Shock mounting IDE hard drives in automobile application 6

[erasmu]

I need to mount two IDE drives in an automobile application. The goal is to prevent drive damage from shock and vibration caused by auto contact with potholes, curbing, etc. I was looking at EAR brand mounting devices. These are glorified grommets that can mount easily in sheet metal brackets with special sholder screws. Is this approach adequate? Is anyone familiar with this problem?

 

[Tunalover]

erasmu-
I've spent time wrestling with the issue of shock-mounting mobile disk drives. I'm afraid it's not all as easy as you might think. The grommets certainly help but that's no guarantee the disk drives won't see G-levels beyond their rating.

You need to find a random vibration PSD profile representative of cars and/or trucks. If the vehicle is an off-road vehicle (e.g. SUV) then that's a different story than for plain old cars. A good source of random vibration PSD profiles (G^2/Hz) is MIL-STD-810. Choose a profile for the vehicle and environment then determine Grms (=the square root of the area under the PSD curve). Then contact LORD INDUSTRIAL or other mfgr of shock mounts and they will help you pick one. One problem is the lack of sufficient sway space. If the available sway space is unchangeable then provide the sway space to the vendor and he'll help you pick out the best one (not necessarily a sufficient one) for your Grms level.

To determine the fragility level (in G's) of the disk drive, be careful about G levels and pulse durations published by the disk drive vendors; the G levels appear to be huge (>1000G's) BUT the pulse width is so small that the energy in the pulse is much smaller than the amplitude will lead you to believe. The real-life pulse width will be around 10msec. You'll have to figure out the equivalent fragility level for a 10msec pulse width. Using a vendor-furnished half-sine pulse of, say, 3000g amplitude and .003 sec duration, you can analytically arrive at the reduced fragility level; set the integral from 0 to 3.1416 of 3000*sin(333*t)dt = the integral from 0 to 3.1416 of G*sin(100*t)dt. Calculate G from that equation and it will give you the equivalent fragility at the more realistic pulse width. Provide this figure to the vendor. Unless your an expert at vibration, I would leave any further calculations up to the vendor.

Regards,
H. Bruce Jackson
aka Tunalover

 

[Tunalover]

erasmu-
My apologies. I provided a senseless equation to you to calculate the disk drive fragility level @ a 10msec half-sine wave pulse width that is equivalent to a 3000g fragility level @ a 3msec pulse width. The right equation is:

Set the integral from 0 to .010sec of [G*sin(100*t)dt] and

equate it to the integral from 0 to .003sec of [3000*sin(333*t)dt]. Solve for G.
My mistake. I hope this is helpful.

Regards,
H. Bruce Jackson
aka Tunalover

 

[ko99]

FYI, some drive shock-mount info I ran across:

http://www.chassis-plans.com/shock_and_vibration.html

Good luck


ko

 

[Tunalover]

erasmu and k099-

Don't believe everything you read in the papers. The article referenced by k099 says "hey, why not just hard mount it to the chassis?. The disk drive is so cheap that you can just replace it when it goes bad."

That is an irresponsible position for the following reasons:
1. The data on the disk drive is often priceless. In comparison, the disk drive itself is peanuts.
2. To hard mount the disk drive provides the lowest possibility reliability.
3. The reference article says the disk drive mfgrs have gone to great lengths to improve the shock performance. In my experience they jacked up the G-levels (>1000g's) but narrowed the pulse durations to e.g. .003sec. That is a marketing gimmick because few real-life collisions see such a small shock pulse. By lowering the pulse width it enables the mfgrs to elevate the amplitude which is what most people pay attention to. Don't fall prey to the gimmick!

I have tested shock drives to railway levels and they dropped like flies without adequate shock mounts.

Tunalover

 

[ko99]

Tunalover,

You are quite right about not believing everything you read, but I don't think the article was suggesting drives should be allowed to fail!

I interpreted it to suggest the entire system should be shock-mounted if necessary rather than attempting to further isolate the drives from the system.

However, this is not my specialty and I'll defer to others to interpret the rest.

ko

 

[IRstuff]

The article makes 3 points:

> Most drive isolation systems are poorly designed or implemented, and are simply in place to provide a "feature" and not necessarily better performance

> A drive isolation system will wind up with a relatively high natural frequency, thereby potentially causing drive induced self-resonances that may affect the operational performance of the drive, particularly since the drive is designed for a rigid mounting.

> A "better" solution is to isolate the entire system, thereby maintaining the rigid mounting of the drive to the system, while still isolating the system from the external disturbances. Additionally, since the sprung mass is larger, a lower natural frequency can be employed, thereby providing an overall better solution.

TTFN

 

[Tunalover]

Thanks guys for pulling me back to reality. I didn't read the article fully. It's just that I've seen such idiotic things done with disk drives... I over-reacted.

As for the automotive application, to isolation-mount a next higher assembly will require even more swayspace and beefier shock mounts than the disk drive alone because of the increased mass. I know of at least one mfgr who makes a ruggedized IDE disk drive assembly but IT IS HUGE (for internal sway space) and is targeted to the harshest military environment.

For your application I would recommend shock-mounting only the disk drive and recommend that you call LORD INDUSTRIAL or other competent shock mount isolator. They have lot's of experience with this application and are usually happy to work with you to select (or design) an isolation mount at no charge if you can convince them that there are production orders at the "end of the tunnel."

Regards,
Tunalover

 

[Lambretta]

Hi
Have a look at Sorbothane rubber, I have used it in the audio field as a vibration absorbing material, it can also be varied in density to achieve different damping properties.

 

[erasmu]

Thanks for the suggestion. Sorbothane works great for protecting LCD displays. It makes sense that it could help with isolating disk drives too.

 

[rorschach]

The EAR stuff is much superior to the sorbothane stuff in my experience. The idea of cascade shock mounts is an excellent one. I once had an RFQ from the Navy to build a machine that could take >500G's. Our approach was to shock mount the drives using the EAR grommets, then shock mount the chassis using even larger grommets to the rack and shock mount the rack using enidine cable type isolators to the ship's deck. (they were worried about nuke torpedos going off under the keel...)LOTS of sway space too. We got underbid by somebody that bought one of the cheap korean cases with minimal damping using tiny SBR grommets. I doubt the system would survive, but they had one thing going for them. they were cheap!

The EAR stuff (and to a lesser extent, the sorbothane stuff) are viscoelastic urethane whereas SBR or natural rubber is not. at certain frequencies, the rubber will go into excitation instead of damping. the EAR stuff even at the resonant frequency provides some damping.

 

[Tofflemire]

Hi All,

Would you want to use a soft durometer of hard durometer?
That is always one of our debates here.

Tofflemire

 

[Heckler]

Depending on the application we follow IRStuff's solution and hard mount the disk then place the enclosure on a shock tray that can be tuned for the specific application. One consense reason for this is by the time all the shock & vib brackets, dampers, and such are install to support the disk.....you're run out of volume for anything else you want to put in the box.

It's really a catch 22 with the US Military moving to COTs electronic gear. They want reliability levels that are just not seen in the commerical world. Then you spend lots of money to develop lets say a conduction cooled power supply or processor board then it's end of lifed or the software is outdated.

Best Regards,

Heckler
Sr. Mechanical Engineer
SW2005 SP 2.0 & Pro/E 2001
Dell Precision 370
P4 3.6 GHz, 1GB RAM
XP Pro SP2.0
NIVIDA Quadro FX 1400
o
_`\(,_
(_)/ (_)

Do you trust your intuition or go with the flow?

 

[rorschach]

Tofflemire, the simple answer is: It entirely depends.

It depends on the mass, the excitation frequency, the amount of sway space available, and the amount of damping required. EAR has a nice cookbook formula that will help you determine the best match for your need.

 

[edoneill]

I have tried using Sorbothane for this application. It was a good dampener, but it has a problem with compression set. Do not use it under a constant load or it will experience this problem. The problem is worse with elevated temperature, but it is there at room temperature as well.