the Hidden Realities of Computer Industry in Japan Japanese

8.2 They charge for replacement of defective CPU


The most important computer part is CPU. It stands for Central Processing Unit. It works as brain to control whole computer and operate computing. The performance of computer mostly depends upon CPU. So, we check it first when we select computer.

When a fatal defect is found in hardware or widely used software, it will be reported on newspapers and TV. And the product will be recalled. But, if users don't have enough knowledge about computer, sometimes venders cloud the issue.

Long ago, I made a simple program of 3D CAD on a personal computer. In 1985 I purchased a newly released PC those days and began programming. It required 3D solid geometry calculation, so I needed to speed up trigonometric function computing. For the purpose of it, I had to add an LSI of numeric processing processor (coprocessor). I opened the cabinet, and inserted this LSI in proper position. It was supposed to speed up numeric operation. But the result was quite contrary; the PC stalled. I opened the cabinet again, and looked each part in it. I took out the CPU and looked it closely. Then I noticed its model number was somewhat different. It read "80X86-3", while it should be "80X86". This extra "-3" is the suspicious fellow. Final products don't have such model number. I guessed it was sample release.

At first I called the CPU maker. They gave me some information and phone number of the support office of the sales subsidiary, and then I called the subsidiary next.

In the meantime I understood the meaning of the extra sign of the model number. They told me the CPU with this sign doesn't work with numeric processor. Also they explained briefly about the digital circuit to use numeric processor with normal CPU. Their response was very kind.

I opened the cabinet again and looked at the circuit around the CPU. The parts were found just as they explained. I could imagine that with the normal CPU, high speed processing will be possible. They wouldn't make such circuit to be wasted.

Next I contacted with my fellows. As I expected, some of them were using this model. But strange to say, some received normal CPU from the vender before they ask them, and most of such companies had strong connection with the vender and was leasing the PC for free.

Moreover, I found the genuine processors were sold with free normal CPU. But they were extravagantly expensive.

I thought they would replace this CPU with normal one at manufacturer's retail store in Akihabara. However, they didn't just refused to replace the CPU and even accusingly asked me how I learned it didn't work. Catalog was showing the model number which would work properly, and my PC was containing the incomplete model. It should be replaced for free.

Our talk over the store counter didn't seem to end. And they called the head office about this settlement. I heard the person on the other end of the line saying, "it (normal one) cost 20000 yen. You can't hand it for free." They were going to brazen it out.

It was not until a fairly long time that the normal CPU was mailed to me with a piece of paper with just a short line of apology, saying we are sorry. By then I had already acquired the normal CPU from the CPU maker through another route and resumed my work.

This isn't the end of this story. The sale of that PC was not favorable. In those days, it was common in such case to donate them to some famous universities. The PCs were donated to a certain section of Tokyo University where I used to visit regularly. Those machines were used for computer literacy education. Two professors from that section asked me to take care of these computers. A graduate of this section produced the first microcomputer game in Japan. As this fact showed, there was nothing I could teach technical issue there. I'd rather learn something from them. I guessed that they actually didn't ask for technical support. They wanted me for other part of maintenance work.

The PCs were free, but other options weren't. The numeric processors were necessary to do complicated technical operations for their research. In those days the numeric processors were expensive, so they had to decide how many PCs should install them.

Meanwhile, the PC maker came, adjusted the machines, and set them in the computer room. I was surprised to know that the CPUs were replaced with normal ones. Tokyo University must have been on their list for free replacement.

This is very unpleasant fact for common users. In those days TV programs reported another PC maker for its bad user support, but it was quite trifle comparing to this case.

Later I met a staff from a PC magazine publisher. We talked about this case, too, considering if it should be written on the magazine. But it wasn't realized after all.

Computer is made by human who is not perfect, and it's not possible to eliminate every possibility of malfunction, calculation error, and damaging data. But it's a different story when someone is trying to cover up a known defect. They try to cheat users openly if they think the users don't know much, and sometimes they even threat when the users find it out.

This kind of case rarely happens to me. The reason may be they know my job title or knowledge level of computer. So they accept me politely from the beginning. Generally computer engineers, especially those in casual clothing, kindly respond to users. If it's on the contrary, it seems to be wise to beware of their response.

If you don't know enough, you'll never know if they're giving you all they have to.


Copyright 1996, 2000 Hirofumi Fujiwara. Translated by K.S.
No reproduction or republication without written permission.

the Hidden Realities of Computer Industry in Japan Japanese