The Great Bug Hunt
Also in the last article, I was on a bit of a bug hunt and found a couple of goodies with the KT7A and the KA266. Dean seems to have found a couple as well. If you want to know what they are, I suggest you read his June Industry Update.
I have been investigating the partitioning problem I had with Windows 2000 and the KA266. Here is what I found.
Your hard drive geometry (size, interface speed, access method etc) is defined in part by the drive and in part by the BIOS of your motherboard. Most (if not all) large (< 528MB) hard drives use a method of storing data on the drive called Logical Block Addressing (LBA). The specifications for recognising large drives using LBA do not specify interoperability between different motherboards or BIOS’s. Here are a couple of very interesting points:
http://thef-nym.sci.kun.nl/cgi-pieterh/atazip/atafq-6.html see point 6.2 and 6.3.
In the real world (there’s that term again), you should be able to transfer hard drives set up on one PC and use them on another without problems. Should. Maybe. And I have – many times. But there is no guarantee that this will work all time, every time without error. And guess what – I ran into the error. I also had a look at the IWill support bulletin board and found many users experiencing errors when the had transferred their hard drive partitioned on one PC to there new IWill motherboard. This is not IWill’s fault. It just is. A pain? You bet. But it isn’t IWill’s fault, or a problem with their product. As I said before, there is no guarantee that simply transferring drives from one machine to another will work, just count yourself lucky if it does. BTW, the most common solution to this problem is to re-format the drive. This won’t solve this problem.
Unfortunately for me, simply repartitioning the drive did not work. In my attempts to format the drive, bad sectors had developed. For an IDE drive, this is usually a fatal sign. But not always. IBM and WD both (other manufacturers may also) have a utility that will completely blank (write zeroes to all parts of) the drive. I downloaded IBM’s utility, and it dutifully told me that I had bad sectors. It then asked me if I wanted to blank the drive. I said yes. If you have read this far and think that this form of drastic hard disk surgery is for you – Warning, this will destroy all data on your hard drive without a means of recovery – backup your data first.
I dutifully did as told and at the end of the (lengthy) process, I had a virgin bare drive. Time to re-partition and re-install. Again, Windows 2000 refused to partition and format the partition. Windows 98Se happily created and formatted it though and Windows 2000 was happy to use this partition. Since this time (almost two weeks now), Windows 2000 has behaved itself with aplomb. No more disk errors.
So now that I have regaled you with my partitioning prowess, whose fault is it. Must be IBM’s right? Wrong! It might be a faulty drive – only time will tell. At the moment it doesn’t appear to be. But there is one other complicating factor. If you go back and look at the way the KA266 was originally setup, you will see that the I used the Ultra memory timings. So what, I hear you say? This is where having intimate knowledge of BIOS performance timings and skill in electrical engineering comes in handy. I don’t have these skills, but MS over at Lost Circuits seems to know what he is talking about. He has made mention of the tRAS BIOS setting in a number of his articles – including his latest on the EPoX 8KHA. After reading this article, the penny dropped. Breaking out WCPREDIT, and again making reference to information at Lost Circuits, I had a look at the tRAS settings. The KA266’s Ultra memory settings have set tRAS to 4 – which is too fast for a 133MHz FSB Athlon. The Ultra setting should work just fine with a 100MHz FSB Athlon – but you are courting disaster using this setting with a 133MHz FSB Athlon. I may have been the instrument of my own stupidity.
I am not totally convinced that the drive isn’t in some way faulty (as the inability of Windows 2000 to create a viable partition attests), as the partitioning scheme isn’t outlandish, and Windows 2000 will happily partition and install on a 15GB 75GXP, a Western Digital 30GB Caviar 300BB and on a 20.5GB IBM 34GXP. But I didn’t help things :(
So I really have to give IWill back the half-point I subtracted for this supposed “bug”. It’s annoying, and not as convenient as just being able to shuffle drives around, but it isn’t a fault of IWill’s.
Be the first to discuss this article!