I have a Samsung HD203WI attached to a Asus P5Q Premium motherboard. I checked the SMART info the other day out of curiosity using HD Tune and it has some CRAZY rising G-Sense error rate in the RAW data column. I confirmed it with CrystalDiskInfo and Speedfan to make sure it was not just HD Tune. The rate is increasing by the second and it in the hundreds of thousands. I would not say the case it is in has tons of vibration. But surely this is not normal? The funny thing is HD Tune has 'ok' next to the column so it's not even marked it up as a problem. Speedfan says on the drives webpage that it is to be watched though.
Is the drive getting read to go bye bye or is this just some stupid oversensitive sensor or driver issue?
AFAICT your drive is racking up one G-sense error every 6 seconds. The Write Error Rate is also very high. I suspect that the write errors are directly related to the G-sense errors.
Hard drives usually have one or two shock sensors on the PCB. These are usually mounted at 45 degrees in order to detect vibrations in both horizontal axes. If the drive detects an excessive shock event during a write operation, then this would result in a write error (or so I would expect).
I suggest you remove the drive from your PC and place it on a soft static-free surface. Then run HD Tune's write benchmark, and record the SMART readings before and after.
I can't perform the write benchmark whilst there is data on the drive. Also I put it on elastic bands outside the case and it still racks up G-Sense errors. Not just every 6 seconds. But 50-200 per second! Everytime I hit refresh the raw data has increased by that amount. Originally I was running the drive vertically mounted in an eSATA dock due to lack of space btw. But moving it into the case (Or on elastic bands) makes no difference. The G-Sense rate just continues to increase exponentially.
EDIT: OK I took the drive downstairs and put it in an eSATA dock attached to the JMicron controller and the G-Sense errors stopped dead! So something is amiss when connected to the P5Q. I suspect it's a chipset or driver issue. I will have to try it using MS Standard IE driver and see.
EDIT2: I put it back in the PC upstairs and the errors increase again. So I took it out and put it back in it's original eSATA dock which is connected to a different chipset. Guess what? The errors stopped rising. So it appears that there is an issue with the ICH10 controller on the P5Q and this particular hard disk. I don't think there is anything wrong with the drive other than when being connected to that chipset and a few unsafe disconnects using the eSATA dock (Which probably accounts for the Write rate errors). It's a shame I can't clear the SMART info as it looks like there is a massive issue with the drive when there probably isn't.
I wonder whether your drive has a dry solder joint at one of the shock sensors. That may account for the difference in behaviour when the drive is horizontally mounted. Try running the drive upside down. I'd also inspect the PCB with a magnifying glass. Many techs are reporting intermittent failures due to the lead free solder that is mandated by RoHS. This malaise affects all consumer gear. Lead free solder also suffers from "tin whiskers". RoHS is very bad news. :-(
I'm going to copy the data off onto an Hitachi when it arrives later this week. Hopefully it will not die before or whilst doing so. Once that is done I shall experiment more with the drive. If it is a problem with the sensor then I shall rma it.