- how to locate the basic test points of an unfamiliar HDD PCB
This tutorial in intended to show how one may determine the test points for the onboard DC-DC converters, spindle motor, VCM, and preamp supply rails of an unfamiliar HDD PCB. We also identify any protection devices. I have done as much as possible in the absence of power in order to simulate a non-functioning PCB.
Our example will be a Samsung SP0411N IDE PCB (PANGO REV06).
It is assumed that the reader understands how to use a digital multimeter.
(a) Date Codes
These can be observed on the drive's label and on the various ICs and semiconductors. The latter are most often in YWW or YYWW (Year / Week) format.
In our example the most obvious date codes are ...
0320 0324 0327 K323 Y325
This suggests that the drive was manufactured during or after the 27th week of 2003, ie some time in early July.
(b) SDRAM and EEPROM
These ICS are two of the easiest to identify. Their datasheets should be easy to locate, and they will tell us the supply voltages and supply pins of each chip.
In fact the SDRAM (IC42S16100-6T) requires a 3.3V supply on pins 1 and 25, and the EEPROM (M29F102BB) is a 5V device.
Locating the DC-DC converters
DC-DC converters are of two types, either linear or switchmode.
Linear supplies will either have a single regulator IC, or a pass transistor that is driven and monitored by the motor controller IC.
Switchmode DC-DC converters will be identified by an inductor, either a coil or a 2-pin rectangular package. In the vicinity of the inductor will be a MOSFET and Schottky diode. Sometimes the MOSFET and diode are combined in a single FETKY package. The MOSFET is usually controlled by pulses from the motor controller (Pulse Width Modulation), but some designs use a separate PWM controller IC.
In our case the PCB has no coils, so there don't appear to be any switchmode supplies. This is not surprising since switchmode converters weren't regularly used in HDD applications of that time.
Instead we need to focus on the larger discrete semiconductors, specifically the 3-pin parts marked LT25, X2B, DA and CF. These would be power devices of some kind. Our initial guess would be that these are transistors (NPN or PNP), or MOSFETs (N-channel or P-channel), or 3-terminal LDO (low drop-out) regulators.
In fact the marking code of the LT25 device would suggest that it may be a 2.5V LDO regulator (we have already determined that the K323 marking is a date code). A reasonable assumption would be that this IC may supply one of the voltages for the Marvell 88i6522-LGO MCU. LDOs have an input pin, an output pin, and a ground pin. Using the 4-pin Molex power connector as our reference, a quick continuity test with a multimeter confirms that the uppermost pin connects to +5V and the lowermost to Ground. This would suggest that they are input and ground pins, with the middle pin being the +2.5V output. However, contrary to our initial guess, the +2.5V pin does not connect to the MCU.
We now need to determine which device provides the +3.3V for the SDRAM. This is done by testing for continuity between the SDRAM's supply pins and each of the pins of the DA, CF, and X2B devices. This time we find that the leftmost pin of DA is the +3.3V supply, and its centre pin is +5V which makes it the input. The +3.3V supply also powers the MCU and is probably its Vio rail.
The CF device now looks like it may be a third regulator. If we assume that its middle pin is an input, then we find that it connects to the centre (output) pin of LT25. The lowermost pin of CF connects to a heavy copper trace, so it would probably be an output. This means that CF down-regulates the +2.5V supply to yet another supply rail. More testing reveals that this lower voltage powers the MCU via numerous pins, and is probably its Vcore rail. There appears to be no continuity between any pin of the MCU and the +5V input.
Datasheets confirm that CF and DA are in fact NPN transistors, and their pinouts are BCE (Base / Collector / Emitter). My first guess would be that the bases of both CF and DA are controlled by the HA13627, but CF is in fact driven from the MCU. This would now suggest that CF may not be a regulator as originally thought. Instead it may just be a switch, and +2.5V may in fact be the Vcore supply.
An actual measurement confirms that the output of the LT25 device is indeed +2.48V, and Vcore is +1.37V. Therefore CF is a pass transistor under the control of the MCU.
+5V o----|LT25|----+------o +2.5V
|____| C| +2.5V regulator & Vcore
| \| B
| CF |------o from MCU pin #24
| +3.3V Vio regulator
DA |------o from HA13627 pin #26
Connected directly in series with the PCB's +5V input is a 2-pin device with a marking that looks like "1X1", the second "1" being upside down. Its resistance measures around 0.15 ohms. This would suggest that it is some kind of protection device, possibly a 1A fuse. In fact it is actually a PolySwitch, a device which behaves like a resettable fuse.
The centre pin of the X2B device connects to +5V and the leftmost pin to Ground. One's first impression might be that it is an LDO. In this case the rightmost pin would be the output, but then one would have to ask why there is a 102 (= 1K) resistor between output and ground. In fact X2B is not an LDO, but an SCR. It's function is to crowbar the +5V input to ground in the event of an overvoltage. That is, when the input voltage exceeds the breakdown voltage of the zener diode (C2 marking code), the gate of the SCR is triggered on, causing the SCR to latch into conduction, and resulting in a short-circuit across the +5V input. The polyswitch then goes open circuit and effectively disconnects the supply. When the overvoltage condition is removed and the drive is power cycled, the polyswitch should recover and the drive should continue working.
| z Zener
X2B V A Diode
SCR - G | Overvoltage crowbar circuit
| | |
| .-. |C
| | | ---
| R| | ---
| '-' |
| | |
Spindle Motor test points
The spindle motor has 4 terminals -- 3 phases plus a common. The phase-to-phase resistance is 3.8 ohms, and phase-to-common is 1.9 ohms.
A simple continuity test identifies the motor test points adjacent to the HA13627 motor controller IC.
Pin #64 = common
Pin #63 = phase A
Pin #61 = phase B
Pin #59 = phase C
Common (Pin #64)
| | |
| | |
C| C| C|
1.9 ohms C| C| C| Spindle Motor windings
C| C| C|
| | |
| | |
o o o
A B C
Pin #63 Pin #61 Pin #59
||<- Q1 Q3 ->||
| | VCM H-bridge
HA13627 pin#51 o----)-------+ | (internal to HA13627)
| ___ | |
HA13627 pin#48 o---+--|___|-+--o VCM o--+---o HA13627 pin#46
| current |
| sense |
||<- Q2 Q4 ->||
HDA Connector test points
The VCM pins have already been identified. I have chosen to number them as 19 and 20.
Most ground pins are obvious by inspection, but pin 4 needs a continuity test.
The Read and Write differential pairs are also fairly obvious. However, I can't tell which is which.
I believe pins 10, 11, and 12 are the serial data, clock, and enable pins, although not necessarily in that order. I say this because pins 10 and 11 both connect to the same 82 ohm resistor (820) and another 8-pin device, probably a capacitor array.
The preamp's +5V supply is cleaned up by an LC filter (serial inductor plus parallel capacitor). A similar low pass filter connects to pin #2. It is for this reason that I suspect pin #2 to be a second supply rail, and I've called it Vcc2. That said, I confess that I'm having difficulty understanding this circuit. A diode test on my DMM suggests that the two 3-pin devices marked "P" and "N" may be dual Schottky diodes. The "P" device has a common anode, and the "N" device has a common cathode. The cathode of N is grounded.
An actual measurement (-5.0V) confirms that "Vcc2" is the second supply rail.
+5V o----UUU---+----o +5V supply Low Pass LC filter
| to preamp
capacitor C ---
| -5.0V |
| o |
| | |
||-+ +----+----+ +-||
Q1 ||<- | | | ->|| Q2
--||-+ P1 | | | P2 +-||--
| V | V |
| C1 - | - C2 | -5V DC-DC converter
| R1 | |C | R2 |
| ___ || | --- | || ___ |
PWM1 o---+--|___|--||---+ --- +---||--|___|--+---o PWM2 (HA13627 pin#40)
(HA13627 pin#41) | 3R3 || | | | || 3R3 |
| V | V |
| - | - |
||-+ N1 | | | N2 +-||
Q3 ||<- | | | ->|| Q4
--||-+ +----+----+ +-||--
| | |
| | |
| | |
| | |
| | |
The -5V negative supply for the preamp appears to be a "regulated charge-pump inverter" topology. I don't have a datasheet for the HA13627, but I'm guessing that its internal circuit could be represented by an H-bridge. Contrary to my initial guess, this circuit is actually a switchmode power supply. It works by charging capacitors C1 and C2 via Q1 and Q2, respectively, during alternate half cycles. The charge on each capacitor is then transferred to capacitor C on opposing half cycles via Q3 and Q4.
IC42S16100-6T, 512K x 16 Bit x 2 Banks (16-MBIT) SDRAM, 3.3V, Integrated Circuit Solution Inc.:
M29F102BB, EEPROM, 1 Mbit (64Kb x16), 5V, ST Microelectronics:
PC peripheral power connector pinout:
2SD1664, NPN transistor, marking DA, Rohm, 32V, 1A, 0.15V VCE(sat), Q marking = 120-270 hfe, R = 180-390:
2SD2150, Rohm, NPN Low Frequency Transistor, 20V, 3A, 0.2V VCE(sat), marking CF, R = 180-390 hfe:
X0202BN, ST Microelectronics, marking X2B, SENSITIVE GATE SCR, 1.4A:
PolySwitch Resettable Devices, Raychem Circuit Protection:
ASCII circuit diagrams created by AACircuit v1.28.6 beta 04/19/05:
HDD IC Database: