All produced HDDs since around 1989-1990 have linear (rotating around axis due to magnetic field) HSA actuator. Such structure allows achieving faster positioning speed than a discrete mechanism has (with stepper motor), but instead it requires creating servo loop to determine actual angle of actuator’s turn. That is why positioning HSA in HDD is carried out according to recorded to disk servo information. Prepared for operation disk surface contains servo markings for head positioning, so-called low-level format that includes sector markings with its addresses and identifiers, and operating system logic format. Operating system has access only to sector data fields where it organizes its logical architecture. Storage medium’s recording surface contains several thousand (11550 in Quantum FB EX) ‘physical’ cylinders (cylinder is a collection of identically distributed tracks on all surfaces of stack of disks), divided into several areas with different number of sectors per track. When working in a standard user mode the physical format is transformed by the controller into standardized logical with which BIOS and operating system works.
All electronic parts of HDD, except for HSA amplifier and switch chip, are located on the printed circuit board of the storage medium. Controller is a specialized computer consisting of control processing unit, ROM (often built-in in the processor), RAM and chips for controlling motor and actuator coil, data encoding and decoding, data conversion and pairing with external interface. This micro computer operates under the control of its software consisting of several modules and forming a specialized operating system. Part of controller’s operating programs is stored in ROM and another part (in many HDD models) – on disks, in dedicated service area, beyond the reach of a user. Storage medium operating without loading these programs in the controller’s RAM is impossible. Service cylinders also contain another necessary for HDD operation information (table of surface latent defects etc.). Information reading from service area is performed at the beginning of drive initialization as well as on an as-needed basis during its operation.
When powered on the controller automatically lands the heads – moves them to a special landing zone that is not used for recording information, where heads lower to the disk surface. Automatic landing is performed owing to the energy of disk stack rotation and the spindle motor is used as a generator.
Troubleshooting HDD IDE
So, ‘the worst happened’. HDD doesn’t want to start. How to determine the cause hiding behind a malfunction? A general algorithm of primary diagnostics could look the following way:
Connect the drive to power cable; do not connect the interface cable. Upon powering on you should hear the sound of spindle motor cranking followed by actuator (initialization, initial callibration), then the sound of disks should become even and 'select storage medium' indicator (if it is there) should go off. If everything happens this way, then proceed to point 2. Any other behavior indicates about possible malfunction. Options of such behavior could be the following:
а) Nothing happens at all. If the motor is not cranking upon powering on (no sound), then its most probably a problem with printed circuit board. The source of malfunction could be in power supply, motor control, any chip connected with control processing unit and micro controller (on top of everything else the processor controls motor start and stabilizing its rotation speed). Sometimes you can detect failed component visually – burnt out chip due to voltage overload or overheating might have bubbles and cracks. Similar happens most often due to incorrect connection to power supply outlet – mixed up 12 and 5 volts wires or interface connector overload while connecting a storage medium ‘in action’ (when computer power is on). Cases of spindle motor malfunction are very rare (coil rupture), nevertheless the possibility of such case is not zero. You can make sure that motor coils are fine by testing for continuity on the relevant connector. Coil resistance is usually around 2-3 Ω.
b) Disks' spin-up doesn’t happen, but you can hear attempts to spin (buzzing). Such effect may be caused by stuck heads (particularly in old Seagates, WDs, Conners and Quantum Sirocco). In this case you can try and strongly rotate a storage medium (holding it in hands and, naturally, having disconnected all cables). It might help to ‘unstick’ heads. Truth is, this defect can manifest itself later again and you won’t be able to get rid of tendency to sticking heads in old devices for good. Mechanical assemblies tend to irreversibly wear out. You still can witness this in new HDDs, if for some reason automatic parking didn’t work or heads moved from the parking zone for some other reason – for example, due to shacking during transportation.
c) Motor spins-up, then several clicks and it suddenly stops. Possible options:
• Conner drives and related Seagate (ST31276A, ST31277A, ST31722A, ST32122A and co.): motor spins-up, then stops (no clicking). And it does it many times.
• Western Digital drives: motor spins-up and you hear frequent even clicks of actuator hitting the stopper (better turn it off immediately because the reason behind it might be damaged heads and disk surfaces).
• HDDs of other manufacturers usually spin-up the motor, then you hear several clocks and it stops. After that a LED indicator might flicker (if it is there) telling the error code.
Most frequently encountered malfunctions with such manifestation are: head failure, concentric scratches on disks (consequence of wearing out) as well as malfunction of read/write channel chips (often as a result of different kinds of short circuits organized by amateurs who love to tinker about powered on computer). The reason behind the clicks is a head stack assembly backend hitting the stopper because of absence of reading, i.e. malfunction of head stack assembly, read channel or destructed disk servo markings: positioning system cannot find the outer track that contains relevant identification code, and after several attempts in vain the control processor stops the motor. In case with Conner and those Seagates that continue Conner model range, in case of no reading from disks, heads are not moved at all, since their operation algorithm provides for stabilization of spindle motor rotation speed according to servo markings in the landing zone and if servo markings are not found there, it simply doesn’t look for the track.
d) Motor spins-up, then you hear several faint clicks and the motor continues spinning though the drive is not ready (busy indicator doesn’t go off and doesn’t react to calls from computer). Or, it readies, LED indicator goes off but it is not recognized by BIOS and does not respond to commands. It means that the control processor ‘freezes’ due to incorrect reading of service programs stored on disks. These programs could be destructed because of drive failures (inclusive of low-level format incorrect attempts) or because the surface of service cylinders has been worn out.
e) Very loud and disturbing noise upon spin-up (gritting, whistling, howling and so forth). Either the motor is faulty or disks are displaced (due to shock), or the heads moved beyond the disk surface due to stopper malfunction.
In the described above cases the drive is clearly faulty and the degree of its faultiness is quite serious. In the majority of such cases it is impossible to take care of such issues at home. Deeper diagnostics and repair works usually require special equipment (oscilloscope, PC3000 complex etc.). You can localize the malfunction (circuit board or enclosure?) by replacing electronics. Please note, you can replace the board only with an identical one (same model and firmware of microprocessor), if you don’t know for sure that the models are compatible.
Connect the hard drive to computer (to the primary or secondary controller). In case of correct connection of interface cable, the drive should initialize the same way as it is described in point No 1. Then you need to run BIOS Setup and check automatic parameters of HDD (reading drive’s certificate by running ‘identify drive’, see description of ATA standard), a relevant field should display its parameters corresponding to those that are written on the enclosure, or at least similar to it. If it displays something like ‘-93456 cylinders, 0 heads, 274 sectors’, or parameters are not defined at all, then you need to check all contacts and the operability of signal cable. Quite often people find loose contacts in adapters for connecting 2” drives. Better replace such adapter with a normal one because most probably you will not be able to squeeze normal operation and reliability from it. The first pin of 44-pin IDE interface of 2” form factor drive is usually the one closer to the pair of jumper wires master/slave, and the first pin of adapter is the one that is located further from the power cable (view from the side of HDD board).
If all contacts are reliable and computer cable and interface are intact and working, then incorrect parameter definition or absence of such usually indicates the malfunctioning of drive’s electronics (microcontroller chip in particular). Sometimes partial destruct of service information can lead to incorrect reaction to commands, but the likelihood of such case is rather low. Working IDE drive has to be available in the system as a physical device irrespective of logical format. You can check it, for example, with the tool like Diskedit from NU, or by running Checkit test.
Diskedit in working with physical disk allows checking data read/write (write can be checked by filling several unused sectors with different codes on 0 side of 0 cylinder), Checkit checks the readability of sector headers (low-level format check) and correctness of data fields checksums. Usually HDD devices are reliable enough and they rarely have hard-to-detect and unobvious defects. If read/write work on the entire disk space - the drive is nonfaulty. In this case you need to look for the problem in software or incompatibility of some devices present in the system. One of the most common and obvious malfunctions of HDD is emergence of bad blocks on magnetic surfaces – the result of normal wear or careless treatment. You can reveal defects by running tests like Checkit, PcCheck and the like. Do remember that file system check programs (Ndd, Scandisk) take initial information about surface defects from FAT, i.e. clusters marked as bad in FAT (with F7FF code), in reality can be perfectly fine, and vice verse. Delays in operation accompanied by clicks and computer ‘hanging’ when accessing HDD (in case there are no apparent surface defects) reflect an unstable read/write on some areas. i.e. soon emergence of bad blocks.
And one more note: one of the probable causes of drive malfunction is a loose contact in power connector as well as deficient power unit that does not provide necessary stability of power voltage. It is particularly relevant for modern storage media that represent quite precise and tender electronic and mechanical devices and at the same time they consume high currents especially at the moment of stack of disks spin-up. Electricity supply instability can lead to serious damages (just remember a mass failure of Quantum HDDs of new series due to this particular reason). That is why contacts reliability testing and power supply testing has to be the first step taken during the troubleshooting of issues related to HDD.