Atomic I/O letters column #139Originally published 2012, in Atomic: Maximum Power Computing
Reprinted here February 27, 2013 Last modified 16-Jan-2015.
My old ThinkPad's 20V, 65 watt power supply died. The laptop still worked from my Piece-O-Crap(TM) car cigarette lighter power supply, so I bought a new off-brand mains PSU for it for $9.49 delivered on eBay (instead of $38 for a real Lenovo one).
The sticker on the new power supply actually says "19V 5A", though. I remember the thing you wrote years ago about high current capacity not being a problem, and 19 volts times 5 amps should add up to more than enough power even if the sticker is lying, but I didn't like the look of that lower voltage spec.
I got out my $10 yellow multimeter that I bought after about the tenth time you mentioned that everybody should have one, and the actual voltage seems to be about 22 volts! This alarmed me a bit so I checked the car charger as well, and that's outputting 21.6 volts, though it says 20 on the sticker!
I remember you saying that it's normal for plugpacks to deliver more volts when "open circuit" than when actually plugged into a load, but these things seem to be all over the place. The car supply cost even less than the new mains one so I wouldn't be a bit surprised if it was killing my laptop - is it? Am I a fool not to have gotten brand-name everything?
You're almost certainly fine.
Dirt-cheap laptop power supplies might be more prone to catch fire or screw up in such a way as to deliver mains voltage directly to the laptop, but in reality the worst they usually do is just drop dead harmlessly after not very much use.
Specifically, the higher voltage probably isn't a problem. Many modern laptops are tolerant of a surprisingly wide input voltage range; their power supply is actually split into two, half in the actual separate AC adapter box and the other half inside the laptop. The exterior portion only needs to deliver sufficient current at something vaguely near to the rated voltage. It's DC-to-DC converters and regulators inside the laptop that feed the actual computing hardware, and they can take all kinds of weirdness and filth in the incoming power in their stride.
You probably can't run a nominal-20-volt laptop from twelve volts, and greatly exceeding the expected input voltage is not a good idea. But pretty much everything ought to be fine with an input voltage ten per cent over or under the specified value, and realistically, 20% is likely to be OK too. (As usual, though, all care, no responsibility. When a fragment from your exploding over-volted laptop kills the passing grandson of a Mafia don, you're on your own.)
In the olden days, this was not the case. Old laptops commonly have several pins on their power plug, delivering the same several voltages that a desktop-PC PSU outputs, and possibly also weird stuff like a clock signal.
My Windows 7 64-bit PC can no longer see the Internet, or anything else on the local network, and I don't know why.
I've got a Netgear DGN1000 combination DSL modem/access point/ethernet switch, and two other computers - a Windows 7 laptop, and a WinXP old PC that I use as a print server when I use it at all. The laptop sees the Internet and the old PC via Wi-Fi, the old PC sees the laptop and the Internet via wired Ethernet, but my main PC doesn't see anything anywhere. I plugged a USB Wi-Fi adapter into it and that didn't work either, though it does on the other two computers.
If it was just the wired Ethernet then I'd figure the motherboard network adapter had died and replace it with a $10 network card, but I'm now thinking it's a software problem. I just don't know what the software is.
I got Flynn to download (using the laptop...) a Linux LiveCD, and try booting the offending computer from that. And lo, Ethernet suddenly worked fine again.
My first suspicion, given this, was that someone had set Windows Firewall or some similar program to block everything. But that wasn't it.
It was bloody commercial antivirus software.
In this case it was an update to McAfee AntiVirus that decided to protect Flynn's computer from Internetually Transmitted Diseases by the 100%-effective means of total abstinence. But McAfee is not the only antivirus program that can do something like this. I don't know when it was that paid security software for home users became more trouble than it's worth, but it's been that way for years now.
It is a good idea to have some kind of anti-malware software on a Windows PC even if you're not prone to click on anything that tells you to do so, but I'd stick with the free, and well-behaved, option of Microsoft Security Essentials if I were you.
(Windows 8 has anti-malware software built in. I'm sure Symantec already have a very plausible explanation for why you should use their stuff instead.)
I've got a Core i5 PC with 4 CPU cores, but I've noticed that if something's clogging the first CPU up, a lot of software doesn't seem to use the other cores very well, so the computer gets all slow and jerky.
This is not usually a problem, because it doesn't happen very often and I can change priority and CPU affinity of offending processes in Task Manager, but recently it's been happening ALL THE TIME and I cannot find the process that's causing it. If I run other CPU-intensive stuff I can see that it all only adds up to 75 to 80% CPU utilisation in Task Manager, meaning one CPU is being completely used up (which I can see in Task Manager -> Performance, too), but what's using that extra CPU time is a mystery, and now my PC sucks.
Now I'm thinking there's some stealth botnet thing running, though Microsoft Security Essentials says my computer's clear.
Lousy performance when only one CPU core is heavily loaded may not actually be because of software that insists on using that same core, but be because the processor is overheating and throttling its clock speed down to avoid a crash. A single fully-loaded core can easily cause this, even if it's the only one that's overheating; the whole CPU slows down when even one core needs to be throttled. You can track this sort of thing with a CPU usage and temperature monitor; for Intel CPUs, I use Real Temp.
If it's a heat problem, then it may be insoluble without improving your computer's cooling, or just killing whatever task's using all that CPU time. Moving the task to another core and reducing its priority won't make much difference, but that's not hard to do, so you might as well try it first. Once, of course, you've figured out what the heck the offending task is in the first place.
The culprit is probably a service, not a normal program.
Services are, essentially, background tasks with which users are not meant to directly interact. Task Manager has a tab to display services, but neither it nor the Control Panel "Services" interface will show you if one's hogging the CPU or flogging the hard drive.
In general this isn't a problem, because services aren't meant to do that sort of thing, and usually they don't. To see if one is misbehaving, though, use Process Explorer, the utility so good that Microsoft bought out the guy that made it. Process Explorer shows services alongside normal processes, and lets you alter task priority and CPU affinity for them, just like normal processes.
Shortly after I sent the above, Piper got back to me with the news that the culprit was indeed a service - "CarboniteService", which feeds the Carbonite online backup system, which Piper subscribed to a while before this problem started. Setting CarboniteService's priority to Below Normal and its affinity to CPU 3 solved the problem, but Piper now needs to do that every time she restarts the computer. If you only restart your computer every several weeks, though, that's not a major problem, so you don't need to mess around with outboard utilities or executable-editors to do it automatically.
The share house I'm living in is kind of long and skinny, and one Wi-Fi access point can't cover all of it. So we want to extend the network with a second AP in... extend mode... (not really a networking expert, here), but my Zyxel ZyAir G-1000 AP and one housemate's Apple Airport Express don't seem to want to do that, no matter which one's the "master" actually connected to the Internet and which one's the extender.
What should we do?
There's a name for the protocols involved with extending a Wi-Fi network, "Wireless Distribution System" or WDS. But there's no actual standard for it, and it's normal for different brands to not work at all with each other, or to not work if you want any security better than the extremely-crackable WEP, or to have other personality disorders. The same goes even for purpose-built "wireless extenders" that can't operate as a normal access point by themselves. I suspect your Zyxel AP - which dates back to 2003 - is too old to be extendable by anything.
Any two recently-made and identical wireless access points that mention WDS in the manual will probably work. Any two recently-made access points (or one AP and one extender) from the same manufacturer are a pretty good bet, too. So if your share-house budget stretches that far, getting a couple of new Netgears or Belkins or whatever, or some cheap off-brand access points that the seller swears have a working WDS and will give you a refund if they don't, should solve your problem.
A cheaper alternative is to run an Ethernet cable down the hallway, from your Internet adapter or a network switch connected to it, and plug any old access point in at the other end. Then you can set up two separate networks, which shouldn't be too much of a nuisance as long as nobody wants Internet access with no interruptions while walking from one end of the house to the other.
I just noticed that when I do a Google search in Firefox, the ads on the results page are not the usual Google ones. The ad-links all go to google-co.com/... now, which obviously wants you to think it's a Google server, but isn't.
Do you know what's going on?
You installed the "IE Tab Plus" extension for Firefox, didn't you?
As well as letting you put an Internet Explorer browser session in a Firefox tab, that extension hijacked Google ads. Its user reviews on the Firefox add-ons site were, for this reason, sharply split between happy users who didn't notice the hijacking, and very unhappy users who did. (In the months since this column was published in the magazine, IE Tab Plus seems, unsurprisingly, to have been removed from addons.mozilla.org.)
When I wrote this for the magazine, google-co.com was already down. As of February 2013 it's still down - there's still a server there, but it's one big 404 error, plus this odd little snippet, for some reason.
Without google-co.com to serve its nasty ads, IE Tab Plus may do the exact opposite of its author's intent, and leave your search results completely ad-free.
This particular browser hijacker was, I think, pretty benign, which is why it managed to get positive reviews. Your typical hijacker - which has a pretty storied history, by malware standards - replaces your home page and search engines, pops up ads and causes "you" to "click" on them, and wedges itself into place using the same extremely annoying tricks as much other modern malware. Something that just surreptitiously replaces ads and also, say, quietly harvests passwords, can fly under the radar.