Atomic I/O letters column #15
Originally published in Atomic: Maximum Power Computing Reprinted here 29-Nov-2002.Last modified 16-Jan-2015.
Promedia plugging
If I buy a set of Klipsch ProMedia speakers from North America via eBay, will they be compatible with our power outlets?
David
Step-down transformers let you run US appliances in Australia, but
one powerful enough for a decent set of speakers won't be cheap.
Answer:
Nope. Different pins, different voltage. You can run US appliances from
Australian power with a step-down transformer, but the price of a transformer
powerful enough to run your speaker system, plus the amount of money you'd
have to pay to get the things here, plus the
GST you'd stand a good chance
of having to pay as well, would certainly cancel out any savings.
Overclocking inquiry
I was wondering how to overclock an AMD Athlon 900 using a Gigabyte 7IXE4 motherboard. I'd like some tips so I don't f**k it up.
Alexander
Answer:
Since that board doesn't have multiplier adjustment, there's no point in
unlocking your CPU multiplier. All you can do is wind up the Front Side
Bus from the stock 100MHz. To do that, use the "SW1" block of DIP switches,
as documented on page 7 of the manual. If you don't have the paper manual
handy, you can download a PDF format version of it from
here.
There's no CPU core voltage adjustment on the 7IXE4, but its FSB adjustment only goes up to 115MHz anyway (1035MHz, from your CPU), so it wouldn't make much of a difference.
Verti-drives
I forgot if it is OK to mount HDDs on their side. Is it?
Steven
Answer:
Yes, it's fine.
In the olden days, hard drives had to be operated in the same orientation as they were low level formatted, because their stepper-motor-based head positioning systems would put the heads in slightly different places when you aligned the drive differently. But modern drives can't be low level formatted, and their voice-coil-based positioning systems have feedback that lets them compensate for gravity in any direction. So they don't care what way round you install them.
UDMA unavailable
I've got a P4 machine with a SIS 650 chipset (Asus P4S333-VM). I've spent roughly 8 hours finding out that Ultra DMA mode doesn't seem to work with this motherboard and my spiffing Seagate Barracuda. I was wondering if this is because of the MB/HDD combo or just a shitty chipset, is this a common problem and would I be wrong to assume that a 3rd party IDE card like the Highpoint RocketRAID 133 would fix the problem?
Oh, even though this "feature" is not mentioned anywhere on the ASUS site, the BIOS revision 1005 seems to disable UDMA at on the HDD, not the CDR/DVD though, what's up with that?
Anthony
Answer:
There apparently were, past tense, some issues with P4S333 boards
and UDMA, but according to Asus' BIOS file info the problem was solved with
BIOS version 1.005.
There are some other possibilities, though.
If you don't get any UDMA mode reported on startup - if you're seeing some Programmed I/O (PIO) mode reported next to the drive's name on the boot display - then you'll be running the drive at half UDMA/33 speed at best, with considerably more CPU load, and something is amiss.
If you only get "UDMA 2" reported, then that's UDMA/33. Modern hard drives that support UDMA/66 or 100 or 133 will run in UDMA/33 mode if you connect them with a 40 wire cable, or with an 80 wire cable the wrong way around - with the end hard drive connector plugged into the motherboard, and the motherboard connector plugged into the hard drive. The hard drive connectors are the two that are closest together on the cable.
It's also possible to manually lock an IDE channel to one UDMA mode in the BIOS setup, though that's unlikely to be done by default.
All modern IDE controllers have independent timing registers for the two devices that can be plugged into each IDE channel, which means that putting drives with different maximum speed modes on the one cable won't restrict the faster device to the slower one's best transfer mode. That used to be the case, and many people still think it is, but they're wrong. When the slower device on a modern controller is doing a transfer the faster device will have to wait, but that's as bad as it gets. So your drive shouldn't be held back by some other device on the cable, if there is one.
If the problem you're seeing is that you can't activate DMA mode for the hard drive in Windows, then try reinstalling the motherboard drivers and making sure your Windows install is fully patched and polished.
And yes, adding a separate IDE controller card will deal with problems like this. You'll need drivers for that as well, though. Your operating system may or may not have drivers built in, depending on what you're running.
16 colour hell
I seem to have a problem updating the Detonator drivers for my Asus AGP-V7100 GeForce2 MX. Every time I update the standard Windows XP drivers to anything else I end up only getting 4 bit colour. Can you tell me why?
Mark
Answer:
If you're not actually installing the right drivers, that could explain
it. Nvidia offer
different driver flavours for different Windows flavours; install the wrong
one and you're likely to see symptoms like this.
Alternatively, your Windows video drivers may just be hideously messed up. This was (and is) a common enough problem for Win95-series OSes (95, 98 and ME), but it's possible for the NT-series Windows versions as well (NT, 2000, XP). Generally, the solution to this problem is either painful manual file deleting and registry editing or a Windows reinstall, but WinXP's "repair an installation" option may fix it for you without requiring the full nuke-from-orbit routine.
Dual XPs and homicide
I recently had an argument with a fellow computer geek while playing a game of Counter-Strike. The argument was over AMD Athlon XPs. This particular person was convinced he was purchasing a dual XP system. I on the other hand was convinced that it was not possible to have dual XPs as they do not support SMP. He continued by saying that ASUS made a motherboard specially built for XPs, but I ended the argument by placing a piece of lead in his brain base with a Steyr Scout.
Although I won the argument on the day, I would still like to know who was right.
Simon
Answer:
You may have had the might, but you weren't actually right. Athlon XPs are
likely, but not certain, to work fine on many dual-CPU Socket A boards.
So are Durons, for that matter.
The difference between the XP and the MP is that the MP is certified by AMD to work in multiprocessor configurations, and the XP isn't. There is a real difference beyond what's printed on the processor and guaranteed by the manufacturer; people fooling around with the early Tyan dual Socket A boards found considerably more... quirks... when trying to make XPs and Durons work than when they used MPs. There are other dual boards around now, which have better XP compatibility, but the MP certification does still mean something. If you're doing something mission-critical, drop the extra dollars on proper MP chips.
Most current XPs, from the 1700+ on upwards, have a bridge cut to stop them from working in MP mode. Re-join the bridge and they work fine, and all of the newer dual Socket A boards seem to stand a good chance of working with them. Dual XP compatibility is still an unofficial feature, though.
Dead drive?
It all happened when I tried to format. I got the 98-boot disk and booted with CD-R support. Got into FDISK and deleted primary DOS partition. All fine and well - deleted. I then recreated the primary DOS partition and it came up with the normal "restart and format" screen, all fine and well. Now the next part is a bit of a blur but I'll try to re-create it.
I inserted the Windows ME disc and restarted. It came up with the normal screen asking what you wanted to do - start WinME install, boot with CD-ROM support, boot without CD-ROM support. I picked the Install ME option. It got in and I think there was an error or something so I decided to restart. Suddenly I couldn't get into setup, so I tried to run FDISK again. And all it came up with is "no fixed disks found". And in the BIOS it says there is no primary IDE installed.
I have tried auto-detecting the primary IDE device on startup but nothing happens. So I am in a bit of a pickle :(.
My mate has suggested that it might be a problem with FAT32 and NTFS?
Sascha
Answer:
Whatever it is, it ain't anything to do with the filesystem(s) you've been
using.
I would surmise that your computer just picked that moment to suffer a hardware failure. Probably the drive, maybe something on the motherboard, conceivably something else. If the basic BIOS drive detection can't find your drive, and no cables have been yanked, then that drive has probably just decided to become a paperweight.
You can check this, of course, by trying the drive in a different computer, or a different drive in your computer.
Mobo hawking
I'm soon going to upgrade my computer and sell my old parts. What is a decent price to sell an "A-Plus AP-941" Pentium 3 motherboard? It's only three months old and in excellent condition. I'm planning to sell it in the Computer Trader.
Jonathan
Answer:
"Aplus" motherboards are a craptacular rebadged pseudo-brand which, I think,
only exists in Australia. That's why you can't find a Web site for this
"manufacturer". The local distributor, or rebadger, seems to be
Achieva Australia.
I think your board's actually a PC Partner product (itself hardly a famous name). Just going by the model number, I'm thinking it might be this one or this one, but I'm guessing. It could be a whole different manufacturer.
Shift it for whatever you can get. Schnooks on auction sites will pay $AU80 or so for a board like this.
Slow ripping
I have a Diamond Data 36X CDROM drive that works fine for most applications, but I cannot for the life of me get it to rip audio CDs to WAV files above about 2.5X. This seems ridiculously slow, even given the parlous state of the rest of the box (200MHz Pentium, non-UDMA hard drive). Using Exact Audio Copy I can get a perfect quality rip, but only at about 1.0 speed. Audiograbber will do about 2.5 speed, but at terrible quality. I have upgraded the CD-ROM firmware, driver, and flashed my motherboard BIOS with no success. I have also fiddled with all drive setting in System Properties and EAC with similar results.
It is obviously not an earth shattering problem but it is bugging the hell out of me. Any help would be appreciated!
Matt
Answer:
The general consensus about this seems to be that most if not all Diamond
Data drives (made, or at least sold, by
Mitsubishi Electric)
just aren't very good. Well, not for audio ripping, at any rate.
I don't know whether this'll work or not, but it has been unreliably suggested that if you try ripping right after closing the drive door - while the drive's still doing its new-disc spin-up routine - you may get much better performance. Or you may not.
It's cheaper to try this than to buy a new CD-ROM drive, anyway.
Network irritation
I'm networking two PCs with a crossover cable. One computer is an Athlon runnin WinXP, the other a Celeron running WinME. The workgroup is called MPC. The Athlon is showing itself and the Celeron, which is good, but I can't access the Celeron; it says "the network path could not be found". The Celeron is showing itself but not showing the Athlon at all. I've had them going before, but since I reformatted both machines, I cant remember how I did it. Is it possible a hub could have made my networking the first time easier?
Shanan
Answer:
Point 1: This isn't likely to be a hardware problem. A crossover cable is
a perfectly OK way of connecting two (and only two) Ethernet devices. Some
NICs don't like crossover cables, but if that's the case, the computers
won't communicate at all. Cable problems will usually either cause the network
to work more or less properly but very slowly, or prevent it from working
entirely.
Point 2: There are lots of things that can go wrong with Windows networking, particularly when you've got different Windows flavours trying to talk to each other. Commonly, it Just Works. Sometimes, it Just Doesn't.
Making older boxes visible to WinXP ones, with only 283 clicks!
My first guess about why this is happening is that you don't have a local DNS server (with just these two machines, you probably don't), and NetBIOS over TCP/IP is disabled. That can make Win95-series machines inaccessible to WinXP boxes.
To check this, go to Control Panel -> Network Connections on the WinXP machine, right-click the LAN connection and select Properties from the menu. Select TCP/IP and click the Properties button, click the Advanced button, go to the WINS tab, click the Enable NetBIOS over TCP/IP radio button, and OK your way out. Now go outside and meditate for a few minutes, because it may take the XP machine that long to see the WinME one. Or just reboot everything in sight.
If this doesn't help, there are other things you can try.
Update both machines to the latest Windows patch version using Windows Update, and update your network card drivers as well, if you can.
It's possible to turn WinXP's firewall feature on for local network
connections. Don't.
Make sure you haven't accidentally turned on WinXP's firewall feature in the Properties for the Athlon's LAN connection.
Disable any other firewall software you're running on either machine.
Disable other extra software, like virus checkers. Virus checkers don't usually cause problems like this, but they can.
Also, if the Celeron can't see the Athlon in Network Neighborhood, that may just be because the Athlon doesn't have any resources shared. Machines with nothing shared don't show up.
Oh, and you know what's really irritating about this situation?
After I e-mailed all this to Shanan, he replied and informed me that he'd figured it out. It actually was the bloody cable all along.
