Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

Recently I've had some issues with my hard drive, where a file extracted from a compressed archive would get corrupted when it was extracted. Extracting the same file from the same archive once more would give me a healthy file.

Today I diffed the broken and healthy file and found that there was only a single bit difference between the files.

I've tried to run the Windows check disk utility to see if there was any bad sectors on the disk, but there wasn't.

What could the cause of these issues?

share|improve this question
1  
First I would use SpinRite. I then would purchase a new hdd, duplicate the existing hdd, and install the new hdd. –  Ramhound Dec 30 '11 at 13:08
3  
I'd be more suspicious of your RAM and the hard drive adapter than of the drive itself. Single-bit errors on a modern hard drive are pretty much impossible -- if an error sneaks through it would be multiple bits. –  Daniel R Hicks Dec 30 '11 at 13:13
    
I had a very similar issue when moving files from drive to drive. the Ram Timings were one number off from 150% stable. I was lucky to discover by using Comparitors before deleting original files. In my situation about 1-in-50 Files had a very tiny corruption. I had done a Memtest86 on the memory, before, but then changed the voltage a bit for the ram. I didnt bother to test it for Hours. After getting these corruptions added via its trip through ram, I made sure that I did Full overnight ram testing, and did not ignore any single ram error. It Could be a minor ram issue. –  Psycogeek Dec 30 '11 at 13:14

2 Answers 2

up vote 3 down vote accepted

I've had errors similar with bad ram that only showed an error in memtest86+ after about 24 hours of testing

share|improve this answer
    
Running memtest86+ did indeed show errors. This should probably be a question for itself, but as this is the third time I've had memory errors, could my motherboard be broken, and is it possible to investigate this? –  Decko Jan 2 '12 at 18:29
    
You did investigate , with the memtest. Chances are good you could just tweak your memory timings, and get it back to fully stable. If you want to blame your motherbaord and or ram, then test first with either higher timings, or higher voltage, or running the ram at a lower Ratio. If it fails when going "slow" then you might have bigger issues. Which wouldnt explain such minor problems. –  Psycogeek Jan 3 '12 at 2:56

Individual blocks of data on the disk have ecc, so at the point of reading from the disk, you have very high probability of detecting an error.

For better or worse the entire rest of the process of copying a file is unchecked, you're completely dependent on all the 99.99...% reliability of lots of subsequent components, and if one has a soft failure you can be screwed without knowing it for a long time.

Back in the stone age of computers (ca. 1978) we had one of the first semiconductor memories made by ampex for pdp10's, which it turned out would very very occasionally miss a write and send 36 bits to the wrong place. Every file that had been written in the last few months was "possibly corrupted", as were all of our backup tapes. What a mess.

My best advice to you if you have a documented case of a single bit error like this is to take all the hardware involved and send it to a landfill.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.