18 Sep
2006
18 Sep
'06
10:42 p.m.
On Mon, 18 Sep 2006 13:58:47 -0500, John Foust <jfoust at threedee.com>
wrote:
At 02:35 PM 9/16/2006, CRC wrote:
My intent was to address classic computer equipment. MTBFs of Laser
Diodes (LDs - not LEDs) has definitely increased over the last 10
years. However, access to information of the ones used in current CD
equipment is privy only to large scale manufacturers. Looking at LDs
that are used in fiber optic transmission shows a substantial
increase in MTBF over the past 10 years. However, I would hold off on
buying into the BlueRay recorders - blue LDs are still have low MTBFs.
CD equipment production has progressed to commodity class and every
penny represents a margin change of $10k / 1M units - design is done
by taking out things until the unit doesn't work :=)). Adding the
circuitry to turn off the laser is not difficult, but it costs money...
The basic problem with commercial CD
burners/players is that they
keep the laser diodes on when powered, although at the low power
required for reading. The life of run-of-the-mill laser diodes is on
the order of 10,000 (10.000) hours. Consequently, if you keep your
system on all the time you can expect one to two years of useful life
from the beast (some longer, some shorter). From <http://
www.wtec.org/ loyola/opto/ad_rohm.htm>:
That document dates from 1994. You'd think that manufacturers would
want to increase MTBF and eliminate failures as quickly as possible,
so I find it hard to believe that LEDs are left on inside today's
CD/DVDs just because it's hard to turn them off and they don't want
to improve lifetimes. This info may be entirely relevant for
1980s drives, who knows? What sorts of other ancedotes and
claimed facts do we have?
From the CD-R FAQ: http://www.cdrfaq.org/faq05.html
Subject: [5-2] How long do CD recorders last? (1998/04/06)
The MTBF (Mean Time Between Failures) on these drives is
typically 50,000
to 100,000 hours, and they come with a 1 year warranty. Compare
that to
hard drives rated at between 500,000 and 1,000,000 hours with a 3
or 5
year warranty and that should give you some idea. Most of the drives
available today weren't meant for mass production of CD-Rs. The only
exceptions are the venerable Philips CDD 522, Kodak PCD 600, and
Sony CDW-900E.
I assume that the FAQ uses manufacturers stated MTBF - there is no
citation for their source.
By 2004, MTBF for consumer CD/DVD were up to 60K to
100K hours:
http://www.computingondemand.com/reviews/storage-LiteOn-LDW-411S/
page1.shtml
And perhaps their MTBF was calculated with a 2% duty cycle, which
would still mean they expect a consumer DVD-R to be able to burn
several thousand discs. I suspect consumer losses are due more to
dust and crud on discs... unless of course you simply don't want to
trust manufacturer-provided MTBFs.
In most cases, the criteria for the MTBF is not stated as indicated
by your suposition. One has to look at MTBF at temperature since
Optek shows a factor of greater than 25 in MTBF between 25 and 70 C.
If the tests were run at 20C and you run the device at 50C you can
expect 6K and 9K hours vs 70K and 100K hours respectively (based on
ratios from the Optek MTBF data for their OPV200 series high
reliability LDs). At 70C it's around 1.5K and 2K hours. It's not
that I don't trust the manufacturer, it's just that there is
insufficient info and the manufacturer's shills will present their
most favorable data.
Besides the mechanical/electrical failures, which there are many, the
LDs power loss generally results in the inability to record and read
reliably. Note that the LD does not touch the disk and unless it is
used in a smoky/cruddy atmosphere there should be no failure in the
optical link. In trying to repair a good number of CDR/W I have
recovered only several units by cleaning the LD lens. Recording
generally uses virgin disks and an occasional brushing of the lens
does help.
Your average $50 Wal-Mart DVD drive has a 70K MTBF:
http://www.liteon.com/prod/getProduct.do?
cid=1_7_13&xml_id=4_2&menu_id=4_2_7
Note that the stated MTBF is in POH - power on hours - not usage
hours. I guess if you use it all bets are off :=P
CRC