PCB swap of old Toshiba MK3017GAP

[michael chiklis]

Hi, i have this old Toshiba drive with damaged PCB.

MK3017GAP, HDD2159 Q ZF01 T
PCB number: B36021752017

At the moment i don't have donor pcb for this old drive, but i can find it.
The problem is that i don't know which is the ROM chip of this PCB, does someone know which is?
I can't find any 8 pin chip on the pcb...

Also, i received the drive from my customer without chip (IC2) because he unsoldered it, he sent me IC2 chip in a plastic bag.
It's possible that IC2 contains ROM code?

Here some pics...

[Tomset]

U12 microprocessor with Flash.
if dead - dry oars.

[michael chiklis]

Hi, thank you for your answer Tomset.
I can't find any U12 chip on this pcb...

On this PCB seems that IC4 is the MCU.
Could it be that the rom is inside it?

[data-medics]

U12 is a WD ROM location. I think he's confused.

[pcimage]

I had a virtually identical board the other day on a MK3018GAP, one of the fuses (top left) was blown.

The ROM is masked in the MCU on this model.

So I would check the fuses first and if one is blown then replace it and try and put the removed IC back.

Good luck!

[fzabkar]

Hi, thank you for your answer Tomset.
I can't find any U12 chip on this pcb...

On this PCB seems that IC4 is the MCU.
Could it be that the rom is inside it?

ISTM that IC2 is the main processor. The sticker usually denotes a programmed part, so it would have embedded flash or OTPROM.

The Agere MS452 (IC1) would be the read channel processor and the square Toshiba chip would be a slave processor that handles the IDE I/O.

I'm not certain, but the motor controller at the top left appears to be associated with transistors Q3, Q4, Q13 and a fourth transistor below Q13. I don't know whether these transistors are regulators or motor drivers, but I would measure their voltages before I would replace IC2.

It might be worth asking how the board came to be damaged. If the user inverted the IDE cable, then that would potentially damage any IC that is connected to pins 1 - 4, ie the Toshiba chip.

A datasheet for the RAM would be useful. It would tell you the RAM's supply voltage, and that would help to answer questions in regard to the abovementioned transistors.

Could you tell us the markings on the chips and transistors?

[michael chiklis]

I talked to customer to get more infos about how the pcb got damaged.
He said that native pcb just did smoke while the drive was working, after that drive didn't work anymore.
So he bought donor PCB (which is the one you see in posted pics) and he tried to do IC2 chip swap with his hotair soldering station, it didn't work so he decided to unsolder it again.
The chip you see on pic it's from native PCB!

I look carefully IC2 chip under microscope, i found that some pins were bended and few pads on board where shorted by welding residues, probably this explains why his "repair job" didn't work.
I have also found that one of those 2 fuse near top left corner is opened, very likely because he powered on the drive with shorted IC2 pins or because he had not placed it properly on board.

Fuse.JPG (27.41 KiB) Viewed 12248 times

Now i have removed welding residues from IC2 pads and very carefully straightened chip pins under microscope with the tip of a small screwdriver.
Tomorrow i will try to find a similar fuse to replace the bad one.

I'm not certain, but the motor controller at the top left appears to be associated with transistors Q3, Q4, Q13 and a fourth transistor below Q13. I don't know whether these transistors are regulators or motor drivers, but I would measure their voltages before I would replace IC2.

IC2 at the moment is out of the board, should i measure those voltages now or after i will soldered it on pcb?

[michael chiklis]

RAM (IC21): Hynix 214A HY57V161610DTC-8 KOREA 5W65572U
I don't know about Q3, Q4, Q13 and fourth transistor below Q13, on top of those are printed only numbers, see pics below:

[fzabkar]

I don't believe any test will be conclusive, but the SDRAM is a 3.3V part, so at least one of the transistors should have +5V and +3.3V on its pins.

HY57V161610DTC-8, hynix, 2 Banks x 512K x 16 Bit Synchronous DRAM, 125MHz, 3.3V:
http://www.hynix.com/datasheet/pdf/dram ... 610DTC(4.0).pdf

According to the following pinout, the blown fuse is in the 5V logic circuit, so you would have to be concerned about any chip that gets its power directly from the 5V supply.

http://pinoutsguide.com/HD/Ata44Internal_pinout.shtml

Code:

41   +5VL       +5 VDC (Logic)
42   +5VM       +5 VDC (Motor)
43   GND       Ground
44   /TYPE       Type (0=ATA)

I suspect that IC2 needs a Vcore (+1.2V ?) and a Vio (+3.3V ?) supply. I would look for both of these voltages at the transistors. I believe that +3.3V should be present even when IC2 is absent, but Vcore might require IC2.

That said, it appears that pcimage's experience is that these fuses sometimes fail due to "fatigue" rather than a genuine fault.

[michael chiklis]

I soldered IC2 on PCB and did some voltage measurement:

[fzabkar]

ISTM that the fuse failure may have been due to fatigue, as pcimage's experience would suggest.

Q3 and Q4 appear to be connected in parallel. They constitute the 3.3V Vio supply.

Similarly, Q13 and its companion also appear to be connected in parallel. They constitute the 1.2V Vcore supply.

You can determine the current draw from the 3.3V supply by measuring the voltage drops across the 1.2 ohm resistors. It appears that the voltages at either end of each resistor are 3.38V and 3.25V, in which case the current draw would be ...

(3.38 - 3.25) / 1.2 x 2 x 1000 = 217 mA

That seems reasonable.

Similarly, the current draw from the 1.2V supply could be determined by measuring the voltage drops across the 1.0 ohm (1R0) resistors.

In short, I can't see any reason for the fuse to have blown other than old age.

BTW, 4.1V should appear on the base pins of Q3 and Q4, ie the top pins, not the centre pins.

[fzabkar]

I think this is how the power supplies are wired:

Code:

           +5V
            o
            |
        +---+----+
        |        |
       .-.      .-.
   1.2 | |      | | 1.2 ohm
       | |      | |
       '-'      '-'
        |        |
        |        |
         \|   Q4  \|
       Q3 |--------|----o from motor controller
         <|       <|
        |        |
        |        |
       .-.      .-.
   1.2 | |      | | 1.2 ohm
       | |      | |
       '-'      '-'
        |        |
        +---+----+
            |
            |
            o
           Vio
           3.3V

[michael chiklis]

BTW, 4.1V should appear on the base pins of Q3 and Q4, ie the top pins, not the centre pins.

You're always right
I mistakenly reported the red line on Q3 and Q4

Similarly, the current draw from the 1.2V supply could be determined by measuring the voltage drops across the 1.0 ohm (1R0) resistors.

Voltage it goes up and down from 1.81V to 1.82V on both side of 1R0, so i can't tell the current here.
Something wrong?

P.S.
The drive still doesn't spin up

[michael chiklis]

Something has changed on Q13 and its companion...
Now i don't get 1.18V but 1.81V.

Maybe i just inverted number 1 with 8 before

[fzabkar]

Vcore voltages have been trending downwards over the years. Current drives seem to be around 0.95V. WD's older Scorpios commonly used 1.2V/3.3V Vcore/Vio combinations. I suspect that 1.8V might be OK for your 2002 model.

You could measure the feedback voltage at the junction of the two sense resistors (Rx and Ry). If the regulator is working correctly, the feedback voltage should be equal to Vref. Vref is the reference voltage inside the motor controller. Typical values are 0.8V for modern designs and 1.25V for older designs.

Code:

                 Vcore
                   o
                   |
                  .-.
                  | | Rx
                  | |
                  '-'
      /|           |
     /+|-----------+ Vfback
  --<  |           |
     \-|-- Vref   .-.
      \|          | | Ry
                  | |
                  '-'
                   |
                  ===
                  GND

Attachments

[fzabkar]

Similarly, the current draw from the 1.2V supply could be determined by measuring the voltage drops across the 1.0 ohm (1R0) resistors.

Voltage it goes up and down from 1.81V to 1.82V on both side of 1R0, so i can't tell the current here. :?

Just measure the voltage across each 1R0 resistor and apply Ohm's Law. That should tell you the current draw. If you measure 0V, then that would suggest that the MCU is open circuit.

[michael chiklis]

Just measure the voltage across each 1R0 resistor and apply Ohm's Law. That should tell you the current draw. If you measure 0V, then that would suggest that the MCU is open circuit.

I get 0V across each 1R0 resistor, i guess that MCU got damaged from customer... or maybe not (i don't know).

You could measure the feedback voltage at the junction of the two sense resistors (Rx and Ry). If the regulator is working correctly, the feedback voltage should be equal to Vref. Vref is the reference voltage inside the motor controller. Typical values are 0.8V for modern designs and 1.25V for older designs.

Vref= 1.58V
Vcore= 1.8V

[fzabkar]

It appears that I was wrong about the MCU being a split-rail type. In fact I'm told that it is an MB90F245.

MB90F245, FUJITSU SEMICONDUCTOR, F²MC-16F, 16-BIT MICROCONTROLLER, 3.3V, 192KB flash:
http://www.datasheetarchive.com/dl/Data ... 023283.pdf

This MCU has a single Vcc of 3.3V, ie the core and I/O voltages are the same. This means that the 1.8V (?) supply is used for some other purpose. In the absence of comparative measurements from a working PCB, ISTM that the absence of any current draw from this supply would suggest that ...

1) its load is open circuit
2) or its load is in standby mode
3) or its load is in the HDA, ie the preamp.