Problem with WD Marvel Demo

[Spildit]

Hi Spildit,

Thanks for your detailed response and the provided link. I'll read through and see if i can make sense of it all. If i'm able to come up with a solution i'll be sure to let you know.

Thanks again for your help.

Ok !!!

Good luck and let us know of your progress !

Also you can buy a full version of WDMarvel and play with it !!!
Remember the DEMO will not write anything to the drive.

Regards.

[Severence]

OK, bit of an update on this.

I think the drive is likely well beyond repair at this point and I'm sure in the process of trying to fix it i have made it far worse.... but hey, it was either play around and give it a go, or throw it away, as i don't have the money or inclination for professional data recovery. in any event hopefully the following info will prove useful for somebody - although some of this may be common knowledge for those in the know - it definitely wasn't for me, nor could i find much info online in regards to some of the problems i was having. bear in mind allot of what i say from here on could be factually wrong so please don't take it for gospel, this really isn't my area.

Drive Details
Drive: WD Caviar Green 2TB
Model: WD20EADS-22R6B0
Board: 2060-771642-001 Rev-P1
MCU: 88i8945-TFJ2
DCM: HARCNV2ABB

OK so after the issues mentioned earlier in the thread i came to the conclusion that it may be a ROM problem, more specifically a translator problem, which as far as i understand is a module Incorporated in ROM - although i dumped the ROM with WDMarvel (and various other programs) and they all matched the original dump taken at the very beginning, i figured it there was no harm in re-writing the rom to the drive...maybe the MCU has a parallel flash alongside the External Rom chip and maybe theres something wrong with one or the other and re-writing may solve the problem i thought...... So all lights are green and i write the ROM file, the write has been successful apparently...

Ok but, after this the drive basically acted as if it wasn't even connected at all... in pio mode or otherwise....shit. next thought was I'll check the external Rom and see whats up, checked it with my programmer, and the ROM is completely blank...what the. so i manually write the ROM dumped by WDMarvel to the external ROM chip with my programmer - i have high hopes at this point that at the very least it'll take me back to square one, but alas... nada, nothing, zip.

Ok, so my next train of thought was maybe the external Rom chip was blank all along and the drive uses an exclusive internal Rom, so now comes the question of how do i write to the internal ROM without PIO drive access. Some searching points to a program U-Command which writes the ROM via UART adaptor - again i appreciate this is common knowledge, but what wasn't was which pins enable UART "X-modem" mode on my specific drive. I found some pins (details on the pins a bit later) which seem to enable UART terminal access, and i could read and write address's if i so wished but either the drive still isn't in the so called "X-Modem" state or U-command simply isn't compatible with my particular drive/board. there very well maybe a specific command which can be issued via terminal to put the drive into a special state but hell if i know it and i could find barely any info elsewhere, I'd also guess that you could manually write ROM/modules provided you know there addresses/locations but again... hell if i know.

What now... i tried the common Kernel mode by shorting the known pins on the 8pin header and got nothing. hmmm, i got to thinking that maybe there is something wrong with original ROM dump made by WDMarvel and even the subsequent dumps, so i manually checked them all with a hex editor and sure enough they're all the same. Now i download some ROM files from near identical drives and notice that in comparison my ROM file seems to be missing an awful lot of information - specifically the data which is comparable to mod 0A,0B,0B,47 - in my Roms flash file these areas are completely empty. i have no idea if this is normal or not..... luckily? i had taken a backup with another program which i completely forgot about until now... the Sediv trial, not only does the Rom file appear to have all needed data in reference to the previous mentioned modules but Sediv also backed up the modules independently - which is great as WDMarvel didn't - again i don't know if this is normal behaviour for the program or what, seems strange to me though.

Right then, i took the Rom file dumped with sediv and again manually flashed external Rom with my programmer and once again i had high hopes I'd at the very least end back up at square one, but unfortunately no cigar, no PIO/Kernel/U-command access or otherwise, right - maybe kernel mode isn't defined by just some of the pins on the header but maybe i have to enable/short another point on the board to give me true ATA kernel access, so as far as i can tell this does in fact seem to be the case and i can once again get access to the drive over sata with the following pins/points shorted but there appears to limitations depending on which "mode" your in, i have no idea the technical terms for these modes or if they're even considered modes at all.... who knows.

So Here's a very crude image based on an even cruder drawing i made while i was trying to figure this all out, you can see I've drew specific areas of interest which are generally unpopulated resistors and I've drawn it a manner in which it kinda aligns to the actual orientation of components on the board, As for the MCU pin numbers - they are likely not in the right order as there are two pin one "Spot" indicators on the chip - one big and one small, and i had to basically guess as to which one was actually indicating pin 1 (as there is no data sheet) irregardless just pretend my MCU pin numbering is correct as it is all relative to the orientation of the writing on the chip together with component location, spots etc.

UART Terminal access
TP-E62 > short > 3V3
TP-E6 > 10K > 3v3

"Kernel Mode"
Short the relevent pins on 8-Pin 2.54 Header As well as TP-E62 > short > 3V3
In this mode and with my drive in it's current condition the drive will now respond to reset commands and indicate but not much else.

Same as above but without need to use 8-Pin 2.54 Header pins
TP-E61 > 10K > 3V3
TP-E62 > short > 3V3

semi detect... i don't know name mode
JTP36 > 22K >3V3
JTP20 > short > GND
TP-E61 > short > 3V3
In this mode the drive is detected by WDmarvel, albeit incorrectly - showing giberish in the info boxes, it claims it has succesfully read the rom but upon checking the files is approx 16mb in size (instead of 192kb) it errors when attempting to write the Rom, when attempting to load overlays it errors "incorrect buffer size" when reading modules the data is filled with the same giberish Y (like) character shown in the drives info boxes etc. with WDR pro tool the drive identifies also but unlike WDmarvel it reports an LBA size which is interesting although i beleive it's wrong, also it will write the Rom (and independent rom modules) without error but it doesn't seem to stick. should be noted hower that this "mode" didn't seem to to work when i used WDmarvels incomplete Rom dump on the external Rom chip, it had to be the complete Rom dumped by sediv.

I tried a multitude of other things to no avail, like i said earlier i am pretty sure i've made a fair few mistakes here along the way but i do hope this info will be even a tiny bit useful. it's like i won't be completely satisfied until this drive is completely dead

[Peter_M]

UART Terminal access
TP-E62 > short > 3V3
TP-E6 > 10K > 3v3

I have the same CPU 88i8945 and this not work checked 5 min ago.
Are you sure that terminal work with they configuration?

[Severence]

Yes, provided you have your UART adatper hooked up to the correct tx, rx, gnd on the 8-pin 2.54mm header. i tested with realterm & hyperterminal - baud rate should be set to 115200 - to check for reception, ensure software is open etc, sata is plugged in and the relevant cables are connected - apply power to the drive ... you should get a response, if not try reading an address > r00000000

If it still doesn't work, check that your test points are going to the same unpopulated resistor points marked in the img and in turn the MCU pins - if your board is a different model/revision to mine then theres a chance that the test points/resitors numbers could be different, so you you'll want to cross reference them with the the MCU pins illustrated in the img.

If you still have no luck i can take a snap of my hookups on the board aswell as a screen cap of the terminal.

[Peter_M]

Yes, provided you have your UART adatper hooked up to the correct tx, rx, gnd on the 8-pin 2.54mm header. i tested with realterm & hyperterminal - baud rate should be set to 115200 - to check for reception, ensure software is open etc, sata is plugged in and the relevant cables are connected - apply power to the drive ... you should get a response, if not try reading an address > r00000000

If it still doesn't work, check that your test points are going to the same unpopulated resistor points marked in the img and in turn the MCU pins - if your board is a different model/revision to mine then theres a chance that the test points/resitors numbers could be different, so you you'll want to cross reference them with the the MCU pins illustrated in the img.

If you still have no luck i can take a snap of my hookups on the board aswell as a screen cap of the terminal.

Well, my PCB is 2060-701640-001 REV A from the same MCU 88i8945 TFJ2, and my CPU have BAD internal ROM, but work with J-tag interface, and trying start with External Flash. All attempts to Boot from external Flash, the terminal failed because the CPU does not respond.
Your suggestion also not work, and annalize your pic witch config resistors, looks like that overlap in large part, and Test points also.
So I Soldered R60 - 10K, and R75 - 30Ohm
And Terminal not work with puttly, checked with 2 terminals, Bauthrate is 115200 8N1
Here is my PCB:

[Severence]

Ok, looking at your photo it looks like R75 is going to pin-19 (or close to it) on the MCU, it impossible for me to tell though based purely on the image your provided, if this is the case then UART/terminal is obviously not going to work.

On my board R75 goes to pin-23 of MCU (check my img)

Put your multimeter in continuity mode, put one probe on pin 23 (based on my img) and find and confirm which pad it's going to. do the same with pin 37 on the MCU (again based on my image) although again looking at the image this time R60 does in fact look like it's hooked up the same as my board.

Also, don't use a 30ohm resistor - use a 0-ohm resistor, a 1-ohm resistor or just jump with a wire. although 30-ohm is pretty low and probably absolutely fine to use, it's best to rule this difference out.

What is your end goal here? are you just experimenting? like i said in my previous post, i was not able to find a solution so.....

Also, are you working with the board uninstalled/not connected to the motor etc? in my case everything was installed as normal.

Also please give me details on your UART adapter, and rule out that that is indeed working as it should by doing a Tx/Rx loop test in terminal - you also might want to rule out a software issue and try realterm and/or hyperterminal.

[Peter_M]

Ok, looking at your photo it looks like R75 is going to pin-19 (or close to it) on the MCU, it impossible for me to tell though based purely on the image your provided, if this is the case then UART/terminal is obviously not going to work.

On my board R75 goes to pin-23 of MCU (check my img)

Well, i analized your pic in my PCB, and Connections of resistors are the same like R60 connect to pin 37 pin, R75 also connect to pin 23,

Also, don't use a 30ohm resistor - use a 0-ohm resistor, a 1-ohm resistor or just jump with a wire. although 30-ohm is pretty low and probably absolutely fine to use, it's best to rule this difference out.

Ok i will try. So i trying solder R75 0ohm and also terminal not respond.

Put your multimeter in continuity mode, put one probe on pin 23 (based on my img) and find and confirm which pad it's going to. do the same with pin 37 on the MCU (again based on my image) although again looking at the image this time R60 does in fact look like it's hooked up the same as my board.

Connection pins are the same with my PCB.

What is your end goal here? are you just experimenting? like i said in my previous post, i was not able to find a solution so.....

well i trying to run my PCB in internal ROM, because in external ROM my PCb not work thats all

Also, are you working with the board uninstalled/not connected to the motor etc? in my case everything was installed as normal.

differently, first uninstalled , not connected to motor, second installed with disk.

Also please give me details on your UART adapter, and rule out that that is indeed working as it should by doing a Tx/Rx loop test in terminal - you also might want to rule out a software issue and try realterm and/or hyperterminal.

My Terminals are homemade building
First terminal are build in MAX232 TTL converter it is standard application 3,3Vio
Second terminal are also bulding on PL2303HX TTL USB TO TTL CONVERTER 2,5Vio, and also standard application.

[Severence]

Ok, did you do a loop test with your TTL converter and observe the output in terminal?

[Peter_M]

Yes, and if i short RX-TX pins then i can write symbols, and these terminals working fine with seagate drives.

[Severence]

hmm, i don't know then.

Maybe there is a third config resistor installed by default on my board which is not installed on yours... I'd suggest looking at images of my board model on google and seeing if your can spot a resistor which is missing on your board...that could explain it - in my picture i have drawn pretty much all unpopulated resitors (give or take a few which didn't seem relevent)... so i would check your board to see if there are any more unpopulated resistors which aren't in my drawing, this might be key.

Double check that your getting approx 3V on the previously mentioned resitors with your meter, if not source the voltage directly from the regulator on the board.

Provide me with a photo on how your connecting up your TLL adapter to the board - all connections.

In the software terminal what is your output exactly, is it blank or $ or >
when issuing commands do you see Tx light up?

try issuing > r00000010 or > S or > s

Have you got a valid rom file on the external rom chip? i can't remember if this is neccesary for communication, but it'd be worth checking - if you have, please upload it and ill check it in hex editor and see if it's valid (based on the near zero knowledge i have on the subject)

How are you powering the drive? regardless i'd suggest using the standard ATX connector, also it might be worth checking that voltage is arriving at your MCU, presumably the MCU is powered from the 3V regulator (although i haven't checked) so check for continuity on the regs output pin to a pin (i don't know which) on the MCU, although it could be going via a protection diode in which case you'd have to check that insteasd...

Other than that though, and failing all other options i'd say your MCU is shot.

Do you get any results in WDMarvel software (or other) via SATA when trying my other previously mentioned resistor set "modes" ?

Edit*

Ok, was just having a quick look at the img of your board that you provided and i noticed that R106 is populated .... if this is indeed the case on your actual board then your going to have to desolder it, aswell as any other resistor which is populated which in comparison to my img isn't.

This is important as i found you can't set multiple "modes" it just simply won't work.

on a side note, i was looking for an alternative to u-command tool and found a chinese tool by a company called qtaosoft "Q3 drive repair tool" i think i downloaded a trial copy, it has options for seagate, western digital etc but unfortunately in the version i got, it seems to be limited to seagate drives only, with the WD option grayed out, i went to thier homepage to see if i could grab a more recent version and although they have a download page the site seems broken or is under maintenence. Does anybody have any info on this tool or a more recent version download? be interested to see if the software supports more boards over UART compared to U-command.