Return Of The Dreaded 'Blue Screen'

Click here for reader feedback on this issue.
Are you confounded by your computer or otherwise ticked off by technology? Send your questions -- with hometown and state -- to The Help! Menu.

Question:

A friend brought his computer to me because he was having a lot of problems after putting in 2 (two) 32-meg RAM chips and a new modem. The icons would not display correctly and the blue screen of death would appear at random and when trying to install any new programs. The computer is a 133 (MHz).

The following would appear each time the blue screen came up: "An exception OE has occurred at 00:28 c0026080 in VXD Vfat(01) + 00008324. This call from 0028:C00267F9 in VXD VFAT (01) + 00008A6D."

I suspected it might be caused by the new memory, I removed the two 32-meg RAM chips and the blue screens have not been displayed since. This was odd, since at startup there was no indication of a memory problem, and it did detect all of the memory.

Is there a way to test the memory in the computer to see if it is bad? I don't have any other machine to install a single 32-meg chip and try and duplicate the errors.

Any advice or guidance would be greatly appreciated. Thanks in advance.

Bill, Vermilion, Ohio

Bill,

The memory tests that are performed at startup are more cursory than exhaustive. Defective memory can initially pass the startup check only to fail later. There is a wealth of memory testing tools available, but unless you're moving a lot of memory, it doesn't make much sense to invest in memory testing hardware. I suggest a complete diagnostic package like CheckIt that performs more comprehensive tests on memory and other components.

If it's convenient, you might want to return the chips to the retailer and ask them to test the memory and replace the defective chip. Anyone who sells memory expects a certain failure rate. They should have hardware to check the memory and to conclusively determine which chip is bad.

Bad Memory Or Slow Memory?

However, it's possible that the memory isn't defective, but instead is the wrong speed. Memory speed is measured in nanoseconds (ns). It's likely that the initial memory is slower than the two new chips. When the computer attempts to write a new "page" of RAM, it crosses over into the newer, faster memory and "freaks out" very badly. If you remove the initial memory and leave in the two new chips, error-free operation would indicate that you have a speed difference problem. You probably need to buy RAM with identical speeds for error-free operation.

Readers Rush To Defend The Mac

Last Friday I wrote about John and his Mac with the slow Web browser. (Click here to read Friday's column) Many "Digital Daily" viewers wrote in to offer advice to John about how to fix his Mac, and to offer advice to me about how to get my head examined. Before I get to the viewer mail, I want to share some facts about John, his computer and his solution to his problem.

• John is a real person. He works here at Digital Daily Headquarters as the Director of Creative Services.
• John uses a Mac G3 mini-tower with 256 megabytes of RAM. He is connected via a switched network of over a dozen T1 lines. His connection is fast.
• John says that his browsing problems are solved now that he uses iCab as his Web browser. He complains that he still needs to use Netscape for mail, but I'm hoping he'll use Eudora instead.

Now, on to the viewer mail:

From A.: "Regarding the misinformation you gave to John Barton of Saint Paul. What boat did you get off? Browser speed depends on, of course, 1, speed of modem connection - a cable modem is just fine!! -- and 2, how much RAM you have in your machine. What is with this trying to have someone switch to a PC? The Mac is very capable of fast browsing. You sound like a pitch person for PCs!"

From Perry: "Bashing the Mac OS is simply a way to make up for a true lack of leadership and direction in the Windows world. Apple computer was the first with plug and play, mousing, dual monitor capability, built in A/V, and holds several of the major tech patents that were filed before DOS & Win 3.1 could use them (QuickTime, for an example)."

From Steve: "When I use an IDENTICAL G3 Macintosh at a T1 equipped office to the G3 Macintosh I have at home, THE ONE CONNECTED AT THE OFFICE SCREAMS and consistently is as fast or faster than ANY other personal computer (or Unix box) in the place. (As a side note, the rendering engine in Netscape is the best comparison on multiple platforms.....admittedly, Microsoft has optimized the rendering engine for Windows and NOT done so on the Mac.....what would their incentive be to do so?)."

In all fairness, I was poking fun at John but I did not intend to "bash" the Mac. I respect Apple and the Macintosh computer, and I have no great love for Microsoft or the Windows operating system. This "Help! Menu" column alone proves how many difficulties people have with these products. I will let Mac aficionado John Barton have the last word by including a link to this GISTICS Study that he sent me. This document shows greater Return on Investment (ROI) for Mac users. You will need an Adobe Acrobat viewer to read this document.

The Self-Help Menu

After this column was published, more advice came in reader Michael Sommers:

Regarding "that guy who had the 'blue screen of death' message...I had a very similar problem, and it wasn't the memory that was defective, it was the memory refresh timing states in the BIOS.

"That will have a lot more effect on memory performance than anything else it could be. Normally, if one is having fatal OE errors, one should try increasing the wait states in BIOS. In my case, since I was going to faster SDRAM I had to decrease the wait states (from 3/3 to 2/2). Anyways, he should try this before investing in CheckIt or sending the memory back."

And this from Jon Sender:

"I, too, faced recently faced this problem when I added 128 k to my 32 k Pentium pro. Totally frustrated, I looked at the motherboard manual and noted that the motherboard required memory modules to be installed in a particular order (e.g., the two 64 k modules had to be in bank 0 and the two 16 k modules had to be in bank 1). After switching the modules around, everything works great.

"I should also point out that the manufacturer's Web site had other constraints regarding memory modules (tin connectors vs. gold, etc.). My advice would be to check out those manuals and Web sites."

Thanks for the help.