Hi...I'm still working on my hardware approach to fixing my Maxtor 5T060H6. Before I say any more, I'd like to ask for help in finding firmware for that drive. Here is the data from my drive:

Model: 5T060H6
HDA: 26A
PCBA: 04A
Unique: 11A
Code: TAH71DP0
Letters: K,M,D,B
Code by 40-pin connector: B4DTA
Code on barcode by serial #: DT26A

I want to thank everyone who responded to my post regarding a hardware problem. I received a lot of good information and certain people have helped me an awful lot. You know who you are...thanks.

I have traced 90% of my drive electronics board. When I finish it, I don't mind sharing it with anyone who is interested. The processor on the board, the Texas Instruments D741667APGF, was made especially for Maxtor. I have a request in to Maxtor for information on that chip, and others, and for the firmware for the drive. They are considering my request but are taking their time responding. At least they ar considering it.

The spindle motor and the voice coil motor are both controlled by the SH6770C chip. The spindle motor is 3-phase and controlled directly by Mosfets in a push-pull arrangement. The gates on the Mosfets are driven by the SH6770C. The SH6770C is driven directly by the processor but it has input from the Agere MS453, which I am guessing by default, is the encoder/decoder for the read/write preamplifier chip in the sealed head chamber.

The voice coil motor is driven by its own Mosfet, and the gates on that Mosfet are driven by the SH6770C. There is a 10-pin chip involved with the SH6770C but I can't identify it (A171F).

There is a neat little DCDC converter that converts +5volts to -5volts. The -5 volts is fed to the preamplifier chip. I'm guessing they are using that for the RLL encoding/decoding for the data written to the drive. I have checked all the voltages, +5, +12, +3.3 and -5, and they are all good.

I have put my hardware approach on hold temporarily while I investigate the firmware issue. At the same time, I am reading a lot on hard drive theory because I need it for the hardware investigation. The Maxtor board is well marked with signal information but I did not recognize a lot of it. It's making a lot more sense now that I understand better what is going on.

I have narrowed the problem on my drive down to two issues: firmware in the service area or either the preamplifier chip or the heads themselves. I realize that is pretty general, but I'm going on the drive's behavior. When I boot it, the drive starts from a position next to the hub on the platters. When it is addressed, it seeks to mid-platter and goes back to a new position about 1/4 inch from the central hub. That means it is taking instructions from the main processor.

If I leave it in safe-mode, it will sit in the position 1/4 inch from the hub. If a reset is issued, it starts to seek back and forth between the 1/4 inch mark and mid-platter. If I boot it while not in safe-mode, it oscillates back and forth between the 1/4 inch position and mid-platter. I put that down to it not being able to read the firmware or to a problem in the heads themselves. Of course, it could still be the preamplifier.

Anyway, the search goes on.

Hello OldTech,

I think that input of the MS453 is the Reset input. the Power On Reset (-POR on the board) is generated by SH6770 and fed to the processor and other components needing to be reset, in this case only the processor and MS453 (and of course the SH6770 itself).
U may check that interconnection to the -POR signal on the board.
And also if U pull down -POR the drive will be reset.

regards,
pepe

thanks for the tip Pepe. I always thought of a power-on reset as being part of the processor. You got me thinking (which can be dangerous) and I checked. The connection I thought went to the MS453 is on pin 60 and it connects to the SH6670 at pin 25. This is actually the +3.3 volt feed. Pin 27 on the SH6770 goes to -POR and that connects to the RAM chip at its pin 34 which is CKE (clock enable).

The CKE signal, when de-activated, puts the SDRAM into power down, suspend or self-refresh. It's an active high signal normally, so it seems to require a negative signal to de-activate it. I think that's why the POR is marked -POR. So, the actual power-on reset must be another signal.

Since pin 27 is related to POR, however, I'm guessing adjacent pins are too. Pin 26 goes to +5 volts and might be the POR source. I don't think they would use +3.3 as a POR source since it is derived from +5v through a regulator. It seems the SH6770 runs off +3.3v and the +5 is used to determine POR.

There is a RESET pin on the processor...pin 49 or so....and it comes in on the 40 pin IDE connector. I'd have to look, but I imagine that reset is a software reset.

It's tough guessing all this without the actual device pinouts.

Introduce friend....

For the first time, i was surprised read your quotation at this forum.
A WONDERFULL QUESTION MAN....
Could you please, are a RECOVERY AGENT or HDD ENGINEER???
For this question, i need your answer....

But, i'm so sorry for this time i can't answer your question in this forum.
For a while i don't enough time for to solve your problem, i was busy to recover data from my customer.

Maybe next week i can give my idea to solve your problem, but for the first i must study over and open my maxtor paper about MAXTOR RIGEL as you have.

Best regards,
Along

_________________
HDD is a COMPUTER...

Pepe...did some more checking on the Reset. Maxtor's manual on the Rigel series (5T0x0Hx) says the following:

"Reset Handling...One of three different conditions may cause a reset: power on, hardware reset or software reset. All three cause the interface processor to initialize itself and the Task File registers of the interface. A reset also causes a set of the Busy bit in the Status register. The Busy bit does not clear until the reset clears and the drive completes initialization. Completion of a reset operation does not generate a host interrupt".

The software Reset comes in on pin 1 of the 40/80 pin IDE cable and goes straight to the processor. A hardware Reset, to me, is one of those momentary pushbuttons you see on some computers. I don't see where there would be a hardware Reset on a hard drive. Of course, if you put your own pushbutton between pins 1 & 2 of the IDE connector, that would cause a hardware reset. Apparently you can test the processor using that method, by reading the ALE line on the IDE connector right after a hardware reset. If the clock crystal is good, and a reset is issued, the processor should send out signals on th ALE line.

The power-on reset, from what they say above 'causes' the processor to initialize itself. The spindle motor can't spin up till the processor tells it to and the processor tells the SH6770C when to start the spindle motor. It does that through a serial bus, if I understand correctly. Also, all the devices have to work in a synchronized manner, and normally its the processor that is in charge of the synchronization. The processor would have to have its own power on reset in order to initialize itself.

I'm still looking for a drawing of the SH6770C and /or all the other chips I have mentioned. I have one for an Athena DSP, which uses the same basic chips as the Rigel, but it isn't marked as to which pin does what. If anyone has some data on the SH6770C, the D741667APGF, the Agere MS453 or the Agere pramplifier 541008D6/63109741, I'd appreciate you posting it or sending them to me. There also a little chip with the lettering A171F I'm trying to figure out.

I notice a lot of people read the posts but hardly anybody responds. If it's a language thing, don't worry. Write in your own language and tell me what it is so I can use an online translator.

ciao

How's it going, friend??

nothing that impressive...not even an FBI agent. I'm just an electronics technician who has worked as a computer technician in the past. I'm also an electrician, but don't tell anybody. I don't work in the data recovery business either.



since I wrote that post, I've done some more testing. I measured all the read/write heads with an ohmmetter. The heads are fed in pairs from the preamplifier starting on pin 14. Pins 14/15 and 16/17 go to the top head on the top platter. One is probably a read head and the other a write head. One is about 17 ohms and the other about 50 ohms. All the heads read good. The voice coil itself is about 12 ohms on my unit. All the measurements were done with an accurate DVM.

Pins 18/19 and 20/21 feed the next head down, under the bottom side of the top platter. The other heads are fed in order from pins 22 to to 38. All the other pins are fed from the ribbon cable attached to the head actuator where the preamplifier sits.

Here's a tip for anyone trying to use a voltmeter (DVM) on such a small chip. Connect the meter leads to test leads with alligator clips on them. Take a couple of ordinary sewing needles, about two inches long, and clip them into the alligator clips on the other end of the test leads. Be very careful to have your hands grounded. Keep one hand in contact with the metal case and try to touch metal with the other one as well. Use the sewing needles to touch the tiny pins on the chip.

Some more info. The magnet over the actuator voice coil is stuck on by magnetic force. You can pry it off, but I'd advice anyone to be very careful doing that. Magnets have a way of swinging around suddenly, and if it catches the voice coil, it will probably destroy it. Also, if you put something under the magnet to pry it off, make sure you don't catch the voice call too. And be sure you don't catch the voice coil and swing it off it's parking mechanism. It could scratch the platter surface if it is suddenly moved.

You should actually try all this on a drive you can afford to lose. If you get fingerprints on the platter, you'll need to use 99% isopropyl alcohol to take them off. Also, the platters are covered with a self-lubricating carbon surface which caused dust particles to stick to it. I've had success wiping dust, etc., off with isopropyl alcohol and buffing the surface with a very soft, lint-free cloth. Be sure to use 99% isopropyl alcohol since the other percentage alcohols have water in them.

I would appreciate any help you can give me. I haven't used my oscilloscope yet but I plan to start looking at the signals coming from the preamplifier. I'm getting a better idea of what they should look like.

I'm still looking for a good method to position the heads I'm thinking of QBasic from the old DOS package. I don't know yet if it works like the real Basic language, but if it does, it's possible to write a small program on-screen and execute it in real time. That way, you can make the drive do seeks, or put it in a seek loop while you analyze the signals being generated. Of course, I still have to get past the point the drive is stuck at now, or find a way to override the control the processor has over the heads.

What you're trying to do with this method is pick a head through the program and check its reading ability using an oscilliscope. That way, you can eliminate each head, one at a time, and verify that it's data is getting through the preamplifier chip. At this point we're not worried about writing, We need the heads to be able to read so they can read the service information on the disk. Otherwise, the drive wont finish it's initialization and it wont ever be ready for further testing.

Of course, if the drive wont finish it's initialization, you wont be able to run a program to operate the heads. That's why I want to find a way to bypass the control the microprocessor has over the head actuator. Most of these system are servo-systems, meaning the processor sends out a signal to do something and expects feedback that it's instructions have or have not been carried out. All the spindle and voice coil motor control is done within the processor on Maxtor Rigels. I'm hoping to fool it into thinking everything is cool with the actuator by sending it the signal it wants to hear. Something like what you'd do with your girlfriend, or your wife.

Anyone have any ideas? I don't know the signals well enough yet and which ones to intercept. One thing you defintely don't want to do is interfere with the spindle motor, so any hacking has to leave it alone.

Hi,

regarding the power on situation:
upon power on the supply voltages are not stable for some time. During this time -POR is held low (by a non-controlled circuit in SH6770, I mean non controlled by the processor, of course controlled by supply voltages, etc), disabling every component having such input and attached of course
Then when things settle down the POR circuit releases the -POR line (it is probably an open collector or an open drain output) and things begin to happen in the processor and other chips.

Regarding the HW and SW reset:
HW reset is the pin in the IDE connector labelled RESET.
it doesn't do the same initialization as -POR, but probably has high priority.
SW reset is bit2 in the Device Control Register (0x3F6 for Primary channel).

I hope it is useful info...

regards,
pepe

Old Tech-

For modern hard drives, what type of scope do you suggest? e.g. analog vs. digital, storing vs. not, up to what MHz, etc.?

-brendan

howdy...I received a reply from a fellow and I'm sorry not to remember his name without looking up the old post. He claims you need a storage scope and I understand it needs to have about 100 Mhz bandwidth. Mine is a dual-trace with only 30 Mhz bandwidth. You would need a storage scope if you wanted to analyze the data properly, especially if the heads were seeking a lot.

There are adapters out now that are actually A/D converters and use a PC for storage and analysis. I've seen them listed for a few hundred dollars but did not check the bandwidth. The data coming off a harddrive, right at the head, is analog. But the preamp for that data, which also shapes it, is placed right on the head assembly. By the time it gets to the drive electronics board, it should be a well-formed square wave, but it still wont be in an intelligible format. It has to be decoded from prml, or whatever encoding it had.

So, it would depend on where you're reading the data. The guy who replied about the storage scope was referring to the RDN/RDP test points which are located right where the data comes into the drive electronic board from the preamp. The frequency of that data would also depend on where the heads were reading on the disk. Modern drives use zone bit recording, meaning the outside tracks are stacked with more data, therefore a head would be reading a lot more data on the outside tracks than on the inside tracks. Then again, the inside tracks are turning much faster with respect to angualr velocity.

What I'm thinking, or I should say 'hoping', is that regular patterns in the signal, like index marks, will be recognizible. I'm just trying to verify that the heads are reading more than noise, so I'm hoping to get away with a dual trace analog scope at 30 Mhz. Other than that, I can't give you good advice on what kind of scope you'd need. It would depend on how deeply you want to analyze the data being read.

There should be ways to sync the data using a dual trace scope. If you can sync one channel to the track timing index, for example, it might hold the rest of it to make it visible. Sorry I can't give you more than theory at this point. There might also be a way to divide down the frequency.

The processor uses the track index mark and I'm wondering if it filters it out by itself or has it supplied. If one processor pin has the index available, it might be a good candidate for sunchronization. More theory.

pepe...thanks again for response. I realize the Reset line on the processor that comes from the 40 pin IDE line accomplishes the task of resetting the processor, but in the digital world, a hardware reset 'usually' refers to the resetting of a device by another piece of hardware. That can be a push button or another device like a transistor or chip. For example, some logic chips have a Reset input on them which can be tied to the output of another chip, or chips.

Unless I have missed something, the -POR line to which you refer does not go back to the processor 'on the Rigel'. It goes only to the memory buffer chip. I have a pin-out for the SH6770C on an Athena model and pin 27 on the SH6770 does go back to the processor on pin 48. The next pin to it on the processor, pin 49, is the RESET input. On the Rigel model, pin 48 on the processor goes to pin 4 on the 10 pin chip (A171F) I was unable to identify. The line connecting it to the processor is marked SHK. I think that might be a shock sensor circuit.

Whereas pin 27 of the SH6770 on the Athena goes to the processor at pin 48, pin 27 on the Rigel is the -POR line. The Rigel and Athena use pretty well the same chip set but use some chips differently. It's odd. That's why I need a set of data sheets for these devices.

I'll look at it more closely when I have the chance. I appreciate you giving me the feedback because it gets me to look at things a lot closer. It bothers me that the POR line is negated as in -POR. That means to me it is an active-low signal which will be on when it's low. That's exactly what the CKE line on the memory buffer is looking for. A reset line would have to be straight POR since it would have to go high with the application of power.

BTW...I heard from Maxtor and the reply was typical. They can't help....translated, they wont help. I suggest to anyone out there that they don't buy anymore Maxtor products. I certainly wont. Their tech support people are all brain-dead and it seems their management is too. I hope Seagate knew what they were doing when they took Maxtor over. Quantum used to be a good company till Maxtor ruined them.

Hi Old guru... =)

I have just had an maxtor hdd crash and I'm desperately trying to fix it, since my profession is HW/DSP designer, I've started to fiddle around with the PCB.. I think the PCB is the bad one in my drama.

I have good connections with salesagents that markets agere and I hope to receive a datasheet on th ms453 chip. Could you confirm one thing for me?
My ms 453 gets very hot when the power is turned on, and within a few minutes it starts to cool down, that is a positive sign that it might be the ms453 chip that is broken on my board..
I've tried to get a new one, or at least a replacer.. The closest thing I've come across is that it should be a dynamic memory. Though I know there also is a sdram (the one from micronics.. mt48lc1m16a1 on my pcb) there might be two mems on this pcb??

I have some q's that you might answer faster than it takes for me to read about it. Is there any calibration data that might differ between diffrent samples? Or is it possible to exchange pcbs and through that, repair the device?
I think the DSP might be very similar to the other DSPs TI manufacture.. if so, one should be able to connect an emulator and through that chech what te DSP is doing and if it's working properly. It's a bit of hard work, but I have the tools. (I also have a 4 ch 4 Gsamples Oscilloscope, so I can check if you and I have the same signals, and in that way we maby can help each other?)

I think you should be very careful to take apart the hdd, since just small dust psc may ruin the whole disc. I have access to a clean-room for this, but if you have detailed pictures from your hdd I would very much appreciate if you share them.
My last way out of this would be to get an exact copy of the hdd, and then in a clean room exchange the discs in it. But first I have to find one, then I can check if it's enough to exchange the PCB or not, then I can change the dics..

If you have more specs on the other components, please share, I will try to get samples of the ms453 so I can exchange mine. shall I get two? one for me and one for you??

keep up the good work, I presume you aren't from sweeden as I am, or we could meet and work together. I hace 5 years of vacation pics on my disc plus work worth a couple thousand dollars, so I really whant to save my discs.. (my backup disk broke 3 days before this... I hate my bad luck)

best regards

Hey young fella...I used to be a young tech myself.

The MS453 is not necessarily the problem. I noticed mine running a bit hot, but my rule of thumb has been the finger test. If you can keep a finger comfortably on the device, it ain't too, too hot. I've seen power transistors so hot you could fry eggs on their heat sinks, and they kept on chuggin'. In fact, my MS453 ran hot and cooled off like yours.

The MS453 is the encoder/decoder for the digital data, as far as I can make out. Data being written to and read from the disk must be in a certain format. The preamp takes care of decoding the head select signals from the processor and the recovery and writing of head signals.

The reason I say that is based on the lines I have traced from the MS453. They go directly to the preamp, which is mounted on the head positioner, as do several lines from the processor. The RDN and RDP lines, which are the read channel data from the preamp go directly to the MS453. I know the RDN/RDP signals are an unencoded signal from the disk after it is amplified and conditioned by the preamp. The fact that they go directly to the MS453 from the preamp means the MS453 is the decoder of the PRML-encoded data.

The 80 gig version has a separate flash RAM chip, but I haven't seen one on this model. The Maxtor manual specifies that the processor takes care of heads select and start/maintenance of spindle motor/voice coil motor.

The processor does not take care of the encoding/decoding of the PRML coded data to the head preamp, however, and the MS453 is the only chip in between the processor and the preamp that could. Also, Agere is in that racket.

If the MS453 is running hot, it could be a problem in the preamp, which is more likely. These chips work on a basis of sourcing and sinking. Sourcing is a reference to supplying current to another device and sinking refers to a device acting as a load for another device. If a condition exists in the preamp that is causing a higher amount of current to flow than normal, the MS453 will be forced to supply more current than it would like, hence the heating. If the MS453 was creating an overcurrent situation, that would also cause it to overheat, but the Maxtors have a history of preamp problems and I would guess the problem is there.


I'm no expert yet, but from what I've learned, the PCB's are pretty well interchangeable between drives of the same family. That doesn't mean that swapping them will solve your problems. In Maxtors, part of the boot firmware is kept in the processor ROM and part is kept on the service area of the disk. It's not clear to me yet that all models of the same family use the same processor ROM. I don't know if Maxtor uses the same ROM for a drive with one platter and two heads as it does for a drive with 3 platters and 6 heads.

It seems to me, however, that the families of drives like the Rigel family, were designed for future expansion and may use the same processor ROMS. Maxtor even suggests in their Rigel family manual, that the preamp MAY have the capacity for 8 heads, even if it's on a device with only two heads. I do know there is a difference between the microcode stored on disk between families, but I don't know if that extends to boards.

Recalibration would be a function of the servo formatting on the disks and the processor ROM. From what I've read, part of the 2 meg RAM buffer is partitioned as a scratch memory for the processor. That should mean the firmware on the disk is read and stored in the RAM buffer. I'm pretty sure DSP's have their own memory area, don't they? If that is the case, then the rest of the PCB should have nothing to do with recalibration.

I have read that not all PCB's are interchangeable in the same family but I don't see why. They all use the SH6770C spindle/voice coil motor controller and the MS453. They use the same mosfet setup to drive both motors. They all use a 176 pin DSP. The only thing that could differ is the ROM on the processor and the firmware on disk. I don't see why they would vary firmware between processors because that's the idea of keeping part of it on disk, since it's more accessible.

I found a 5T060H6 recently and there are several differences between it and mine, even though they were made in Singapore only two months apart. The 'unique' is different, as are the letter sequence and the number on the 40-pin connector. Even the processor is a different variety of TI DSP. I put the newer board on my drive and it spun the spindle motor up fine, but the old problem persisted. The fact that the motor spun up is encouraging because it's the processor ROM that controls the motor initially. The Maxtor motors don't use Hall effect sensors on the motors for feedback, all the startup motor control is done from data supplied by the processor ROM and intelligence built into the SH6770C controller.

Sounds good. I'd like to get a sampler scope setup for my PC if I can get one that is affordable. I have checked your theory somewhat and have done some similar reasoning re the TI DSP. I focussed on the 176 pin DSP's but so far I haven't found anything close enough. I spoke to someone at TI who indicated these DSP's were made specifically for Maxtor.

If you look at the layout, it makes sense. All the ATA connector inputs go to one side of the chips pins. All the RAM buffer pins go to another side. The third side is for the preamp, the SH6770 motor controller and the MS453 encode/decoder. The fourth side takes in the clock crystal and a few circuits I haven't traced yet. It seems to have been designed especially for the Maxtor board layout.

You would think, however, that a DSP is still a DSP. Hard drives differ somewhat from most applications in that their input is a 16 bit parallel data bus with a few control and address lines. Their output is serial, however, for data, and they have a few control lines as well for the preamp and motor controller.

I haven't finished my drawings yet, but maybe if i sent you a copy, the DSP portion would make more sense to you.

I appreciate what you're saying and I'm cognisant of the dangers. I have been keeping tract of the surface on the upper, exposed platter and I don't see any scratches on it yet. I have run the drive a lot since having it open and it hasn't crashed. Another fellow revealed that he took his apart, fiddled with the heads to unstick them, and his drive had not failed in the year after.

I don't have a digital camera but I may be able to borrow one. I don't mind sending my circuit drawings once I have cleaned them up.


changing the platters onto a good drive is an option I am considering. There may be a problem with that too. You have to realize that different models may have different servo formatting on them. I don't know. There's an awful lot to think about. I'm trying to find out more about the problems involved with mechanically realigning the platters once they have been removed.

It would be easier to change the head/positioner assembly than to change the platters. There is a good article on this site on changing heads.

I would appreciate a copy, but while your at it, try to get a copy of the preamp. It's an Agere and the numbers on it are: 541008D6/63109741. I have partially traced it and the six heads connect directly to pins 15 to 38 of that chip. There are 4 conductors per head and they run consecutively from pin 15 starting with the head on the top platter. The rest of the pins are +5v, -5v, gnd, and the data channel (read/write) with head select lines. Since it has an 8 head capability, it requires 3 select lines.



No...I'm from the west coast of Canada, being born in Scotland, I still want to meet Agnetha Faltskog of ABBA...so if you can help me... Not many women have turned me on with their voice alone, but she is one of them.

Be patient!! The worst thing you can do is rush the recovery. Even if it takes you a year, don't take chances. I plan to use my oscilliscope to see what signals are available on my working drive and compare them to the bad drive. I'll share them with you when I get them, but I don't have a digital camera.

Hi ther..

Sorry for my l8 answer, and unfortunately, I'm not that young any more.. =(, lets just say 35+... but less than 45..
I must have missspelled my email since I didn't get notified when you replied, hence the l8 answer.
I also have done some reasearch, and found out a bit more about the ms453, you summarised it up quite a bit in your answer.. They didn't have a pdf copy of the datasheet, however, they did have a papercopy so they will mail it by snail mail to me asap he told me..
you say it's easier to exchange heads, well I can't see how that will help since it's probably my electrics that's burned.. however, your idea about the preamp is quite intresting, I will check if I can get hold of a sample and exchange.. I have managed to read the contence of the flash and if I get hold of another pcb, i will use that flash content and compare what differs in that way I will be able to see if there are any calib-data stored.. I will (when I get the time) put that in my simulator, probably it's the same assembler, and we will eaier understand what the code does.
I don't think it's very strange placement of the pins on this dsp, so it would suprice me if the DSP differ alot from there normal dps.. unfortunately, when they order their own DSP that usually means that they place a bootstub in the DSP ( normally DSPs needs external mem to boot from, only big companys (like maxtor) can get their own dsps with their own start up code). This way they can crypt the contense of the flash so it's unusable for us to read, then we can't understand what it does..(well can't and can't, one can always grind down, and use a electrone microscope to look at it, but, that will take more time than it takes to redo all the lost work on the disk...It might even be funny for me to reshot my vacation pics again..)

I will try to get hold of a new preamp, exchange that and see if it works. Probably exchange the ms453 as well.. My ms453 gets much hotter (74 degrees C) than on my other maxtor discs, so I think that's at least one of my problems..
Not much new info here yet, it will come however. Unfortunately I have a bit too much to do at work, so I don´t have lots of time to spend here..

Cheers!

I thought maybe you had taken up my request to meet Agnetta from ABBA and had run away with her.

You haven't specified what your problem is yet. I said it would be easier to exchange the head assembly than to resolder the preamp. It's tiny and there's no way you could get at it to resolder it unless you took the head assembly out of the drive.

Is your motor powering up? Can you hear the heads seeking at boot time? What are your symptoms?



I don't understand what you are saying. First of all, I assume you have a 5T060H6 Rigel. As far as I know, there is no external flash ROM as there is on the 80 gig model. I don't refer to ROM as 'flash' ROM unless it is flashable. Even at that, flash is not a good description

There are basically three types of ROM: there's PROM, EPROM and EEPROM. The first is the write once type. The EPROM is erasable with UV light and can be reprogrammed electrically. The third is an electrically erasable PROM and that's what is normally called flash ROM. I don't know how flashable the ROM is on the microprocessor, but if you are able to read it, I'd appreciate knowing how you did it. I realize there is software available to do that, but I'm wondering if you are able to do it by reading the chip from the 40 pin bus line.

Also, I'd like to know what language is being used in the ROM code. Is it assembler, or do DSP's have their own language? I haven't had the time to check that out.

There is another matter, however, that I think you are missing. The code from the processor is not that easy to read. A lot of it is data for the motors. Also, much of the code for running the drive is kept on the disk, in the service area. That's the code you should be worrying about, and it varies from drive to drive.

You just can't compare the ROM contents to any PCB you might find. As you will find out, they vary significantly. Besides, I wouldn't worry about the processor code since it's hardly likely to fail. It's more than likely the code on the disk that is the problem, or the inability of the heads/preamp to read it.


I don't know how your simulator works. If it's like PSpice, or one of those simulators that allows you to mimic an electronics circuit independently of the hardware, that would be cool. I would be careful, however, of messing with the motor controls, especially on the spindle motor. If it loses speed suddenly, while the heads are out, they could crash. I know there are fail safe methods built in to retract the heads in an emergency, but you might be overriding them in your emulator.

On the older hard drives, they used a simple capacitor to retract the heads. If there was a sudden power fail, the capacitor would discharge through the positioner motor in such a way as to retract the heads to a safe position.




I wasn't implying strange as much as convenient. The chip seems to have been designed with the circuit board layout in mind. Not all DPS's have a 40 pin connector coming into one side of them and a memory buffer into the other. At least, the 176 pin DSP's I looked at did not resemble this one in any way.


the problem with decryption is that it has to be decrypted. That means they would have some kind of decryption code in there too. Since hard drives don't have proprietary drivers, and have to comply with the ATA standard, I don't think decryption would be all that difficult. They are not going to waste valuable ROM space with a decryption routine, I wouldn't think.

Besides, if they are using decryption, they would need a scratch pad buffer, which is most likely found in the external RAM. I'm wondering if DSP's have single-stepping functionality. Can you force them to single step through their code?


We call that shotgunning a circuit over here. It's easier to isolate what is actually wrong rather than changing chips on a guess. When you see the preamp, how small it is and where it's located, you might reconsider. Some of them are glued in place with a conductible glue and I'm not sure how easy it would be to remove them, They are mounted on a fine mylar-type harness that I don't think would take kindly to a lot of heat. The type of preamp varies from drive to drive as well.

Keep in mind what I said about the interaction between the MS53 and the preamp. Maxtors have a problem with preamps blowing and the overheating in the MS453 could be related to an overload condition in the preamp.