Intel U-turn finally means Cape Cod mobo recall

We don't want to say 'I told you so', but...

Regular readers will be aware – bored even – by our continuing coverage of the Caminogate horror story. Since October last year, we have been pointing out numerous performance and design shortcomings in Intel motherboards using the 820 chipset, as evidenced by the long list of related stories below.

On 18 April, we ran a story Intel recalls dodgy Cape Cod mobos, after an Intel Europe spokesman confirmed to us that this was the case. Intel immediately issued a denial and dissatisfied users of the mobo were officially told there were absolutely no plans for a recall.

On 21 April, Intel posted the following reply on its motherboard support forum to a user query on The Register's story on the Cape Cod mobo recall:

"Thank you for posting in our forum.

"No, it has been erroneously reported that Intel is recalling the CC820 board. While Intel does not comment on specific articles or publications, we can confirm that the CC820 desktop board has not been recalled nor are there any plans to recall the product for this issue. Due to the specific characteristics of the synthetic memory stress test, we believe the exposure to this issue is small for customers running real-world applications. If a board is found to exhibit this issue, it should be handled through the standard warranty repair process.

Regards
John S. MCSE, A+
Intel Technical Support"

Yesterday, user Ed Rutherford wrote in the same forum:

"Umm... isn't this the same problem that The Register reported last month that Intel fell all over themselves denying. Although there may not be an outright recall of Intel's 820 boards with the MTH it would have been nice if Intel had owned up to any potential issues with the 820 chipset/MTH at that time. A flat denial of problems then a flip-flop within 1 month on exactly the same issue erodes confidence."

And on the same day Intel issued a press release Intel To Replace Motherboards With Defective Memory Translator Hub Component in which it finally confirmed our recall story of almost a month ago, despite vehemently denying it for the best part of a month.

It would be nice to think that The Register had played some small part in this rather embarrassing U-turn on Chipzilla's part, but we believe that our original information on the recall was correct and that Intel was following its standard procedure of holding 27 conference calls weighing up its options before making the decision public.

In other words, Intel's technical people saw the need for a product recall, but the marketing people tried to brazen it out. Now common sense has prevailed and CC820 users are being offered either a refund or replacement – the latter would probably involve the Rambus 820 mobo, the Vancouver VC820. This raises the question of whether Intel will fork out for RIMMs to populate the board.

The release says "Intel plans to reserve for the cost associated with this replacement program when the cost can be determined. Depending upon the user replacement rate, the amount of this reserve could be material."

If the recall involves supplying large quantities of Rambus RIMMs, the cost will indeed be 'material'. ®

Sponsored: How to determine if cloud backup is right for your servers