Tivo/DVRs - HDDs more prone to failure than PC Hard Drives?

On 2006-10-05, wrote:
This is a great question.

I have two Series 2 Tivos. One was purchased in Dec 2003, one in Dec 2004.

The hard drive in the Dec 03 Tivo failed in October 2005. (22 months)
The hard drive in the Dec 04 Tivo failed in September 2006 (20 months).

I am currently operating computers with a total of 6 internal (IDE) hard
drives and 2 external hard drives. The oldest of these is 55 months.
Zero failures among any of them.

Both Tivos are in well ventilated areas, but I suspect an internal
temperature cause rather than ambient cooling.

From my data set: Tivo hard drives are much more failure prone than the
ones in PCs.

I have a laptop that has been running hot for over 2 years at this
point. EVERYTHING in that system should be par boiled. Yet it all keeps
chugging along. I can't imagine any Tivo drive having it tougher.



--

Nothing today, likely nothing since we tamed fire,
is genuinely new: culture, like science and |||
technology grows by accretion, each new creator / | \
building on the works of those that came before.

Judge Alex Kozinski
US Court of Appeals
9th Circuit

wrote:
This is a great question.

I have two Series 2 Tivos. One was purchased in Dec 2003, one in Dec 2004.

The hard drive in the Dec 03 Tivo failed in October 2005. (22 months)
The hard drive in the Dec 04 Tivo failed in September 2006 (20 months).

I am currently operating computers with a total of 6 internal (IDE) hard
drives and 2 external hard drives. The oldest of these is 55 months.
Zero failures among any of them.

Both Tivos are in well ventilated areas, but I suspect an internal
temperature cause rather than ambient cooling.

Quite possible, though I believe S2's are typically considered to be
adequately cooled unless blocked. Some folks add additional cooling
fans in or near the case.

From my data set: Tivo hard drives are much more failure prone than the
ones in PCs.

But your dataset is much too small to be statistically meaningful. You
have a dataset of 10 total drives (2 Tivo, 6 internal, and 2 external)
of widely varying manufacturers, models, interfaces and environments.
It's just not enough to draw any meaningful conclusions from. Others
have presented similar stories with the exact opposite results.

If we could gather data from *hundreds* of Tivo drives and their failure
rates then we could compare that and possibly arrive at some
statistically significant results. I'd say we could compare them to the
published MTBF's, but that would only be valid in a relative context as
the environments are certainly different.

Randy S.

Randy S. wrote:
wrote:
This is a great question.

I have two Series 2 Tivos. One was purchased in Dec 2003, one in Dec
2004.

The hard drive in the Dec 03 Tivo failed in October 2005. (22 months)
The hard drive in the Dec 04 Tivo failed in September 2006 (20 months).

I am currently operating computers with a total of 6 internal (IDE)
hard drives and 2 external hard drives. The oldest of these is 55
months. Zero failures among any of them.

Both Tivos are in well ventilated areas, but I suspect an internal
temperature cause rather than ambient cooling.

Quite possible, though I believe S2's are typically considered to be
adequately cooled unless blocked. Some folks add additional cooling
fans in or near the case.

From my data set: Tivo hard drives are much more failure prone than
the ones in PCs.

But your dataset is much too small to be statistically meaningful. You
have a dataset of 10 total drives (2 Tivo, 6 internal, and 2 external)
of widely varying manufacturers, models, interfaces and environments.
It's just not enough to draw any meaningful conclusions from. Others
have presented similar stories with the exact opposite results.

If we could gather data from *hundreds* of Tivo drives and their failure
rates then we could compare that and possibly arrive at some
statistically significant results. I'd say we could compare them to the
published MTBF's, but that would only be valid in a relative context as
the environments are certainly different.

Randy S.


I didn't claim statistical signficance...just relaying my experience.

The drives in other-than-Tivo enclosures (desktops, laptop and external
USB enclosures) are failing at a zero percent rate over ages of two to
almost six years. The drives in Tivo enclosures have failed at a one
hundred percent rate in less than two years.

Tivo and the desktops are all the same IDE interface (7 of the 10).

This not quite statistically significant sample has tough me this:
even though the non-Tivo drives might fail...they haven't yet.

Even though the Tivo drives might not fail....they already have.

I build my own PCs, but if any vendor (of a whole system or the hard
drive itself) had sold me two consective units that failed withing two
years...I wouldn't return to that well.

Both Tivos came are on lifeterm service contracts....so I'm a happy,
repeat Weaknees customer. I have another year to go before I can draw
any conclusions about the quality of their drives....

I build my own PCs, but if any vendor (of a whole system or the hard
drive itself) had sold me two consective units that failed withing two
years...I wouldn't return to that well.

That's a significant difference right there. I can virtually guarantee
that you were willing to spend an extra $10 per drive to select a
well-reviewed drive with a good warranty. I'm sure Tivo selects drives
on cost as the overriding factor (as any large volume manufacturer
will). $10 isn't that much to you or me, but multiplied by 100,000 or
so it starts making a dent on their balance sheet. I bet you if you
compared home-built machines to any large manufacturer (Dell, Sony, IBM)
you'd find the home-builts have a better reliability.

Randy S.

Randy S. wrote:
I build my own PCs, but if any vendor (of a whole system or the hard
drive itself) had sold me two consective units that failed withing two
years...I wouldn't return to that well.

That's a significant difference right there. I can virtually guarantee
that you were willing to spend an extra $10 per drive to select a
well-reviewed drive with a good warranty. I'm sure Tivo selects drives
on cost as the overriding factor (as any large volume manufacturer
will). $10 isn't that much to you or me, but multiplied by 100,000 or
so it starts making a dent on their balance sheet. I bet you if you
compared home-built machines to any large manufacturer (Dell, Sony, IBM)
you'd find the home-builts have a better reliability.

Randy S.


Your assumption is correct. WD/Seagate with 3 or 5 years (variously)

I'm assuming the that Weaknees replacement drives are higher quality
than Tivo OEM.

If either of those Tivos's were without have lifetime
subsrciptions..they'd be in the recyling heap right now.

I think it was 32 months of operation to make the lifetime subcription a
better deal than monthly....plus the extra months for the replacements:-)

I enjoy the Tivo tv watching experience. But I've bought my last product
from them (especially for the price of the Series 3 HD capable machines)

wrote:
I bet you if you
compared home-built machines to any large manufacturer (Dell, Sony, IBM)
you'd find the home-builts have a better reliability.

Randy S.


Your assumption is correct. WD/Seagate with 3 or 5 years (variously)

What makes you say that? Do you think home builders have quality
control engineers and perform thermal imaging and resonant vibration
testing on a prototype unit to assure there are no problems? Don't
confuse warrantee period with reliability, the former has a large
marketing gimmick to it. It is true that you can purchase parts that
are for the most part higher quality, but that is only part of the
picture, environmentals (power, mechanical & cooling) CAN matter.

As reported by my S2 system settings, my Tivo is running at 40 degrees
C (104F). I don't know where this is, but 40C is definetely hot for a
hard drive, a big manufacturere would take corrective action on this
(even if it was not 24/7 like a Tivo).

On 2006-10-06, wrote:

wrote:
I bet you if you
compared home-built machines to any large manufacturer (Dell, Sony, IBM)
you'd find the home-builts have a better reliability.

Randy S.


Your assumption is correct. WD/Seagate with 3 or 5 years (variously)

What makes you say that? Do you think home builders have quality
control engineers and perform thermal imaging and resonant vibration
testing on a prototype unit to assure there are no problems? Don't

Sure they do. They're all the other consumers. Assuming
you're not the first on your block with a particular piece of
gear you will have ample data. Even if you just avoid notorious
brands or vendors you'll come out way ahead.

confuse warrantee period with reliability, the former has a large
marketing gimmick to it. It is true that you can purchase parts that

OTOH, Seagate has a long and well established reputation
as server systems storage provider. Also, the longer warranties are
not so much a marketing gimmick as pushback against a recent trend to
reduce the stated warranty periods on consumer drives.

3 years used to be the standard.

are for the most part higher quality, but that is only part of the
picture, environmentals (power, mechanical & cooling) CAN matter.

As reported by my S2 system settings, my Tivo is running at 40 degrees
C (104F). I don't know where this is, but 40C is definetely hot for a
hard drive, a big manufacturere would take corrective action on this
(even if it was not 24/7 like a Tivo).

--
....as if the ability to run Cubase ever made or broke a platform.
|||
/ | \

wrote:
wrote:
I bet you if you
compared home-built machines to any large manufacturer (Dell, Sony, IBM)
you'd find the home-builts have a better reliability.

Randy S.

Your assumption is correct. WD/Seagate with 3 or 5 years (variously)

What makes you say that? Do you think home builders have quality
control engineers and perform thermal imaging and resonant vibration
testing on a prototype unit to assure there are no problems?

No, but they can take the time to look at past experiences on well known
models, avoid models known to be problematic, and in general value
reliability over cost. Most large manufacturers value cost over
reliability because they believe they'll save more money that way (your
inconvenience doesn't cost them anything, and unhappy customers are hard
to measure in dollars).

Don't
confuse warrantee period with reliability, the former has a large
marketing gimmick to it.

Actually, if you look, you will notice that *specifically* did not
mention warranty length as a factor because I agree with your statement
totally.

It is true that you can purchase parts that
are for the most part higher quality, but that is only part of the
picture, environmentals (power, mechanical & cooling) CAN matter.

Absolutely, and that was my first statement above. I suspect heat is
much more likely to be the culprit here, if it *is* statistically
significant.


As reported by my S2 system settings, my Tivo is running at 40 degrees
C (104F). I don't know where this is, but 40C is definetely hot for a
hard drive, a big manufacturere would take corrective action on this
(even if it was not 24/7 like a Tivo).

That does seem hot. Are your vents clear of dust and unobstructed?

I don't believe the fact that it runs 24/7 is a factor, these drives are
designed to stand that. Many people believe that turning them off and
on actually wears them out more due to thermal cycling (also bearing
wear is significantly greater at startup and shutdown).

Randy S.

JEDIDIAH wrote:
What makes you say that? Do you think home builders have quality
control engineers and perform thermal imaging and resonant vibration
testing on a prototype unit to assure there are no problems?

Sure they do. They're all the other consumers. Assuming
you're not the first on your block with a particular piece of
gear you will have ample data.

After a few months you might be able to find blatenly crappy individual
components being identified, but thats about it. The subtle things
that make an ultra relable machine are well beyond that of even a
highly skilled homebuilder. You can stack things in your favor by
spending a lot more, but you will never get a better price performance
point then Dell, HP, Lenova.

Randy S. wrote:
wrote:
wrote:
I bet you if you
No, but they can take the time to look at past experiences on well known
models, avoid models known to be problematic,

Yes, but thats 20/20 hindsight.

Most large manufacturers value cost over
reliability because they believe they'll save more money that way (your
inconvenience doesn't cost them anything, and unhappy customers are hard
to measure in dollars).
Well I somewhat agree with what you're saying. Manufacturers HATE
spending warrentee dollars. They will never risk a big warrantee bill
to save a small amount. Long warrantees on components are easy to
weasle out of, since most consumers will not be able to prove date of
purchase.


That does seem hot. Are your vents clear of dust and unobstructed?

Next to 40C is reported 'Normal', it does sit in an entertainment
center, but the local ambient at box inlet is still typically 25C. I
think this is just the way it is. Its probably worth researching where
the sensor is, 40C for the board or components on it is not a problem.
Better yet would be to crack it open and thermalcouple the drive and
try to correlate its temp to the on board sensor, but I'm just not that
motivated....

I don't believe the fact that it runs 24/7 is a factor, these drives are
designed to stand that. Many people believe that turning them off and
on actually wears them out more due to thermal cycling (also bearing
wear is significantly greater at startup and shutdown).


24/7 becomes a bigger factor when the temperature is high. Once the
grease begins to dry out, you're toast.