A while back, a colleague in the IT profession reached out to me asking if I knew what the failure rate was these days for laptop and desktop computers. I hadn’t considered this question of computer hardware reliability for some years since I had read this study by the Gartner Group, so I imagined things must have continued to get better since that publication was released. But instead of looking for a more recent study on the subject, I decided to ask around from others in the IT profession and hear what their thoughts were regarding laptop and PC failure rates and what’s happening with the reliability of different components.
Cheap is cheap
Mark, the operations manager for a company in British Columbia, Canada, that provides on-site computer repairs, maintenance, and consulting offered up the following reflections:
“There are lots of variables affecting laptop and PC failure rates. In my experience, I have found that “consumer-grade” equipment, cheap, low-end systems, both desktop and laptop, have a significant failure rate. But where do you start? Most systems have a hardware failure rate of 0.5 percent-3 percent DOA (dead on arrival). The less expensive systems having the higher rate, higher quality/price systems the lower failure rates. Once a system gets past 30 days (which is including the DOA category) most will get to one year with a hardware failure rate of around 2 percent-5 percent. So the less expensive systems have the higher failure rate, and higher quality/price systems the lower failure rates.
Past one year service, however, I find dramatic differences. Expensive, higher-end systems seem to have a hardware failure rate hovering around 5 percent-8 percent in the second year. Our favorite “El cheapo” systems, however, start dropping like flies once they reach 18 months in service. Forty percent-50 of “cheap systems” fail in the second year of use. A cynical person would think the manufacturers are designing cheap systems to fail once they get past the warranty period. Of course, all manufacturers have their client’s best interests at heart and design all systems for long life!
Cheap systems usually do not get past four years’ service unless they are rarely used and powered down most of the time. High-end systems easily reach six or more years’ service with one caveat: hard drives. Hard drives start failing after two years and increasing the failure rate for each year after two years. I recommend to all business clients that they change their hard drives after three years.”
Mark emphasized to me that his statistics were all anecdotal in nature, but when I ran his comments by some other colleagues who work in similar professions they basically agreed they were in the ballpark. In other words, cheap is cheap. You get what you pay for.
Laptop and PC failure rates: What’s really the problem?
Still, this didn’t really explain laptop and PC failure rates and why so many tended to fail after only a short period of time. Was it a problem with a specific type of electronic component? Or was it because of the hardware vendor cutting corners in some fashion? A colleague named Dave offered these thoughts:
“Laptop and PC failure rates also depend a lot on the components themselves. About 15 years ago there was a big problem with capacitors having bad mixtures, which caused air. Fifty percent of the desktops in 2004 had to be replaced due to sudden failure from these capacitors. Then we have SSDs. Most SSDs that we had used from 2010-2011 have already died. SSDs now are more stable, but all but one SSD from them died completely. The only SSD that did not die was an Intel.
Something interesting, though, is also the misdiagnostic of failure rate. For example, in desktops, the report says motherboards and hard drives, but I am doing an educated guess that more than half of those motherboards could actually be power supplies. A bad (cheap) power supply will kill other components on the computer, but might not kill the PSU itself. Case in point, I had several Dell computers where the motherboard was replaced several times by Dell, but when I decided to start changing the PSU the motherboards stopped dying. Which makes me wonder how many replacements were unnecessary. To this case in point, one XPS system had three motherboard replacements because it had a Nvidia 680L chipset (famous for having a high failure rate). On the fourth fail, I told the technician to replace the PSU, which he refused, so I “Frankensteined” the system with a PSU outside the system (XPS has a proprietary format). This specific system has been running four years with said “bad” motherboard and a new PSU. Since then I have replaced the case so I could fit everything inside the case.”
Dave’s comments about cheap PSUs dorking other components rang a bell with me as I’ve had several PCs (and one server) die in our business due to faulty power supplies. I asked him if he thought this might be a major contributing factor toward computer hardware failure nowadays. Dave replied:
“Power supply is where I see most vendors cutting corners, and this causes most of the problems. Tom’s Hardware did a report on the manufacturing process of PSU, and what causes a “cheap” PSU to fail more. A PSU is like the tire/brakes in the car. If you have a cheap one, don’t be surprised when the car does not stop on time. My own personal computer also had a fried motherboard, RAM, and SSD, but this was due to a bad 12-volt rail. I got all components replaced under warranty. In this case, though, the PSU was neither cheap nor underpowered; the rail just went bad, so Antec even covered the SSD, which was not under warranty anymore.
The other thing that gets me wondering from the report is to separate laptop and PC failure rates from user failure rate. Meaning, cracked screens are usually not a failure on the component, unless a notebook has a flawed design that put pressure on it. I own a Surface 3, and it has a yellow tint on the side. That is for me a failure on the component (because it is decaying, with normal use), but if I dropped the tablet, and broke the screen I would not consider it a failure. Sure, we would like better design to avoid cracked screens from a drop, but that is an inherent problem with portable devices.”
Dave correctly highlights the point that not all system failures are hardware-related. Some PC and laptop failures are simply due to bad behaviors on the part of the user. Unfortunately, while it’s possible to build better hardware components, it doesn’t seem possible to build better users!
Things are not so bad really
Not everyone I talked with felt that computers were becoming more prone to failure. Clare, an IT pro who provisions systems for businesses of various sizes, shared the following observations:
“It is interesting to see people are experiencing such high laptop and PC failure rates on quality machines. I install only Acer Veriton systems for my business clients. One office has 40 PCs and three servers in-house (two Acer and one HP) the other has about 50 PCs and two servers. The servers are new and about 17 PCs are brand new Acers, the rest an assortment of Del Vostro, MDG, local clones, and about 33 old Lenovos.
In the first office (a large insurance brokerage) we try to keep the computers under 7 years old, and I have had a total of three hard drive failures, and four or five power supplies fail over the last four years. I currently have one failing motherboard (about 6 popped caps) on a computer that is overdue for replacement (installed July 2007) The servers are five and six years old and I have replaced four out of six hard drives (in Raid 5 set). None of the PC failures have been within the first two years, and two hard drives have been covered by warranty. The server drives were all under warranty.
At the second customer (a manufacturing facility), we replaced dumb terminals in the plant with used Lenovo PCs almost five years ago. Most were 486 XP systems purchased used from a discounter and were four or five years old when purchased ($99 each with keyboard, mouse, and XP Pro) We are now seeing a high failure rate on the 486 machines, with capacitors popping left, right, and center. The Pentium Dual Core units we bought on the second round (to complete the changeover and for spares) have been standing up quite well. We had some DOAs, but otherwise, they’ve been pretty good. These systems are 9 years old and older. The non-Acer PCs were all installed before I started looking after them — most were less than 3 years old so are approaching five to eight years of age. About half a dozen or so have failed totally in the last 18 months, and many of the others have slowed down significantly (some due to bad caps) — and switching to Windows 7 from XP has initiated the changeover of the rest of the systems (mostly Pentium Dual Core and Core2 Duo processors). I’ve changed a few hard drives and power supplies on these older machines over the last four years or so. One server was DOA.
Of my smaller clients, I have had one DOA, about two power supplies and two or three hard drives over the last nine years on the Acers, and yes, some of the first Acers I sold are still in service. Previous to selling Acers, I sold locally built higher-quality clones from two suppliers with spotty results (and no consistency in product).
Since the defective capacitors have worked their way out of the supply chain, the PCs are pretty reliable — particularly “professional-quality” brand name (Acer in particular) systems — and I won’t sell anything else.”
I asked a few more colleagues their thoughts on laptop and PC failure rates. Erik, an IT Manager based in Manhattan, gave these observations:
“I’ve been using Dell PC’s for the past 10 years and I would say my failure rate for the last three years is averaging 5 to 10 percent per year on 75 computers. Before that, it was close to zero. Laptop and PC failure rates are often high mostly because of bad hard drives, though I have a bunch of Optiplex 990s that fail repeatedly. Otherwise, it is all dead power supplies and cooling fans.”
Erik’s comments about clogged fans causing PC failures reminded me how important it is to keep dust and other particulates away from computer systems, something I wrote about previously on TechGenix here and here. Another fellow IT pro named Tom works with systems from both Dell and HP and he offered these comments:
“From my experience: Desktops — five to eight years, although usually around the five-year point the desktop is too slow for newer versions of software. Laptops — when used as a desktop replacement, three to four years. When used by mobile users, usually two to three years under normal wear and tear. Of course, this can depend on brands. These are mainly for Dell and HP computers.”
What about you?
What’s your own experience and reasoning concerning laptop and PC failure rates? Are they getting more reliable or less so? Share your comments!
Photo credit: Shutterstock