care about vibration, doubt progress have been standing still while rest of disk have been improved. Especially as it's ridiculous amount of precision required to position the head in the right place.

And remember, don't shout at your NAS

 

[SmallGuy]

Yes it depends.
You can think of this kind of device
Or this kind of device (not truly recommend hardware for FreeNAS)
I'm pretty sure the approach is different in both case. ;)

 

[Bidule0hm]

I've seen this video, pretty amazing and funny :)

But the only thing that interest me here isn't the exterior vibrations but the drive own seek vibrations. My question is: Is there any significant improvement in reliability (and performance as a secondary question) if we mount the drive in a rigid cage rather than with some movement allowance?

 

[cyberjock]

But the only thing that interest me here isn't the exterior vibrations but the drive own seek vibrations. My question is: Is there any significant improvement in reliability (and performance as a secondary question) if we mount the drive in a rigid cage rather than with some movement allowance?

I don't think anyone has really evaluated that. ;)

I think it depends on a laundry list of things like I mentioned above. It's one of those "depends on the situation" things.

 

[Bidule0hm]

Yeah, that's what I thought unfortunately.

 

[joeschmuck]

My question is: Is there any significant improvement in reliability (and performance as a secondary question) if we mount the drive in a rigid cage rather than with some movement allowance?

My ex-wife works for WD, I'd ask her but then she may never stop talking to me ;). I'm sure the drive manufacturers know all that kind of stuff.

The seek vibrations you talk of, well you need to temper it with knowing that both the spindle and the actuator are mounted to the same ridged frame. Those vibrations do not affect the platters/spindle at all. All you hear is the actuator coil and heads (both have low mass) moving around very fast and you must keep in mind that when a head is seeking to a location on the pack, it will slow down and from what I've read in the past it will not overshoot it's mark. The disk keeps spinning, the data runs under the head, and off to the next track to find a sector to read data.

If you want to get into it more, do some Google searching on it. You might be reading for a few days to gather all the data but I think you will find that any hard drive made in the recent past will be very solid and reliable. Use the drive warranty period to gauge longevity, after all that is what the manufacturers do no matter how you mount them. Speaking of mounting a hard drive, I recall many years ago a hard drive had to be mounted in one way, then a few years later it was two ways. Now days you can mount most hard drives upside down even (not that I would).

 

[Bidule0hm]

Thanks for this precise answer ;)

"when a head is seeking to a location on the pack, it will slow down and from what I've read in the past it will not overshoot it's mark." This is exactly what I've thought, so in theory there is no problem due to seek vibrations because there are compensated naturally by the positioning system as any other position error would be :)

 

[cyberjock]

They are *supposed* to be compensated for. But that compensation often requires recalibration (which the disks often do themselves).

Don't know if you know this, but when you first power on your disk and it warms up, the hard drive recalibrates itself multiple times over the first 2 hours it's powered on because as the components warm up the actual location of the data moves ever so slightly due to the temperature difference.

So yeah, there's constant calibration and such going on. That's not the *real* problem problem in my book. The *real* problem is vibrations that are so constant and consistent that the disk is having to correct for overshooting the desired sector regularly, resulting in much longer latency... or the ever worse head crash.