Keith Dutson wrote:
> A typical MTBF for EISD hard drives is 500,000 hours. However, this is just
> a statistical benchmark figure used by manufacturers. It is NOT a guarantee
> by any means.
I am well aware of that - the M in MTBF stands for mean, which is a
statistical term.
The guarantee is typically 3 years, which is very different from the 57
years MTBF. The MTBF on some SCSI drives is 1.4 million hours (160
years) but the guarantee period is 5 years.
However, whist I can't find it now, I remember reading on the Seagate
site that this MTBF applies only when the drives are replaced periodically.
The definition is different from what one normally associates with MTBF.
But I guess in the case of things like car cam-belts, the MTBF you get
if you replace them every 50,000 miles is far more important that the
MTBF you get if you run one cam-belt until the belt snaps and the engine
dies. If something is known to wear out in this way, it is not
unreasonable to replace it periodically if you wish to minimise the
frequency of unexpected failures. The same can apply to tubes too, or
anything else where they are known to wear out.
G8WRB
>
> -----Original Message-----
> From: amps-bounces@contesting.com [mailto:amps-bounces@contesting.com] On
> Behalf Of David Kirkby
> Sent: Sunday, September 11, 2005 7:16 PM
> To: David Lisney
> Cc: amps@contesting.com
> Subject: Re: [Amps] Screen Supply for 4CX250 series
>
> David Lisney wrote:
>
>>Hi, a couple of people emailed me off list for the circuit diagram of
>>my transistor stabilised screen supply, sorry I had a hard-drive
>>failure (how do they possibly claim such amazing MTBFs)
>
>
> The MTBF's on hard drives are quoted in an odd manner (well odd to me
> anyway), but I guess it has some logic to it.
>
> A decent SCSI hard drive will have an MTBF of over 1 million hours, but this
> does *NOT* mean if you take a load of hard drives and run them until they
> break, the mean time before failure will be 1,000,000 hours
> (114 years)
>
> The MTBF is based on the mean time for a random failure to occur, assuming
> that you replace the disks periodically in a preventative maintenance
> schedule. You need to replace them at the end of their service lives (5
> years typically for SCSI, less for other disks), even if the disks are still
> working.
>
> So if the service life of a disk is 5 years and the MTBF is 114 years, you
> will (on average) need to replace the disk 22 times over before a random
> failure will occur that will mean a disk fails under 5 years old.
>
> Of course, few of us bother replacing disks that are not broken, which is
> why we don't see average lifetimes close to the MTBF's.
>
> If you make heavy usage of disks (lots of DVD ripping or similar) then SCSI
> are a lot better. I knew someone who was hammering IDE disks very hard 24
> hours/day. These would often fail three or four times during the 3 year
> warranty period. I suggested he switched to SCSI, which he did and has not
> looked back. They are built much better, and have more intelligent drive
> electronics that minimises head movement. Of course, the downside is that
> they are 2 or 3 times more expensive for the same capacity.
>
> --
> David Kirkby,
> G8WRB
>
> Please check out http://www.g8wrb.org/
> of if you live in Essex http://www.southminster-branch-line.org.uk/
>
>
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
>
--
David Kirkby,
G8WRB
Please check out http://www.g8wrb.org/
of if you live in Essex http://www.southminster-branch-line.org.uk/
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|