I bought a brand new Samsung HD501LJ 500gb hard drive yesterday, and used Norton Ghost to copy all data from my 250gb samsung to the new drive. The HD501LJ is now the only hdd in my system. The drive is suspended frm elastic in the lower drive bay of my p180. It is generaly very quiet, I cant hear it in idle and seek is also very quiet. However, it occasionaly makes a louder 'knocking' noise that I can hear quite clearly.
Is there something wrong with my drive?
HD501LJ making knocking noise?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
HDDScan (freeware) can both scan the surface and read SMART data. There's several scans available:
Host controlled scans (application asks HDD to read, verify or write "sector-by-sector"):
- read scan (read data from platter and send them from HDD to application asking it (but HDDScan doesn't do anything with the data, it only counts how much time it takes to transfer it))
- verify scan (like read, but data that is read is not sent through the data cable. HDD only verifies that it can be read and error correction code (ECC) is matching for each sector, then send some kind of "OK" signal to application asking for verification of that sector) (probably equivalent to Windows scandisk or chkdsk surface scan. HDDScan doesn't scan the file system, though)
- erase scan (will delete all data on all sectors tested. Full scan => all partitions and all data erased)
SMART self-test scans (SMART is not just "SMART values" but a set of self-test scans as well):
-offline data collection routine (does something... non-stop?)
-short self-test in offline mode (performs analysis of mechanical subsystem, servo, head amplitude, etc. + scans some sectors (typically N first and M last sectors))
-extended self-test in offline mode (performs analysis of mechanical subsystem, servo, head amplitude, etc. + scans all sectors)
Self-tests run in offline mode (= details of testing is handled by HDD itself), so they cause very little CPU load while the test is in progress.
Read => every sector returned => causes severe traffic over the interface. Noticeable CPU load.
Verify => every sector confirmed => causes some traffic over the interface. Noticeable CPU load(?).
Offline => only periodic messages reporting progress to host system => Negligible traffic over interface and negligible CPU load.
All offline scans are pretty much offline but some are more offline than others. Offline data collection routine send pretty much no periodic messages of progress (and it may actually be a non-ending routine) and is thus purely offline. (HDDScan support only one HDD in the system performing short or extended self-tests, but several HDDs may be scanned simultaneously with data collection routine. Host controlled surface scans (read/verify/write) can be performed only to one HDD at a time... but launching multiple instances of HDDScan may (or may not circumvent that limitation).
Haha! Made you read a lot of technical babble just to say: use "extended self-test in offline mode" (in HDDScan under SMART Self tests tab). That's the most complete scan there is. If you run Samsungs own diagnostics, the software will probably issue that command if you chose the "extended test" option. The difference between HDDScan and Samsung's diagnostics is merely
- under which OS the software runs (manufacturer utilities typically use bootable floppies that include tiny PC-DOS operating systems). HDDScan works with at least with WinXP and probably most of it's relatives. It does NOT work with Vista because Vista sucks. :p
- how the software reports possible problems. Manufacturer diagnotics may just tell you OK or FAIL. They don't let you read SMART data. They read SMART to diagnosize it, but it's not shown to the user. HDDScan displays the data to you. (Just hit the button that reads "S.M.A.R.T." (short for Self-Monitoring Analysis and Reporting Technology))
Host controlled scans (application asks HDD to read, verify or write "sector-by-sector"):
- read scan (read data from platter and send them from HDD to application asking it (but HDDScan doesn't do anything with the data, it only counts how much time it takes to transfer it))
- verify scan (like read, but data that is read is not sent through the data cable. HDD only verifies that it can be read and error correction code (ECC) is matching for each sector, then send some kind of "OK" signal to application asking for verification of that sector) (probably equivalent to Windows scandisk or chkdsk surface scan. HDDScan doesn't scan the file system, though)
- erase scan (will delete all data on all sectors tested. Full scan => all partitions and all data erased)
SMART self-test scans (SMART is not just "SMART values" but a set of self-test scans as well):
-offline data collection routine (does something... non-stop?)
-short self-test in offline mode (performs analysis of mechanical subsystem, servo, head amplitude, etc. + scans some sectors (typically N first and M last sectors))
-extended self-test in offline mode (performs analysis of mechanical subsystem, servo, head amplitude, etc. + scans all sectors)
Self-tests run in offline mode (= details of testing is handled by HDD itself), so they cause very little CPU load while the test is in progress.
Read => every sector returned => causes severe traffic over the interface. Noticeable CPU load.
Verify => every sector confirmed => causes some traffic over the interface. Noticeable CPU load(?).
Offline => only periodic messages reporting progress to host system => Negligible traffic over interface and negligible CPU load.
All offline scans are pretty much offline but some are more offline than others. Offline data collection routine send pretty much no periodic messages of progress (and it may actually be a non-ending routine) and is thus purely offline. (HDDScan support only one HDD in the system performing short or extended self-tests, but several HDDs may be scanned simultaneously with data collection routine. Host controlled surface scans (read/verify/write) can be performed only to one HDD at a time... but launching multiple instances of HDDScan may (or may not circumvent that limitation).
Haha! Made you read a lot of technical babble just to say: use "extended self-test in offline mode" (in HDDScan under SMART Self tests tab). That's the most complete scan there is. If you run Samsungs own diagnostics, the software will probably issue that command if you chose the "extended test" option. The difference between HDDScan and Samsung's diagnostics is merely
- under which OS the software runs (manufacturer utilities typically use bootable floppies that include tiny PC-DOS operating systems). HDDScan works with at least with WinXP and probably most of it's relatives. It does NOT work with Vista because Vista sucks. :p
- how the software reports possible problems. Manufacturer diagnotics may just tell you OK or FAIL. They don't let you read SMART data. They read SMART to diagnosize it, but it's not shown to the user. HDDScan displays the data to you. (Just hit the button that reads "S.M.A.R.T." (short for Self-Monitoring Analysis and Reporting Technology))
Well I ran that program and if it found any errors it's not telling me. I enabled AAM with it, doesnt seem to make any difference. The knocking sound seems to come at the beginning of a seek, just before the (very quiet) 'scratchy sound'. My friend says maybe its the read heads engaging or something but i have no idea.
SMART values are all OK.
Yes, the noÃse could be head parking/unparking (or disengaging/engaging, respectively). Even if we could be certain they caused by (un)parking, it would only take us closer to the real cause but we would be looking the true cause itself. (Un)parking during operation (not just when powering up/dowm) may or may not be normal. For laptop drives, it's a defacto standard power management feature that cannot be disabled, but 3.5" HDDs have similar features disabled by default. They may be enabled by you, if you wish to enable them.
Unwanter (un)parking noises could be caused by HDD reseting itself. Typically HDD is never spun down, just parked and then unparked, in rapid succession. HDD may reset itself if there's problem with media (bad sectors, bad servo information), problem with R/W head or for some other fault internal to HDD.
HDD may also reset itself when there a problem with PSU or power cable or connector.
Happy trouble shooting!
Yes, the noÃse could be head parking/unparking (or disengaging/engaging, respectively). Even if we could be certain they caused by (un)parking, it would only take us closer to the real cause but we would be looking the true cause itself. (Un)parking during operation (not just when powering up/dowm) may or may not be normal. For laptop drives, it's a defacto standard power management feature that cannot be disabled, but 3.5" HDDs have similar features disabled by default. They may be enabled by you, if you wish to enable them.
Unwanter (un)parking noises could be caused by HDD reseting itself. Typically HDD is never spun down, just parked and then unparked, in rapid succession. HDD may reset itself if there's problem with media (bad sectors, bad servo information), problem with R/W head or for some other fault internal to HDD.
HDD may also reset itself when there a problem with PSU or power cable or connector.
Happy trouble shooting!
There's some cosmetic glitches in HDDScan but they don't much limit the usefulness of the utility. Device mode box is not quite real-time, and (regular) Power Management slider doesn't show current value at all. You can move it to desired value and click "Set timer". Timer is set, but slider returns to "disabled" position.
Also, if one use HDDScan to benchmark STR, it's not irrelevant whether "Map" or "Graph" tab is active. It will affect the STR slightly and it is visible. Switching between Graph and Map view makes the Graph look more jagged than what it really is.
Also, if one use HDDScan to benchmark STR, it's not irrelevant whether "Map" or "Graph" tab is active. It will affect the STR slightly and it is visible. Switching between Graph and Map view makes the Graph look more jagged than what it really is.
Aeria, I'm using 8 HD501LJ drives (on 2 PCs) and not a single one makes the loud 'knocking' noise you describe. Unless you've been modifying some settings (APM maybe but I think it isn't supported) I'd say this isn't normal at all. I think all HD501LJ drives get delivered with AAM off, APM off or disabled. Only S166 drives come with AAM set to the lowest value.
The only time I heard some strong clicking noise from the drives is when I had a BSOD : the computer rebooted itself automatically and each HD501LJ used in my RAID-0 array made a single strong clicking noise when reset by the ICH8R RAID bios, then everything was fine again.
The only time I heard some strong clicking noise from the drives is when I had a BSOD : the computer rebooted itself automatically and each HD501LJ used in my RAID-0 array made a single strong clicking noise when reset by the ICH8R RAID bios, then everything was fine again.
"Unless you've been modifying some settings (APM maybe but I think it isn't supported) I'd say this isn't normal at all."
True, but (regular) PM is supported. And like I said, due to a glitch in HDDScan (or possibly the HDDs lacking a feature for polling it's current status): current (regular) PM setting is always unknown. To eliminate that possibility I recommend moving the slider slightly and then moving it back to zero = disabled. Then click "Set timer". (Or move it to some place, enable timer with time N, "Set timer", move it back to zero, "Set timer". This should not be necessary, but since current value is not shown, who knows.)
AAM probably has nothing to do with it. OP mentioned there's a steep difference between seeking and the intermittent loud clicking. So the seeks are noticeable more subtle. It definitely is some sort of reset noise. Media, head or power related... meaning the problem could be anything. Trying different cables should always be tried before diagnosing it as dead HDD.
True, but (regular) PM is supported. And like I said, due to a glitch in HDDScan (or possibly the HDDs lacking a feature for polling it's current status): current (regular) PM setting is always unknown. To eliminate that possibility I recommend moving the slider slightly and then moving it back to zero = disabled. Then click "Set timer". (Or move it to some place, enable timer with time N, "Set timer", move it back to zero, "Set timer". This should not be necessary, but since current value is not shown, who knows.)
AAM probably has nothing to do with it. OP mentioned there's a steep difference between seeking and the intermittent loud clicking. So the seeks are noticeable more subtle. It definitely is some sort of reset noise. Media, head or power related... meaning the problem could be anything. Trying different cables should always be tried before diagnosing it as dead HDD.