WD3200AAKS-00VYA0 how to swap ROM from faulty PCB to new

[Nick_CT]

Hi all,

I have a WD3200AAKS-00VYA0 whose PCB has a hole burn't in the smooth chip. I have a compatible donor PCB to use as a replacement.

I've only ever done physical swaps of ROMs where the U12 is present. On this PCB the ROM is in the Marvell chip. So from the little I know about my newly acquired PC3K I thought I would remove the faulty PCB, connect it up to PC3K, read the ROM, then connect up the new donor PCB and write that ROM to it.

(Sure the HDD itself could have a dead pre-amp etc at this point I'm just trying to figure out how to transfer the ROM)

I connected the faulty PCB in kernel mode but PC3K won't read it at all, absolutely nothing. It seems the PCB is completely screwed. PC3K reads the new donor PCB with no problems. I would imagine if I could read the ROM from the faulty PCB it would be a fairly straight-forward process, but I think the PCB is damaged beyond that.

Obviously the HDD with the donor PCB just clicks as it has the wrong ROM info.

Is there something I'm missing that I could do? What could I try in this situation.

I've been fiddling all day, tried the 'ROM recovery' option in PC3K but no luck. If I had the suitable info in my database would this work?

This is what I get from the donor PCB:

Techno mode key
Techno mode key......................... : Ok

ROM:
Flash Dir reading....................... : Ok
Dir info (0B)........................... : 0 (0)
ROM reading............................. : Ok
ROM Data size........................... : 192 Kb
ROM version............................. : 58.7MD
ROM generation.......................... : 58.7MD
Link table version...................... : 01.0F.
Caclulated Version...................... : 0058007M

Heads configuration..................... : by map
Heads number............................ : 4
Heads number in use..................... : 3
Switched off heads...................... : Yes
Head map................................ : 0,2,3
ROM Firmware version.................... : 0058007M



ROM data recovery
Techno mode key......................... : Ok
SA dir reading.......................... : Unknown error. Code(0)


Thanks

[BlackST]

If the Marvell MCU is totally fried , you can't get anything out of it. You need to study some alternatives.

[Nick_CT]

I see it's possible to rebuild the ROM from the SA...but when you can't access the SA that becomes a problem. I'm guessing I need to power the drive up using a loader?

[fzabkar]

I have a suggestion that relies on the Marvell MCU being intact.

Remove the SMOOTH chip or at least disconnect its supply pins. Also disconnect it from the onboard DC-DC converters.

Wire the MCU's Vcore and Vio supplies to the same supplies on your working donor. Then power up the donor board.

Hopefully the patient MCU will now be accessible via PC3K. If not, then you may need to locate its reset pin and manually operate it, as it will normally be under control of the SMOOTH chip. I'm assuming that the donor's DC-DC converters will have sufficent capacity to power both boards.

If the above is feasible, then you may consider a more universal solution. I would modify a cheap car charger for a mobile phone. Many use the MC34063 PWM converter IC. To adapt it for powering an MCU, you would only need to change the values of one or two voltage sensing resistors. If more current is required, then you could add a pass transistor.

Here is a representative circuit:
http://www.eleccircuit.com/wp-content/u ... c34063.jpg

[Nick_CT]

Here's what the PCB looks like, smooth chip went up nicely.

http://tinypic.com/r/15hoapl/7

To be honest I haven't even checked any of the basics on the board yet, like R64/67, D3/4 etc.
So removing the smooth chip alone won't give me access to the rom assuming there are no other faulty components on the board?

Your idea in previous post is good, however I don't think my skills are up to connecting up the MCU's Vcore and Vio on such a small scale.

[Vulcan]

@fzabkar - That's a very interesting idea As you say, some assumptions (hopes) in there, but IMHO certainly worthy of deeper consideration.

@Nick_CT - before you give up on this idea due to the worries about soldering, here are a couple of comments. Obviously this is assuming that no-one gives you any other way of doing what you want to do "officially" (without needing to use the "patient" MCU), and also assuming you don't want to / can't move that MCU. I don't have one of these boards in front of me, so please take my comments with a pinch of salt

So removing the smooth chip alone won't give me access to the rom assuming there are no other faulty components on the board?

Obviously it's fzabkar's idea, so he may have a different view, but IMHO the answer to your question is no. The SMOOTH chip drives 2 external NPN pass transistors that power the MCU - so a dead (or removed) SMOOTH chip == dead MCU. We just have to hope that the (unknown) exact mode of failure didn't also fry the MCU...

To be 100% sure that the SMOOTH chip is no longer providing MCU power, the emitter pins of the external transistors on the problem board, are the place to be measuring and you can then compare those voltages to the donor board (obviously being very careful to avoid your test probes slipping ) Another thing that faulty SMOOTH chip could be doing, is providing power, but also (wrongly) keeping the MCU held in reset - and the MCU would then also not be accessible via the PC3K.

Your idea in previous post is good, however I don't think my skills are up to connecting up the MCU's Vcore and Vio on such a small scale.

It might not be quite as bad as you think, if you were thinking that you'd have to solder to the pads where the SMOOTH chip is/was. IMHO given the catastrophic failure of that chip, its effect (e.g. loading) on all its external connections is highly suspect, so removing the whole chip is the best option - it's doing no good as it is now, and lifting individual pins on such a chip is possible but non-trivial.

The connections from a donor board's two MCU power supplies which fzabkar suggested, onto the faulty board, would likely be possible onto the emitter pins of the NPN pass transistors (probably Q2 & Q3 - but you'd need to check that). As you can see these are larger (and more robust) than the pcb pads where the SMOOTH chip is located, and so easier for you to connect to.

In order to check that the donor board's MCU power supplies could also power the "patient" board, I would actually go a bit deeper and measure the current drain on the donor board of those supplies (though that would need more soldering) and guesstimate the power dissipation in the pass transistors. Using external power supplies for the MCU "patient" (i.e. not from a donor board) would remove that worry about the current capabilities of those parts of the donor board.

The MCU reset signal (driven by SMOOTH chip pin 18) is more of a concern for accessibility, if you don't have steady hands. I use a microscope for this type of work, though when I was younger, my eyesight was better and I could probably have managed without - but not now You'd need to either use that SMOOTH PCB pad, or trace the signal from there towards the MCU, to see if you have any easier places to get to that signal (e.g. a possible TP pad) and inject a low->high pulse (and keep it high, if there is no onboard pull-up, to avoid that signal floating).

Anyway, those are just a few comments from me, as an electronics engineer but without such a board in front of me. If I was in your situation, and needed to gain access to the MCU via SATA, and had run out of other options, then fzabkar's idea is one that I would definitely be considering, perhaps with externally generated power inputs to the "patient" board.

Good luck...

[Nick_CT]

Thank you for all the valuable ideas. In the end it turns out the heads are dead Ah well, was worth a shot.

[fzabkar]

Below is my impression of the Vcore and Vio circuits. There is also a transistor, Q1, that could be a pass transistor for a linear regulator (+1.8V for read channel ???). You'll need to confirm this by measuring its voltages on the donor PCB. The circuit comprising D2, Q2, and L2 appears to be the negative supply for the preamp and should not be required.

I would leave the SMOOTH chip as is, and then connect wires between Vcore and Ground on both boards. Use the copper planes around the capacitors as your connection points. Also connect the emitters of Q4 (Vio). It would probably be a good idea to lift the base pin of Q4 on the patient PCB, and also to remove both its inductors (L).

Code:

                from SMOOTH

                   |
                Q3 |             L
                   |            ___
                   ===       |--UUU--|        R45
                   |||       |       |        ___
  VCC  o--|---|----+V+--|----|       |-------|___|----|---|---o Vcore
          |   |         |    |  ___  |                |   |
          |   |         |    |--UUU--|                |   |
    C??  --- --- C37    -                        C24 --- --- C25
         --- ---        ^ D1     L                   --- ---
          |   |         |                             |   |
          |   |         |                             |   |
          |   |         |                             |   |
  Gnd  o--|---|---------|-----------------------------|---|---o  Gnd


                    +5V
                     o              SMOOTH
                     |     .---------------------.
                     |     |                     |
                     | c   |                     |
           2SD1664   |     |   /|                |
                      \| b |  /+|-------o Vref   |
                   Q4  |---|-<  |                |
                      <|   |  \-|---|            |
                     |  e  |   \|   |            |
                     |     |        |            |
         +3.3V o-----|-----|--------|            |
          Vio              |                     |
                           '---------------------'


(created by AACircuit v1.28.6 beta 04/19/05 http://www.tech-chat.de)


2SD1664, NPN transistor, marking DA, Rohm, 32V, 1A, 0.15V VCE(sat), Q marking = 120-270 hfe, R = 180-390:
http://www.datasheetcatalog.org/datashe ... sd1664.pdf