Short life of Hard Discs

[Bob Grime]

A bit of general information.
In the early days of my running a Tivo box, I had a problem with failing hard discs. The first two failed in under two years each. I investigated and found that the discs were running hot (case temperature over 50C) I fitted my discs with heat sinks obained from quietpc.com. and stripped back an area of sound proofing above the disc bringing the temperature down to approx 40c. I have had no disc failiors since. One disc is 5 to 6 years old and the other is 2 1/2 years old.
Sun 31/7. The same problem occured in two different Tivo's. So any problems like fans not working is unlikely. The temperature was measured at the hard disc case usng stick on temperature sensors not the temperature displayed in the system information. This is purely my own personal experience which I thought might be of interest.

[SolidTechie]

50 degrees sounds very high - mine run about 37-38 on average - is your TiVo in a closed cabinet?

I have a pair of 120's in my TiVo - been running fine since 2004

[Rachel]

... and is the fan working?

[Pete77]

I upgraded my original Tivo acquired in December 2002 with two Samsung HA250JC 250GB drives in July 2005 and six years later they are both still running. They have been in 24/7 use throughout this time.

These drives were specifically designed as PVR models though and only have a 5400rpm spin speed. They run at a Tivo reported internal temperature of a between 37C in the winter through to about 44C on the hottest possible summer day.

[Furbag]

Bit light on disk life arnt we

Running a Maxtor 120gig bought from Novatech (Product code MAX-P9120) on 11-11-03 for £76.10, yeah got the receipt in my hand now

Switched on about a week later and then runs 24/7 via a ups with only a shut down once a year for a TiVo internal de-dust

Of course now I've said this it will pack up tomorrow

Furbag

[Pete77]

I'm beginning to think that the likely very short life time of hard drives in a Tivo was a vicious rumour carefully circulated by the myopic citrus to encourage us all to upgrade our hard drives ASAP.

He used to tell us how Tivos positively ate hard drives for breakfast and also how the power supply was bound to need replacing after only two or three years (causing all spare Tivo power supplies in the world to soon be gobbled up and trade at a premium price), Strangely enough six years after I replaced them the replacement hard drives in my Tivo are still running without signs of any problem and I bet the power supply I took out of the Tivo and put in the drawer when I replaced it at only four years old would run fine too if I ever needed to put it back. In all these years of running numerous electrical appliances I have only had one power supply ever actually fail and fortunately that was an external brick so was easy to replace.

[Tcm2007]

The TiVo power supply doesn't need to fully fail to cause problems. Lots of issues, especially around sound, seem to be related to the supply being on the margins of being able to deliver enough juice to power disks, network cards and the like. As it gets older, it's capacity seems to slightly reduce.

I had to change the supply on one of mine because of intermittent sound problems; changing the supply cured them.

You can't draw any conclusions about hard disk like for the sample size any one of us has running in our homes, as any statistician will tell you.

[Pete77]

You can't draw any conclusions about hard disk like for the sample size any one of us has running in our homes, as any statistician will tell you.

You can't deduce anything from just one owner. A sample size of 20 or 30 owners in relation to the number of Tivos sold would actually be quite reasonable as long as they were selected at random.

[Tcm2007]

That would still be too small a sample, I suspect. To have 95% confidence of the mean time to failure of the factory shipped TiVo disks from a pool of 30,000 would need over 100 I would estimate. A statistician could do the calculation to be sure.

[Pete77]

That would still be too small a sample, I suspect. To have 95% confidence of the mean time to failure of the factory shipped TiVo disks from a pool of 30,000 would need over 100 I would estimate. A statistician could do the calculation to be sure.

That would equate to sampling one in 300 all Tivo owners.

That is way, way higher than the sample frequency of one in 60,000 used in most General Election Opinion Polls (which have a sample size of 1,000 people out of a population of 60,000,000) not to mention the fact that out of the original 30,000 Tivos manufactured only a maximum of about 5,000 are probably still in some kind of use. On that basis my proposed sample size ought to be more than accurate since hard disk drives will tend to faily on a very uniform although random basis across the Tivo using population as a whole. The main difference will be between the failure rate of the original hard drives and the replacement Tivo hard drives.

[Furbag]

I'm beginning to think that the likely very short life time of hard drives in a Tivo was a vicious rumour carefully circulated by the myopic citrus to encourage us all to upgrade our hard drives ASAP.

He used to tell us how Tivos positively ate hard drives for breakfast and also how the power supply was bound to need replacing after only two or three years (causing all spare Tivo power supplies in the world to soon be gobbled up and trade at a premium price), Strangely enough six years after I replaced them the replacement hard drives in my Tivo are still running without signs of any problem and I bet the power supply I took out of the Tivo and put in the drawer when I replaced it at only four years old would run fine too if I ever needed to put it back. In all these years of running numerous electrical appliances I have only had one power supply ever actually fail and fortunately that was an external brick so was easy to replace.

I dunno, the reason I swapped both mine and and my mums drives was due to failure (picture breakup and freezing) and even a new TiVoland drive Dave supplied died within a few months, Dave replaced it double quick and since then hers has been as good as gold and mine is equally fine.

PSU's can be funny old things (yes I do have a spare ) it does depend on the spec of the components used, I've had Swtiched mode stuff give up within 12mths in fact one TV we owned had 3 PSU's in as many years, a PC PSU went pop in a very spectacular way as well so they do fail.
One of the main issues was the bad capacitor issue (google bad caps) this problem has been the bain of many electronics for years sadly and I have replaced caps on many bits of electronics due to their failure when in fact most caps if rated correctly should last for decades or even centurys

Furbag

[sstman]

I had two HDD failiers on my TiVo unit, but no problems with the PSU, how ever this thread reminded me of a conversation I had with an old friend, who used to work with a 'large well known supplier' of satellite set top receivers that were made in Wales, the 'design life' of the switched mode PSU in these receivers, was quoted as 14/16 months!!. the fix was a simple 'upgrade' of capacitors, but the boxes were built to a price.....and were 'lifed' at 14/16 months, so at least they lasted for the 12 month warrenty.

[Pete77]

I had two HDD failiers on my TiVo unit, but no problems with the PSU, how ever this thread reminded me of a conversation I had with an old friend, who used to work with a 'large well known supplier' of satellite set top receivers that were made in Wales, the 'design life' of the switched mode PSU in these receivers, was quoted as 14/16 months!!. the fix was a simple 'upgrade' of capacitors, but the boxes were built to a price.....and were 'lifed' at 14/16 months, so at least they lasted for the 12 month warrenty.

Were they Pace Sky satellite boxes by any chance?

My Pansonic TU-DSB20 Sky box is at least 13 years old and still going strong running the latest Sky firmware.

[Tcm2007]

That would still be too small a sample, I suspect. To have 95% confidence of the mean time to failure of the factory shipped TiVo disks from a pool of 30,000 would need over 100 I would estimate. A statistician could do the calculation to be sure.

That would equate to sampling one in 300 all Tivo owners.

That is way, way higher than the sample frequency of one in 60,000 used in most General Election Opinion Polls (which have a sample size of 1,000 people out of a population of 60,000,000) not to mention the fact that out of the original 30,000 Tivos manufactured only a maximum of about 5,000 are probably still in some kind of use. On that basis my proposed sample size ought to be more than accurate since hard disk drives will tend to faily on a very uniform although random basis across the Tivo using population as a whole. The main difference will be between the failure rate of the original hard drives and the replacement Tivo hard drives.

I'm sure an intelligent man like yourself knows that the number required to get statistical significance from a sample is not a simple linear relationship to population size.

To be valid of course you'd have to sample the TiVos no longer in use as well, as otherwise you'd be excluding a whole load of devices who probably went out of use because there drives failed. And you'd exclude replacement drives as that would be measuring an entirely different brand of disk, rendering the research pointless.

[Pete77]

And you'd exclude replacement drives as that would be measuring an entirely different brand of disk, rendering the research pointless.

Not if one established the brand and model of the replacement drive and also ran some statistics on how well they had lasted.

Of course I know that would clearly not be acceptable to you since you repeatedly made it quite clear that you think that any Tivo owner ought to have moved on to a more modern system like MCE or Sky HD at the point at which their hard drive expired and indeed your view appears to be (as admittedly extrapolated by myself from your various other comments) that they were stealing Lifetime service from that point on by not doing so.

Also I am sure that a hard bitten commercial man like you would know that the research questions you ask will generally be determined by the answers you want to produce for marketing purposes in most commercially driven market research environments.