Need advice on how to attack a HDD problem

Hi...this is my first post. I'm an electronics technician with many years experience. I have also worked as a computer tech when HDD's were not as dense.

I'm starting out in the English forum, but I'm trying to learn Russian. That will take a long time to be conversant. Meantime, I'd appreciate help from anyone of any language.

I have a Maxtor 5T060H6 (Rigel - Blue-Fish). It is seeking even without the data cable plugged in. I am tracing the circuit and would be happy to share what I learn with anyone who is interested. So far, I have only covered the Voice Coil circuit. I can't find data on certain chips such as the SH6770C and the DSP (D741667APGF) from Texas Instruments, the MS453 and preamp, 541008D6, from Agere, and a smaller 10 pin chip with A17 on it with a 1F underneath the A17. I know the FDS6961A's are dual channel Mosfets and I have a data sheet on them.

I realize my problem could be a microcode problem on the service area of the HDD but I'd like to have a circuit to work with first. I don't like guessing at what I'm doing. The disk spins up fine but the heads are ticking without the data cable in place, and I realize that's a typical problem with a Maxtor drive.

If I can't repair it long enough to retrieve data, I'm wondering about building my own controller to take direct control of the heads. Alternately, I'd like to reverse engineer the Eprom code on the controller chip, once I locate it. At this point, I don't know if any of this is possible.

Does anyone have any advice other than forget it??

Hi Old Tech,

Take my advice - don't touch the PCBA! The problem is not there. There are 3 posibilities:
1. A broken head
2. Defective commutator/preamp (inside the HDA)
3. Critical module of SA is not readable
You can do nothing about the first two of them. A specialist can help you in the third case, but for you it will be wasting of too much time.
Good like.

BGman wrote:Hi Old Tech,

Take my advice - don't touch the PCBA! The problem is not there. There are 3 posibilities:
1. A broken head
2. Defective commutator/preamp (inside the HDA)
3. Critical module of SA is not readable
You can do nothing about the first two of them. A specialist can help you in the third case, but for you it will be wasting of too much time.
Good like.

Hi...and thanks for responding. Your suggestions gave me something to think about.

Please don't take anything I say as a personal attack on you because it certainly is not intended as such. But I'm getting a bit frustrated with this "don't touch' the innards of the hard drive.

I pointed out that I'm an electronics tech with many years experience in electronics and computers. That's what techs do, we repair things. I used to take hard drives apart on customer premises, change the heads and/or platters, and they worked fine. I also pointed out that things have changed. Todays drives have a much higher density of cylinders per inch.

A similar fetish developed around static electricity. I have never worn an anti-static strap and I have never blown a device. I don't trust anti-static devices because they have a high resistance built into them that supposedly limits static current. It doesn't work that way. You'd need copper wire wrapped around your wrist and I'm not about to do that. I have found it is enough to ground yourself to grounded metal 'each' time you move around.

There is a folklore developing in the hard drive community about dust. It has gotten to be a fetish. It's not new to me, however, since I've been hearing about the mysterious dangers of dust most of my life. What the heck is that stuff anyway? We should be using it as a secret weapon. :-)

A quick story. I took a hard drive to pieces on a customer's premises many years ago. I exposed the platter and heads to flour and general grain dust, because the customer's business was bakery goods. I did the same on another customer's premises who used a very clean room with air filters and everything else. Guess whose drive failed?

Do you want to know why? The computer operator in the clean room was a smoker. She smoked while entering data, right in the clean room. The particles from the cigarette is dust, with lots of sticky stuff in it like tar and the metal products the cigarette manufacturers add to affect the taste of the cigarette. Even though the hard drive had it's own filters, the goo from the cigarette had gotten through them and stuck to the heads.

Dust is not composed of one thing. If it has an organic base, say from cigarette smoke, or dandruff and skin matter, then it 'might' pose a risk. There is an article on this site by Stanislav Korb. It's an excellent article on how to change the entire head stack. I followed the article with great interest because it's the first article I've read that tackles modern hard drives. It also confirmed my suspicion that nothing has changed, even though modern drives are much more intricate. If you're careful, use 99% isopropyl alcohol (it 'must' be 99%) to clean up, use lint-free swabs of some kind, and work in a relatively clean atmosphere, it can be done.

Consider the aerodynamics of a flying HDD head. It is held above the disk surface only by the air pressure under it created by the high speed of the drive. That's why Stanislav suggest rotating the platters as you remove the heads, because there is so much relative pressure on them that dragging them forceably across the platters will scratch the surface. I think Stanislav even suggested wedging the heads apart slightly to ease them out.

Anyway, with the disk turning at 7200 RPM, which is double the speed they turned in my day, the heads maintain a relative distance above the surface. I say relative because the heads are prone to vibration and they will wobble up and down as they fly. If they hit a piece of organic matter, they will fly right through it, but the jolt will cause the head to oscillate in it's path. "If' it hits the surface, it's probably game over. The greatest danger there is that the smear caused by the collision will affect the heads ability to read and write. I'm sure that a lot of the problems attributed to opening the head chamber is contamination of the heads themselves by matter. You need to find a way to clean them.

With tars like those from cigarettes, the tar will actually stick to the heads. I've seen the trail left by such a head and the smearing on the head surface. It doesn't take much to ground the head. With other types of dust, like lint, it's hard to say what will happen. I'm trying to say that heads and platters are a lot tougher than people seem to think. In my case, I'm not aiming to put the drive back into use, I just want to recover some data and chuck it. I feel strongly that I have a decent chance of accomplishing that.

You mentioned there was nothing to be done about changing the preamp. Why? The problem I'm having is locating parts. Surface mount technology is admittedly tougher to work on, but there are techniques available to all of us to change even that preamp. The first thing you need is a high powered magnifier. I have terminated lots of fibre optic cable which is a lot smaller than the preamp. I used a 100 times magnifier to see the end of the fibre because it needs to be perpendicular to the length.

The next thing you need is a really good soldering station, with a very fine tip, and high quality, fine tweezers. I can see from the ribbon cable attached to the heads that heat will be a problem. That means careful control of tip temperature and heat sinking. I am not pretending it will be easy, but I can't accept your advice that nothing can be done. At the same time, you have helped a lot by prompting me to consider looking elsewhere first. In my still-primitive approach, I am considering ways to bypass the manufacturer's control over the voice coil mechanism. I have a lot more reading to do.

With respect to the experts you mentioned, the only way to gain expertise is to research, ask questions, and 'do it'. If you fail, big deal, at least you tried.

Hi Old Tech.
it's really interesting to read your answer.
can you share more about your experience it to article like Stanislav Korb's article.
thanks...

tonyht wrote:Hi Old Tech.
it's really interesting to read your answer.
can you share more about your experience it to article like Stanislav Korb's article.
thanks...

I'd be happy to share anything I discover, but right now I'm in a very preliminary stage. I'm trying to discover how the circuits works on the drive electronics board and it's difficult to know where to start.

I contacted Texas Instruments and they passed me back to Maxtor. Of course, Maxtor wont help either, so I'm stuck trying to find data sheets for the DSP controller (D741557APGF) and the spindle motor/voice coil chip (SH6770C). I contacted Agere about the MS453 chip, but they did not have the decency to reply. How do you spell conspiracy? It would be helpful if someone could give me information on these chips and/or on what other models they may be found.

From the time the data comes in on the ribbon cable (16 bit), till it reaches the read/write heads, it has to be changed from parallel to serial. There are only something like 18 conductors going to the head chamber and about half a dozen are for the voice coil motor/mechanism itself. I'm theorizing that the Agere MS453 converts the parallel data from the storage buffer to an encoded form for the read/write pre amplifier, and decodes the data on the way back from the pre-amp.

The pre-amp acts as an amplifier,a multiplexor/demultiplexor for the head select and data delivery, as well as a signal conditioner to clean up the analog signal from the heads. There are some good explanations of that on the net, but I need time to read and digest it.

As I was tracing the ribbon cable entry to the electronics board and the DSP, I discovered something interesting. The drive select lines all go to the same area of the DSP, I think it's pins 44 to 46, but one of the head select pins, the pin by itself, goes to 5 volts. Why would they put 5 volts on a drive select pin? The other select pins like master, slave and cable select all go to ground via the jumper. Is the 5 volt selection there for the 'safe mode' or is it for some other factory purpose? It's next to a pin marked RxD on the board and that pin goes directly to the DSP.

I don't have much time tonight, but keep in touch.

tonyht wrote:Hi Old Tech.
it's really interesting to read your answer.
can you share more about your experience it to article like Stanislav Korb's article.
thanks...

I aggre, u can Title for u article "Harddisk & frustrated smoker"

Keep discovering

It is probably the first (?) time we can see here a man who is doing right: he is exploring hard disk drives by himself and not asking questions like "where to get PC3000" and "how to flash the disk". Well. I cannot comment these answers…

Regarding the original question: if the drive is making "clicking" sounds and pcb change did not help, then you've got heads killed (or preamp killed).

It is possible to disassemble the ROM, it will give you best results (but will not help you to recover this particular drive

). And creating a device which controls the heads will take several years (it is possible).

hi Old Tech
really very great work and analysis and keep going on , and u are really ture about datasheet or technical support from any manufacutre who supply hdd chips .
all are of none help at all
with all the best of luck

Regards

maysoft wrote:It is probably the first (?) time we can see here a man who is doing right: he is exploring hard disk drives by himself and not asking questions like "where to get PC3000" and "how to flash the disk". Well. I cannot comment these answers…

Thanks. I don't understand how anyone can afford the PC3000. I understand it costs several thousand dollars. As far as flashing the disk, in the west, 'flashing' has another meaning too. It's about the dirty old man who wears only a raincoat with nothing underneath. He opens the raincoat to 'flash' unsuspecting women. That's about all you could expect to flash on a hard drive. :-)

maysoft wrote:Regarding the original question: if the drive is making "clicking" sounds and pcb change did not help, then you've got heads killed (or preamp killed).

I haven't tried another board yet because I can't find a working Maxtor 5T060H6 drive. I can't really troubleshoot my own drive until I have the circuit. That's why I'm tracing it, basically. Even at that, I only have an old 10 Mhz oscilloscope and I don't know if the bandwidth is good enough for this application.

maysoft wrote:It is possible to disassemble the ROM, it will give you best results (but will not help you to recover this particular drive ). And creating a device which controls the heads will take several years (it is possible).

I only want to disassemble the ROM to see what they are doing in it. It's possible to disassemble and reverse engineer the BIOS on a computer, so I don't see why that can't be done on a disk drive.

There are 3 address lines on the 40 pin connector by which you can send commands to the DSP. It should be designed to respond to the ATA standard. As far as I understand, you can get it to send it's own data out on the data lines. Of course, I am only theorizing at this point.

I wasn't thinking as much about creating a device as interrupting the present circuit and sending it back the information it wants to see. For example, if the DSP wants to do a maintenance routine, or a S.M.A.R.T. check, it will wait for one of its pins to go high (or low) as a signal that the job is complete. Maybe I can send it that signal (or code) and it will think everything is OK.

If my head asembly is damaged, or the pre-amp, that's a hardware problem and there's nothing I can do but repair it. But, if the ROM code wants to see certain functions performed before it will turn control over to the BIOS or diagnostic app, then I think it should be possible to trick it into doing that. More theory. I have a lot to learn about the microcode and the tables on the service area of the disk.

Do you know much about the ROM on these drives?

Last edited by Old Tech on March 12th, 2006, 23:46, edited 1 time in total.

sawwa7 wrote:hi Old Tech
really very great work and analysis and keep going on , and u are really ture about datasheet or technical support from any manufacutre who supply hdd chips .
all are of none help at all
with all the best of luck
Regards

thanks for the encouragement. It is frustrating that corporations have to be that petty. Microsoft was the same, but people have found how to reverse engineer their code and use it to their advantage.

darkfroce1 wrote:Keep discovering

Or as we say in the west...."keep on rocking". :-)

Hello Old Tech,

one thing about reverse engineer: it is not easy to reverse engineer a code even if U know the opcodes and addressing modes of a CPU, but the main problem with the DSP processor is that U probably cannot gain acces to those opcodes and other stuff like what the ports are used for and how do they exactly work.
Another tricly part is to interpret the code once disassembled.
This is hard as to my current knowledge and I would lift my hat if anyone could do it without detailed info from manufacturers.

regarding your drive: have U tried ebay?
regards,
pepe

pepe wrote:Hello Old Tech,

one thing about reverse engineer: it is not easy to reverse engineer a code even if U know the opcodes and addressing modes of a CPU, but the main problem with the DSP processor is that U probably cannot gain acces to those opcodes and other stuff like what the ports are used for and how do they exactly work.

as long as the code has to work on an Intel machine, it has to conform to standard assembler coding. Of course, on a DSP from TI, they are not confined to Intel standards. However, I have briefly looked at PICs and other microcontrollers and they seems to stick to basic assembler. That means you know all the opcodes because they are standard. Also, if the code is in assembler, there are disassemblers which will disassemble the code to some extent. Of course, it's not easy, and you have to use trial and error and some guessing.

The structure of modern processors require some kind of register system to take in data, shift it around between register and memory and output it, if necessary. Even if you don't know specific codes, many of them can be inferred by direct inspection or by examining the data they are working on.

In the case of my DSP, I know already that it takes in 16 bits of data, 3 addressing bits and other control signals. The addressing bits 'seem' to have a standard to them based on the ATA standard. In other words, the DSP is there to operate a hard drive which must conform to the ATA standard. Where things will get tricky is in the code designed to convert the data that is input to the DSP from the main computer. The read/write heads require a specially-coded signal for storage on the platters. Hopefully that is taken care of by a dedicated chip, like the Agere chip.

A code in machine language, as it would be stored in a ROM, is more difficult to analyze than a compiled code from a language like C. That's because most application written in C are aimed at Windows, and it has a standard format for creating windows, retrieving messages, etc. Then again, the ROM program would be smaller and more specific.

pepe wrote:Another tricly part is to interpret the code once disassembled.
This is hard as to my current knowledge and I would lift my hat if anyone could do it without detailed info from manufacturers.

I agree it would not be that easy, but the scope of the program is quite limited to the functioning of the hard drive. Also, the program is confined to the ATA standard and to the size of the ROM available. The trick to deciphering what is going on is to look for 'jump' statements and other familiar opcodes like for output, etc. By tracing the hardware, you can narrow down where the ports are, and I'd be willing to bet that Texas Instruments did not re-invent their processor technology just for Maxtor. There is plenty of information on TI DSP's on the net.

pepe wrote:regarding your drive: have U tried ebay?

Yes. Most of the drives I saw were already sold. I'll keep looking. The problem with ebay is that you don't know what you're getting. :-)

Old Tech wrote:I haven't tried another board yet because I can't find a working Maxtor 5T060H6 drive. I can't really troubleshoot my own drive until I have the circuit. That's why I'm tracing it, basically. Even at that, I only have an old 10 Mhz oscilloscope and I don't know if the bandwidth is good enough for this application.

Hi Old Tech
actually you do not need to look exactly for 5T060H6, you can use any 5T0x0Hx (Rigel). Anyway you will need to pay attention to the PCBa because in some rare cases there are incompatibles
All this is IMHO

Hi Old Tech,
Read my message carefully! I wrote " don't touch the PCBA", not inhard!
If the PCBA was the problem the drive wouldn't spin up at all. If you want to be really sure just find the same PCBA and try it. You drive belongs to Maxtor Rigel family (as Samo wrote) which was very popular. Each PCBA has a label with 5-character code (in bold). Make sure that it is exactly the same.
Good luck.

BGman
There might be lots of other issues including head clicking when the pcb is damaged, not only "drive does not spin-up" problem.

maysoft wrote:BGman
There might be lots of other issues including head clicking when the pcb is damaged, not only "drive does not spin-up" problem.

As far as I know Maxtor DSP first check the PCB itself and then starts the motor.

That is correct. But there is no way to check all components on the pcb...

Samo wrote:actually you do not need to look exactly for 5T060H6, you can use any 5T0x0Hx (Rigel). Anyway you will need to pay attention to the PCBa because in some rare cases there are incompatibles
All this is IMHO

thanks for info. I figured that but wasn't sure. Apparently yhe SH6770C chip that controls the motor and voice coil is common to other boards as well.

Need advice on how to attack a HDD problem

Need advice on how to attack a HDD problem

keep Discovering

Re: keep Discovering