Computer longevity -- is no protection better than cheap protection?

[Winterpool]

First, my apologies if this is the wrong forum. I opened here because 'General Hardware' gets much more traffic than 'Power Supplies', and it's a topic beyond PSUs per se. Translocate if necessary...

Black Friday often brings deals on cheap UPS, which reawakens in me every year 1.) the desire to purchase a unit and 2.) the anxiety that these things don't actually protect your kit. In fact they may make matters worse. And every year, I read primers on the Intarweb and take away the conclusion that nothing I do will really make a difference, because

1. Cheap UPS produce 'dirty' power: the infamous square wave rather than a true curving sine wave.

2. Cheap UPS won't switch to battery in time (n millisec will pass between power failure and battery coming on line). This can be resolved by permanently online UPS batteries, but then you're always running on less smooth power.

3. UPS does not protect against power surge (but why are there 'surge protected' outlets inside some UPS that are not connected to the battery?). Does 'AVR' make any difference here against minor variations? Don't some UPS claim to come with integral surge protection?

On a related issue, consumer surge protectors aren't anything more than glorified power strips because

1. They too can produce 'dirty' power.

2. 'Clamping' voltage and reaction time may be inadequate to switch off power before damage in event of surge.

3. Design of consumer surge protector insufficient to protect against truly massive surge -- the lightning strike in one's neighbourhood (as opposed to temporary power company variation). If your house isn't grounded properly, all your kit is burnt toast anyhow.

I've read some threads that devolve into name-calling and comparing the length and girth of one's EE degree. The punters (like Westom) who cite the most science often seem to propound a sort of destructive Pyrrhonism that suggests every consumer UPS and surge protector is useless.

My questions is: if my goal is to maximise the survival and longevity of my computer kit, not necessarily maximise continuous uptime, is it better to run with 1.) nothing at all, 2.) simply a good surge protector, or 3.) a surge protector and UPS in some sort of series? And does 'good' exist at a price below hundreds of dollars?

What if electrical power in my location is unreliable (drops and 'brownouts')? Would 'dirty' UPS power be preferable to the likely power failure / restart that drops would trigger?

If uptime does come into consideration, have I just consigned the UPS-plugged computer to abbreviated life expectancy?

 

[ThatsABigOne]

Those are some interesting points. I am going to keep watch over this thread in hoping to get some answers.

 

[AnonymouseUser]

You are thinking way too hard about this. This is what I think about UPS's, and here is another opinion. Get one for each PC and any other expensive electronics.

 

[Winterpool]

Thinking 'way too hard'? I wonder if you're right. Westom brings in lots of thought and detail to seemingly conclude abandon hope all ye who ponder this!

 

[Sir_Loin]

I think they're worth the expense. With as often as power is interrupted in my area, my UPSs save my computers from the ever present threat of a hard shutdown....and keeps my farm running happy. I even have a small one for my cable modem and router.

The trick is finding ones that use the same style 12v batteries that you can find at Academy or Gander Mtn for ~$20, as opposed to what the maker of the UPS wants to charge you for their replacement batteries.

 

[AnonymouseUser]

Thinking 'way too hard'? I wonder if you're right. Westom brings in lots of thought and detail to seemingly conclude abandon hope all ye who ponder this!

Way too much thought. Even if a UPS doesn't protect against lightning strikes (though it sure seems like it protected some of us ), it will protect against sudden power loss, which can be just as fatal (especially concerning data loss). Also, even if they provide "dirty" power, at least it's consistent, and your PSU will handle it just fine.

 

[mfenn]

Oh boy, here we go again. This is likely to be my only post in this thread due to certain member's tendency to raise my blood pressure.

Basically there are 3 types of UPSes:
1. Off-line, or standby: The AC input normally just goes straight through to a glorified power strip as well as to charge the battery. When input power is lost, the UPS quickly switches over to battery which powers an inverter that provides output AC. These are useless IMHO and you might as well get a power strip.
2. Line-interactive: Similar to an off-line UPS except that the input is connected to an autotransformer that can clean up low/high input voltage. These are also pretty much useless unless you live in an area susceptible to brownouts.
3. On-line: The AC input always drives the battery charger and inverter which then provide the output AC. Essentially, AC is always converted to DC then back to AC. These are the only true Uninterruptible Power Supplies in my opinion. The first two are best classified as Standby Power Supplies (SPS). Definitely worth it as the power strip is completely isolated from the mains.

 

[corkyg]

My system runs 24/7. I would not do that without a good UPS. As for lightning, we got a direct hit about7 years ago. Nothing can protect from that. Forunately, we were out of town and all computers were off and disconnected. But, I had to replace all of my security alarm panel and system. Insurance covered most of it.

 

[Winterpool]

Jah, generally, if I know an electrical storm is en route, I disconnect vulnerable kit from the mains.

As to UPS, thanks, mfenn, for clearing up 'line-interactive' v 'online'; I had thought they might be the same.

As to offline/standby, this is the category most often cited as useless; however, some punters have reported their machines remaining powered up with this sort of UPS.

Also: I believe I've read elsewhere that manufacturers discourage connecting UPS and surge protector in series (principally to cover their arse), but if one does elect to do this, it's better to connect surge to mains, then UPS to the surge, and finally your machine to the UPS?

 

[RebateMonger]

It's hard to conceive of a UPS that's so bad that a PC would be better without it. Well, I'm sure they exist, but you can get APC or CyberPower units starting at $40 or so.

Sine-wave UPSes are pretty pricey. Modern PC power supplies don't really care. Sine-wave units are good for non-PC devices that CAN be sensitive.

All UPSes contain some sort of surge protection. No, they usually aren't rated for as much a surge as pure surge protectors.

Most of us can't afford a sine-wave UPS (starting at $100+), nor an online UPS (starting at around $400), so the discussion is pretty much academic.

 

[westom]

I've read some threads that devolve into name-calling and comparing the length and girth of one's EE degree. The punters (like Westom) who cite the most science often seem to propound a sort of destructive Pyrrhonism that suggests every consumer UPS and surge protector is useless.

First, most failures are traceable to manufacturing defects. One popular example was so many electrolytics that failed due to counterfeit electrolyte. So many see failure, do not do the autopsy, then know what must have caused that failure using only wild speculation. But hard reality from too many generations of experience: most failures are traceable to manufacturing defects.

Very little can avert that failure other than simple techniques that can sometimes identify a defect before the warranty expires or failure happens. A simplest tool is heat. Defects that will not appear maybe months later can sometimes be located by operating electronics in what is a normal computer environment: 100 degrees F. Diagnostics executed at what is normal but extreme conditions can identify manufacturing defects long before failures occur.

Second, another source of failure are transients that overwhelm existing and internal protection. Anything that would work adjacent to an appliance is already inside the appliance. Transients that can overwhelm that internal protection must be diverted to and absorbed harmlessly outside the building. Only then will rare transients (typically once every seven years) not overwhelm existing protection.

None of this is complex. As you well know, the underlying reasons why were explained at length elsewhere. And denied only using wild speculation, hearsay, no numbers, and name calling. Meanwhile, experience and hard numbers explain why so much of our hardware did not fail even a decade later in harshest environments. Surprisingly, that means learning and using some of the simplest techniques to identify defective hardware that has not yet failed. Ie heat. Too many will cure heat with more fans rather than fix the existing problem.

Third, are good reasons why certain manufacturers tend to appear more often in reliability charts of consumer magazines. No reason to explain why here. A statistical fact considered when seeking reliable hardware.

Fourth, anything a UPS might do to protect hardware from damage must already exist inside a computer's power supply. Anything that a UPS might do to avert that damage must already exist in circuits inside every supply. For example, design standards even 40 years ago said a 120 volt supply must withstand up to 600 volts without damage. Today, supplies must be more robust.

But when selling supplies to computer assemblers without basic electrical knowledge, then many supplies are dumped into the market missing required functions. Too many assemblers only understand dollars and watts. Do not understand and therefore select supplies missing other required functions that might avert hardware damage. IOW many supplies, that may even violate FCC requirements, are sold to computer assemblers. The manufacturer is not responsible for meeting that and other international standards. Only a computer assembler is responsible for meeting those standards. Most computer assemblers do not know only they are responsible. And so the consumer may end up stuck with the consequences.

Four points to consider when seeking more reliable hardware.

 

[westom]

It's hard to conceive of a UPS that's so bad that a PC would be better without it. Well, I'm sure they exist, but you can get APC or CyberPower units starting at $40 or so.

Nobody said a PC would be better without it. Stated was that a typical UPS output (during battery backup mode) is some of the ‘dirtiest’ power electronics might see. AND (the part that conveniently gets forgotten) that 'dirtiest' power is made completely irrelevant by what must already exist inside every computer's power supply.

A UPS is temporary and 'dirty' power during blackouts. To protect from data loss. It does not and does not claim to do anything useful for computer hardware protection. If it did, then others would have posted the manufacturer spec number that makes that claim. No spec number was posted because it does not exist.

One must define a problem. Only then seek a solution for it. UPS is for one problem - to protect unsaved data. It does not and does not claim to protect hardware. But urban myth - without any hard facts and no numbers - will make that claim anyway. A UPS only does what it was designed to do - temporary and 'dirty' power during a blackout. Computers get some of the 'cleanest' power directly from AC mains. A typical UPS connects that computer directly to AC mains when not in battery backup mode.

Since this reality cannot be explains in a sound byte, then some misrepresented what was posted. Nobody said a UPS is bad for a PC. But that is how soundbyte logic perverted what was defined.

 

[bryanl]

I doubt slow holdup time is a problem since most backups switch over in 1/60 of a second, or less than the holdup times of all but the worst power supplies.

The clamping voltage is almosty always 330V. MOVs are by far the most common clamps and are among the fastest, working in just a few nanoseconds. The cheapest surge protectors use nothing but MOVs, anywhere from 1-3 sets, each set being MOVs in parallel.

 

[RebateMonger]

Nobody said a PC would be better without it.

The Title of the thread is a question:
"Computer longevity -- is no protection better than cheap protection?"

I was trying to answer that question.

 

[westom]

The Title of the thread is a question:
"Computer longevity -- is no protection better than cheap protection?"

I was trying to answer that question.

And I was noting how ineffective protectors can even make appliance damage easier. Maybe quote the first conclusion from Dr Martzloff's IEEE paper is necessary. He notes how plug-in (point of connection) protectors can even contribute to appliance damage (as we engineers saw when a power strip protector earthed a direct lightning strike through the network of adjacent and powered off computers). Dr Martzloff 1994 paper said:
> Conclusion:
> 1) Quantitative measurements in the Upside-Down house clearly show objectionable
>difference in reference voltages. These occur even when or perhaps because,
>surge protective devices are present at the point of connection of appliances.

Plug-in protectors are supplementary protection. But if the 'whole house' protector is not installed, then plug-in protectors can even made appliance damage easier.

Meanwhile, view numbers in a UPS. Protector numbers that are near zero. But just enough above zero as to hype 100% protection in retail propaganda.

Sometimes insufficient protectors can make surge damage easier.

And then another problem with grossly undersized protectors - fire:
http://www.hanford.gov/rl/?page=556&parent=554
http://www.ddxg.net/old/surge_protectors.htm
http://www.zerosurge.com/HTML/movs.html
http://tinyurl.com/3x73ol entitled "Surge Protector Fires"
http://www3.cw56.com/news/articles/local/BO63312/
http://www.nmsu.edu/~safety/news/lesson-learned/surgeprotectorfire.htm
http://www.pennsburgfireco.com/fullstory.php?58339

Earthing one properly sized 'whole house' protector averts all problems – and for tens or 100 times less money. Cheap protection is one 'whole house' protector. Expensive protectors (all plug-in protectors) can also make damage easier.

 

[theevilsharpie]

Sine-wave UPSes are pretty pricey. Modern PC power supplies don't really care. Sine-wave units are good for non-PC devices that CAN be sensitive.

There is a caveat to this advice when using cheap UPS's with Active PFC power supplies (i.e., pretty much any modern power supply).

http://www.apcmedia.com/salestools/RMUZ-7DTKRC_R1_EN.pdf