A company by the name of “DataSlide” had released a prototype storage device that they claim may one day replace SSD and traditional HDD. It’s apparently capable of acheiving 160,000 IOPS (Input/output operates per second) and 500MB/sec performance levels all while consuming no more then 4 watts of power.

 This all sounds well and great until you hear how it manages this level of performance. The device uses a piezo electric actuator to keep the media in a precise motion, and a diamond solid lubricant protecting the surfaces (Which are in direct contact to each other) It then utilizes a massive 2D array of heads (Up to 64) to perform read/write operations.

It will remain to be seen how this technology develops, but from my experience I have enough problems with magnetic media that is utilizing only 1-8 heads right now; I can’t imagine how a device that uses 64 heads could be reliable.

      I thought it would be a good idea to cover some “what if” scenarios. This post is primarily for someone who’s hard drive may have failed in a specific way and they are wondering if there is anything they can do to fix it themselves. While I am not against DIY solutions, I want to stress that if you are attempting DIY solutions on a drive that has valuable data on it that you are putting the data at your own risk, and it is ALWAYS better to send it to a professional immediately to ensure proper care.

What to do if: You knock over your external drive while it is on and now it isn’t recognized.

    This is a very bad scenario. In this case the only good course of action is to NOT attempt to turn the drive on again and talk to a pro ASAP.

    Why? Because significant damage has probably been caused to the internal mechanics of the drive that, if not treated properly, can render your data permanently inaccessible and gone forever.  

     There are a few different ways that this can happen:

      1) The heads might crash into the platter causing a radial scratch on the media surface which possibly kills the head and renders that area of the disk permanently un-readable.

      2) The Spindle Motor suffers damage and seizes. It will no longer spin up unless it is somehow loosened. Even if it can be loosened many times there is “eccentricity” that will cause the heads to be unable to read the media.

      3) The heads may stick to the media, causing damage to the surface, the heads, and the motor. This can be made even worse if the user continues to attempt to re-power the drive.

What to do if: You plug the wrong power connector into your external drive or you plug in the Molex/SATA power connector in upside down.

    In this scenario you have at minimum damaged the PCB (Printed circuit board) of the HDD. You will need knowledge of circuits and be handy with a VOM (Voltage-Ohm meter) or DMM (Digital Multi meter) to repair this. Typically the only damage that will occur to your drive in this case is to the power protection components of the drive (assuming they did their job and did not fail). You will need to repair these components to return the drive to service. Even if you manage to repair the drive it is never recommend you use the drive again; transfer your data to a new drive and back it up.

Our second data recovery seminar has come to a close and I am just getting back into my normal activities. I wanted to share my experiences with the seminar and give a preview of what is to come in the future.

This year we went a slightly different route with the seminar presentation and it was an overall success. We started Day 1 lecturing about hard drive anatomy. We began by covering the various components of the PCB, and then moved to the anatomy of the HDA and explantions about how data is stored and accessed on the media. We then explained in detail the workings of the “Push-down” defects list, and “Grown” defects list and covered HDD firmware structure, and finally walked through the start up operation of a HDD.

At this point we gave a full demonstration of the Atola insight; we demonstrated what it can do and how it works. This is the device we would be using for all of our demonstrations for the rest of the seminar, so we felt it was important that our attendees understood the inner workings of it before we began diagnosing and repairing drives with its help.

We then went into the basics of diagnosing a HDD. We covered basic PCB diagnosis, Head-stack diagnosis (i.e. the nature of the clicking noise) and also spindle motor diagnosis and firmware diagnosis. In addition, we covered the various ways that firmware can become corrupted.

After covering basic diagnosis, we then dove in to general theory of recovering PCB and firmware problems.

Next, we spoke about hard drive security. We covered ATA passwords (including a demonstration on how to remove passwords on Maxtor, Hitachi Travelstar, and Seagate drives). After that we talked about FDE (Full Disk Encryption), HPAs (Host Protected Areas), and DCOs (Device Configuration Overlays).

After security, we moved to a detailed discussion about imaging. We started with the overall objective and purpose of imaging and then went on to investigate the requirements for a good imager and the required attributes and parameters that are needed. We then moved to a basic explanation of the file recovery process and talked about some of the hardware/software solutions that are available.

Lastly, we talked about RAID arrays; their inner workings and relevance to data recovery.

At the end of Day 1, we changed the focus from lecture to hands-on demonstrations. Using various broken drives, that had generously been donated to us, we used the Atola Insight for on-the-spot diagnosis. We had not looked at any of these drives yet, so this was a perfect way to demonstrate a live HDD diagnosis procedure. We invited our attendees to get involved in the process and try to guess what the problems were to give them practice with diagnosis with the Atola Insight.

Day 2 and Day 3 were dedicated to demonstrations and hands-on activities.

We began Day 2 with a short lecture about the various aspects of the data recovery business and talked about the tools and facilities needed to start up a data recovery business.

We proceeded to the demonstration portion of the day starting with a demonstration of PCB recovery. We had previously prepared two HDDs - one with a failed TVS (Transient Voltage Suppressor, or Zener diode) and another with a failed fuse (sometimes called “ferrite bead”). We demonstrated the proper diagnostic procedures for these types of failure – showing good use of a Digital Multi-Meter (VOM) and proper recovery techniques for these failures.

Next, we demonstrated head stack replacement using a simple single platter Seagate drive as an example. We then showed the same procedure but instead on a 2 platter / 4 head Western Digital Scorpio HDD.

After the demonstration, we walked the attendees step-by-step through the process on their own drives, which were provided to them with the seminar cost. Each attendee received his or her own Seagate single platter and Western Digital 2 platter HDD).

After each attendee had completed the removal and replacement of the heads back onto their drives, we explained the procedures for “screw ups”, or fixes for common mistakes (like dust removal and fingerprint removal) from the platters. After the attendees and cleaned and reassembled their disks appropriately, each attendee had the opportunity to test their own drive on the Atola Insight to see if they were successful. To our excitement, many of the drives were still in working condition! Western Digital Scorpio drives seemed to fair the best in the dust filled environment. Even without a cleanroom, our attendees were able to remove the heads, put them back, and still have a working drive.

We ended Day 2 with “Hard Drive Fact and Fiction”, where we discussed some common myths about hard drives and data recovery. We also discussed the feasibility and facts and fictions behind extreme recoveries (such as dropping a HDD in water or rescuing data from a notebook that had been in a fire, and even a drive that had been dropped from a large height).

Day 3 was dedicated to spindle motor repair and platter swapping. We covered all of the various techniques involved in diagnosing and repairing a spindle motor, and then moved onto platter transplants. We went over some of the various techniques developed by other technicians, as well as some common commercial solutions such as the Salvation Data HPE. We had the Salvation Data HPE on hand to demonstrate and to allow attendees to try themselves.

We ended with a final lunch where we presented each attendee with a Certificate of Completion acknowledging the attendees’ completion of the course, signed by Dmitry Postrigan and myself.

Overall I had a great time. I always enjoy Dmitry’s company, and also seem to learn something new when I am around him. At this point in time, it looks as though Dmitry will not be joining us for future seminar’s (This may change). He wants to focus on development of his Atola insight product, and giving his customers 100% of his time and energy. The torch has been passed to me to conduct future training.

I wanted to give a preview of what we are planning here at Ji2. Keep in mind that none of this is confirmed at this time, but if you have some interest please contact us and let us know.

We have been considering offering shorter, more customized training – possibly on a monthly basis, based on interest. We are thinking of offering different level courses, each one tailored to customer needs, level of experience, or focused on specific topics. Examples of this could be offering a Basic Class and a separate Advanced Course, or even more specific topics such as Western Digital Data Recovery. If you have any suggestions or have ideas about class topics you’d be interested in please post a comment or email us directly.

That’s all for now; I will begin posting more frequently from now on so keep an eye out for some good info.

- The HDD Doctor

I am happy to announce that Ji2 will be hosting our second data recovery seminar! It will begin April 8th and continue until the 10th, and will take place in Cypress California near our offices.

In light of this new announcement I wanted to share my experiences of the previous seminar to give you an idea of what you can expect at our upcoming event.

Day 1: I headed over to the seminar location in the morning ~8:30am, and arrived to catered breakfast in the main hall. Dmitry began at ~9:00am and spoke until 12 noon at which point there was a 1 hour break for lunch. The following days would play out very similar, but with much more hands on activities.

                The speaking schedule (roughly)

 9:00am - 12:00pm (Hard drive anatomy) Covered the basic components in a HDD, firmware modules, start-up operation, RAIDs.

1:00pm - 5:00pm (Diagnostics) Covered: “The nature of the clicking noise”, Data recovery flow, PCB anatomy and repair, clean-room (what you need to build one), ATA passwords, Firmware corruption, Imaging. To name a few of the subjects covered.

Day 2: This would begin delving into more advanced data recovery procedures, and also give every attendee the opportunity to open up two hard drives and attempt clean room procedures, as well as soldering.

  9:00am - 5:00pm (Advanced Diagnostics and Data recovery procedures) Physical procedures, Head and preamp diagnosis and replacement, screw ups (fingerprints, etc) this included hands on lessons with removing and replacing headstacks, and screw ups.

Day 3 - This would be the final day of the seminar, and would cover platter transplants, and also some chances to get 1 on 1 QA with Dmitry Postrigan himself.

9:00am - 12:00pm (Platter replacement) Hands on lessons with platter transplants

1:00pm - 2:30pm (2 options) Option 1: Guidance software: Bit-locker drive encryption or Option 2: 1 on 1 QA with Dmitry Postrigan

2:30pm - 3:00pm (Atola insight QA)

3:00pm - 4:45pm (Data recovery QA)

4:45pm - 5:00pm (Closing discussion)

That is a general breakdown of how our first seminar flowed. We anticipate that our next seminar will be even more refined and provide an even broader wealth of knowledge and hands-on experience. It goes without saying that after attending these seminars you walk away nearly exhausted from the amount of knowledge that you have gained in only a matter of days. These seminars are intended to give you the knowledge you need to get you started on your way to becoming a Data recovery professional, or to start up your own Data recovery firm.

Hope to see you in April!

-HDD Doctor

Part 1: Platters

            Probably the largest characteristic of a hard drive is the “disk”. This “disk” is actually called a platter. In fact, current HDDs commonly have two or three platters, some even have four or five!

            So what IS a platter? What is its purpose? What are they made of? How do they work? All of this and more will be answered as we take an in-depth look at Platters.

A hard drive? Whats that?

            Hard disks or “rigid disks” where originally produced by IBM in believe it or not, 1956. They where composed of an astonishing fifty 24 inch platters that where capable of holding a total of roughly 4.4 megabytes.

            We have come a long way since 1956. Currently (as of the end of 2008) a single platter in a desktop hard disk is capable of holding up to 500GB (Samsung as of Dec 29^th 2008). This is made possible by countless different technologies that increase what is called the “Areal density” of a HDD. The “Areal density” is simply the amount of data that can be fit into a physical area. As this density increases, for example: a platter at one time was only able to fit 5GB per square inch, and is now capable of fitting greater than 125GB per square inch.

            The physical size of desktop HDDs has not changed since the 1980s. These HDDs are called “3.5 inch HDDs”. The measurement is based off of the diameter of the platters, and not the actual dimensions of the HDA (Hard disk assembly). These drives where created originally to fit in the same space as the larger floppy drives, which at the time fit in the 5.25 inch bays in a PC. Since then, floppy drives have shrunk to a 3.5 inch factor, and even more recently have been made obsolete by “Flash memory”. There are other “form factors” of HDDs that exist besides the standard 3.5 inch desktop drives. 2.5, and 1.8 inch notebook drives are set to surpass 3.5 inch drives as the standard due to the increasing popularity of notebooks, and even more recently “netbooks”. Even smaller form factors exist, such as the 1.3 inch and 1.0 inch compact flash drives. These drives found their popularity being used in portable media devices such as the apple iPod and Microsoft’s Zune. 

The world of tomorrow, today!

            Platters and HDDs for that matter are one of today’s current technologies that reflect just how far our grasp of science and mathematics has taken us in such a short period of time. HDDs utilize magnetism to produce a signal that can represent binary code, which in turn is translated into data bits which hold hexadecimal code that is translated into the mp3, video, text, and picture files that you view on your computer.

            Platters consist of many layers, and are divided in to sub-micro meter sized magnetic regions. Of these layers the most important are the magnetic medium layers. These layers are based primarily on a cobalt-based alloy; beyond that each of the sub-micro meter sized regions are composed of hundreds of magnetic grains. The reasoning behind using hundreds of grains instead of one consistent medium is because of an occurrence called a “Neel spike”. Without going in to the physical science of magnetism, it is possible for these spikes to cause disruption in the magnetism of the medium, and also reduce the areal density of a HDD greatly.

            Platters are created by first taking either an aluminum or glass substrate and then coating the substrate with various layers using a technique called “Magnetron sputtering” which sounds like it came directly out of star trek. The process begins by first adding the under layers which are made up mostly of various non magnetic metal alloys that are then covered with the final magnetic medium layers which are made of primarily of the grains that we just discussed. Finally, a nanometer layer of lubricant is added for the purpose of protection from head crashes (we will cover this in a later article).

            Many improvements have been made since the HDD became the standard storage medium in the 1980s. None have been as beneficial to the effectiveness of HDDs as a new technology that was just recently implemented in 2005 called “Perpendicular recording”

            “Perpendicular magnetic recording” or “PMR” is a new way of recording data to a HDD, and was actually discovered and proven to be useful as early as 1976 by a Japanese professor named Shun-ichi Iwasaski. This new technology is allowing the areal density of platters to triple, allowing capacities much greater then previously possible. The technology works by utilizing material with “higher coercivity”. Basically a stronger magnetic field coupled with new magnetically soft layers is utilized to allow the bits to be placed vertically instead of horizontally. The data must be written “deeper” so the magnetic medium must be stronger, as well as the heads which read the data must be stronger to be able to extract the data from the greater depth. This is a simple explanation, the exact science is extremely complex, and there are many more variables to consider in the science including temperature levels. The end effect is that each data bit takes up less physical space allowing the platter to hold more bits in a smaller physical area, thus increasing the areal density of the platter. The predicted maximum density allowed by perpendicular recording was said to be ~437GB per platter, but Samsung recently released a ~500GB per platter drive which proves that prediction false.

The tricky part

             Understanding all of the scientific and mathematical processes that are taking place “behind the scenes” helps give you perspective on the complexity of the technology we are dealing with here. Once you begin to grasp the way that the technology works it becomes easy to understand why recovering data from these highly advanced devices is so difficult, and why data recovery companies charge so much to perform these services.

            In particular, platter damage is one of the worst case scenarios that can occur in data recovery. Even a scratch barely visible to the human eye can be catastrophic to the success of the recovery. While I have not seen any such procedures in action, there are rumors that there are techniques that involve creating a material that is very similar in property to the cobalt based alloy that coats the platter, and using it to fill in the damaged areas. This sounds great, but unlikely due to the high precision that would be required to perform such a procedure. This would have to be performed by a machine with precision down to the micron, and the ability to maintain platter alignment to perform repairs on all of the platters and their bottom sides. There are rumors going around that a machine of this magnitude is being developed by a company named Salvation Data, but due to the complexity of the machine I would not expect this machine to be ready in the near future. Until then scratches on the platter are the leading reason for un-recoverable drives.

            Another scenario that can cause a drive to become un-recoverable is platter de-alignment. This is a problem that is becoming increasingly common due to the higher areal densities found in today’s HDDs. Platter de-alignment occurs when the platters shift in opposite directions causing the data to no longer line up correctly. Usually this takes a fairly significant shock to cause this to occur, but because of the increased density of current drives a smaller knock can cause a large enough shift for the drive to not be able to find it’s track. The more platters the higher chance of this issue arising.

            The final scenario is water and fire damage. Assuming the water that caused the damage was of average temperature it would be MORE desirable for water damage to occur than fire damage. Water damage can likely be reversed as long as the drive is rushed to a qualified data recovery center quickly. The clock is ticking because the longer you wait the more residues that can build up and be harder to remove. Fire damage is even worse. The high temperatures can actually scramble the magnetism of the drive, acting almost like a degausser. Another effect of the high temperatures is obviously melting, which the platters will do, and if this happens then recovery is impossible.

So, in conclusion…

            Platters are just one piece of the puzzle. A relatively huge piece, but one of many pieces that fit together to give you the working hard drive that your probably using to read this article with. There are more articles to come, and I promise they won’t disappoint. Keep checking back for the next part of this in depth look at the hard disk assembly.

Just a quick overview of some of the more interesting events going on in the storage technology world:

Western Digital purchases Fujitsu’s HDD division, and now owns roughly 25-30% of the 2.5 inch market.

SSD’s continue to grow in performance, with OCZ, Intel, MTron, and Samsung leading the way. But at the same time many major issues still remain, especially issues with fragmented free space. Causing these drives to slow down by 80%  after only six months of use. Many software companies such as Diskeeper are scrambling to create defragmenting software that can allieviate these issues.

FDE (Full Disk encryption) hard drives have been around for some time, but recently seagate began providing Dell with new generation FDE drives that Dell will offer in it’s notebooks. These drives will be near impossible to recover if the user forgets the password, or the firmware becomes corrupted. While these drives do make your data safe from notebook theft, it makes your data less safe from HDD failure.

Many new Hitachi GST drives are coming with drive stickers that do not cover the center lid screw. This means that you can remove the lid without any trace of doing so. This is bad for Hitachi and for data recovery companies. This means that any user, or low quality data recovery company can open these drives and wreak havoc upon them and as long as there is no visual damage no one would know. Hmmmm.

Intel and Hitachi GST announced that they are jointly developing SAS and Fibre channel enterprise SSD drives that are planned to ship in 2010.

          Hello everyone,

             So we went over all of the parts that I chose for my work PC and why I chose them; and also made some recommendations for alternative components.

              Now that we have all the parts it’s time to put them all together and make it work. I’m going to walk through the steps that I take when building a PC; while giving some pointers that I have picked up over time.

The first thing I always do when building a PC is to prepare the Case. If the wires for the power button and reset button are in the way; move them to a safe, out of the way position for now. Also remove both side panels for easy access, and wire management. Some cases you may want to take off the front panel; as with the case that I got (the sides bulge out so that  it will not sit flat unless you remove the front panel).

Bottom view with front panel removed

Side view looking at the HDD area

 After preparing the case I always start by installing the processor and heat-sink (used to cool the processor) while they are outside of the case. (Some cases have “removable motherboard trays”). If the case you choose does then you will want to remove it and do the installation on the tray while it is removed.

The reason I like to work outside of the case is because installing heat sinks (especially large, aftermarket ones) can be increasingly difficult if you attempt to install them while in the case. Also, many large aftermarket heat sinks (like the one I chose) have a special back plate you will have to install on the bottom-side of the motherboard for extra support so the board does not break.

Heatsink (left), motherboard, processor, thermal compound

Mirror finish on the Heatsink base

 Installing the processor itself is simple; and there are instructions provided on how to do so which explain it as good as I can, so I recommend following those instructions.

Once you have installed the processor, you will need to apply thermal compound to both the processor and the heat sink. There are many “techniques” that are recommended to do this. From my experience, do not listen to the instructions telling you to put a “drop” on the middle of the processor and letting the force of the heat sink spread it. I recommend using the technique I have used for a number of years, and always get amazing results.

First, you will need two “tools”. A zip-lock bag, and a used credit-card/ gift card. Then what I do is put a drop on the heat sink (more then a small drop, but not too much) then you are going to want to put the zip lock bag over your hand, and massage the compound into the base of the heat sink; the reason for this is because there are microscopic machining marks that will be filled that increase the heat transfer even further. Once you have done this, put one more small drop on the heat sink, and use the card to spread it evenly across the base. Once you have done this; put another drop on the processor and spread it evenly across the top of the processor. Make sure that none of the metal on the processor is showing, but that the compound is not too thick (paper thin at most).

Spreading the thermal compound on the processor

                Spreading the thermal compound on the processor    Once you have done this it is time to install the heat sink. You heat sink may have a different mounting apparatus; so my best advice here is to follow the instructions provided, and if in doubt get online and browse around some forums (overclocking forums are a good place to look) for guides and reviews made by other users. You can find some useful pointers about your heat sink, and the best techniques to use while installing.   For my Zalman cooler that I chose; I have found that the best way to install this is to have a partner handy to help you. The reason for the is because the bracket calls for you to screw in two screws alternating back and forth on each side until the heat sink is securely fastened. The problem that the instructions don’t tell you about is that the bracket is curved (sort of like a U shape slightly). This makes for a nice tight fit once fastened, but also causes a lot of resistance while installing; you can never get one screw quite tight enough to stay in place while you tighten the other screw; I actually stripped one of the sockets the first time I tried to install this cooler (on a different computer) and had to re order a new heat sink. So, the best way is to have a partner with a screwdriver screw in one side a little; while you are also pushing down and screwing in your screw, and alternate back and forth until they are both tightened securely. Installed heat sink

Installed heatsink

 Once you have installed the processor and heat sink it’s time to install the motherboard into the case. Make sure you install all of the pegs into their correct positions, and then carefully place the board into place. You will probably have to apply some force to the board (pushing towards the back of the case) while you screw in the peg screws.

The motherboard installed into the case

      Installing memory is extremely easy; just match the connectors and apply force. Usually you will hear a “click” but don’t count on it. Watch the clips and observe if there are in the proper place after you have applied force. You can install the memory before or after you install the motherboard into the case; I chose to do it before in this case.

Installing the graphics card is almost identical to installing the RAM, but make sure you take the back-plate off of the case before you try to install the card.

       Installing HDD is also fairly simple, but every case has a slightly different type of installation. This particular case came with screw-less installation. Basically, it’s some brackets that you install on each side of the drive, allowing it to slide in and out of the HDD are fairly easily, but hold firm enough to not cause any un-wanted vibration.

Installed Hard drives

     Installing the Power supply is straight forward. Just screw the unit into it’s position. I usually try to un-tangle the cables once I get it installed, and plan out which way I want the cables to run; so they do not obstruct air flow, and look kinda pretty to! Any unused cables can be run to the back side of the case (that has no window) so they will be hidden, and not obstructing air flow.

Running cables behind the motherboard tray

And that is all!

You are done building the computer, the only thing you may have to install are the bay devices (Removable HDD tray, CD/DVD drive)

Here is a picture of the finished product:

The finished product

Good luck building, and have fun!

- HDD Doctor

Building a PC is something almost everyone in the world of technology does at some point in time; some do it a lot more then others, and some even take a hobby to it (like myself). Being the hobbyist I am; I keep myself familiar with not only Data Recovery, but with all the latest parts for PCs being released, and all of the latest technologies with processors, graphics cards, motherboards, and all those other components that make up a PC. Recently, I was appointed to build a new Work PC for our Data Recovery Department here at Ji2, and I am going to walk you through the steps that I took to build a good PC for Data Recovery work.

Depending on the amount of cases you see and the way that you/your technicians like to work will play a big part on how much you will need to spend on your work PC; or if you are going to need to purchase duplicate parts to build multiple PCs.

Logical work is always an afterthought in the data recovery community; and I am not sure why because logical recovery is inevitable. No matter if you do a platter transplant or a PCB swap, in the end the goal is to be able to image the drive, and rebuild the files/file system on that drive to extract the data.

A problem I always see in our lab is this: physical work only takes a matter of minutes, but logical work can easily take hours, even days to image and rebuild the file structure. This can cause a major problem to arise; your workload becomes backed up because you have to wait for a drive to image or for your software the rebuild this 500GB hard drive’s file structure. This greatly reduces productivity. All of these things I took in to consideration when building a new PC. I wanted something that was fast, had a lot of expandability, and most of all, reliability. Reliability should be your number one priority. You will be working with customer data on this computer, and you cannot afford for it to crash. This puts customer data at risk, and also lowers productivity.

Alright, so I got started looking for all of the parts I wanted for my work pc. I didn’t really need to research much (I build PCs as a hobby, remember) so I already generally knew what was going to be the best route to go with brand, configuration, etc. Here is a pic of all of the parts after they arrived. Set up all fancy and everything to!

Fancy picture

So, you may be wondering; why did you choose these parts? Why do you need the case with the huge fans on it? The answer is: you don’t. Why did I get the case with the massive fans? Because it was a combo deal with the processor and huge fans mean one thing: huge airflow. Airflow leads to cooler components, and cooler components have longer lives and better reliability. For every one’s benefit; here is a list of all of the parts that I chose, and why I think that this part/brand is important, and if I have any alternative recommendations.

[Processor]-Intel Core 2 Duo e8500: If you have kept up on processors in the last couple years then you surely know that AMD had a big lead with their “Athlon 64″ processors up until about two years ago, when Intel released the beast that is the “Core 2″ series of processors. There is no arguing that these processors deliver, and Intel has been extremely aggressive with pricing. I chose a dual core and not a quad core for a very specific reason; compatibility. At this point in time you may run into compatibility issues with quad cores due to the programming of a lot of software used for data recovery. For that same reason; while a program may not crash while running on a quad core, most programs aren’t designed to run on a quad core and therefore do not utilize all four cores. At this time, a dual core is the best solution for the software DR technicians to use. I recommend getting the highest clocked (GHz) processor you can afford.

[Motherboard] - Asus P5Q Pro: Fairly straight forward; get a board that has a lot of SATA ports, and as many PATA ports as you can; it is hard to find boards with more then one these days. This board I chose because not only does is have 8 SATA ports; but it also has “8-phase power” which essentially helps clean up the currents running to your components; very nice for a DR PC. It also sports “Solid State capacitors” which have much longer life expectancy then standard capacitors.

[Hard Drive] - Samsung Spin-Point F1: For hard drives I looked no further then samsung. The spin-point F1 series is among the quickest 7200rpm drives you can have right now; and it is a general consensus that samsung makes some of the easiest drives to recover, due to their simple, but efficient design. I went with 3 750GB drives, and ran them in a RAID 5. RAID 5 essentially gives you the high read performance of RAID 0, with parity so that if one of the drives fails, it can rebuild itself. Keep in mind that if you want to run a RAID 5 you will need a minimum of three drives (get them identical in size) and that one of those drives is used only for parity, so the total amount of data you will be able to store is the combined total of two of the drives.

[Power Supply] - Corsair VX550: Power supplies get overlooked far too much. A lot of people figure they can go cheap on the worthless power supply to save some money. This is asking for all of your expensive parts and valuable data to go up in smoke. Never go cheap on the power supply. There are a couple things to consider when choosing a power supply. Watts is the most commonly known figure to look for on a power supply. Depending on how much hardware your running in your system you may want to go higher, but ~500watts give or take should be more the enough. The big number you want to look for is the amount of amps on the 12v and 5v rails. This corsair I those can feed something like 50amps into the 12v power rail. In comparison, one of the spinpoint F1 drives uses ~ 2amps on start-up. So you can see how 50amps gives you a lot of head room. The last number you want to look for is efficiency. You want a Power supply that has an 80% + efficiency rating. A lot of people don’t realize that just because your power supply says “500 watts” that does not mean that’s what it can actually output. Depending on heat, and the cleanliness of the current coming into the power supply it usually is much less ~20% less on a good power supply, which gives you your 80%. Keep all of this in mind when looking for a power supply and you should have no problem. Some of my personal favorite brands for PSU are: OCZ, Corsair, Seasonic, Thermaltake, and pc power and cooling.

[Memory] - Corsair XMS DDR2 800 4GB - Corsair has been making great memory for years; it has a lifetime warranty, and it was on sale. Memory is so cheap these days; max it out. On XP and Vista 32bit you will only see ~2.8-3.2gb depending on some variables, but that’s OK, at least cap it out. I would tell you to go with a 64bit operating system, but as with the quad-cores, that is something I would wait a while for because of compatibility issues with data recovery software. Other good brands are: OCZ, Kingston, Mushkin, Geil, Crucial.

[Graphics Card] - HIS ATI 2400pro: Graphics cards don’t do anything to speed up our DR work. Keep it simple and low profile; so it doesn’t use a lot of power. You may want to consider things like the video outputs if you want to run multiple monitors. This card is very solid and was under fifty dollars. Go with ATI or Nvidia, anything else is not very good.

[DVD Drive] - Samsung: DVD burners are only like thirty dollars these days; I picked the samsung, but there are quite a few good brands out there: Samsung, Sony/NEC, LG, Liteon, Asus to name a few. You may want to consider going with a Blueray burner in another half a year or so. With a 40GB capacity, it may come in handy for cheaper delivery media.

[Cooling] - Zalman CNPS 9700LED and Arctic cooling MX-2: As a hobbyist I have learned that keeping your CPU cool will help it live a nice long life. If you can keep you CPU running at around half of its “safe operating temperature” then chances are that processor should last you a good half a decade. The zalman cooler I chose is the most popular cooler on the market for PC Gamers, and games being as strenuous as they are; if you can cool a processor to play games, then you will be just fine cooling it for any other tasks. Arctic cooling MX-2 is what is called “Thermal compound”. This stuff goes in between the processor and the heatsink to promote better transfer of heat. This stuff is ceramic so it is non-conductive, and you don’t need to re apply it for over five years.

Alright, we went over all of the parts; next time I will walk you through the building process and give you some pointers for building a PC.

- The HDD Doctor

One of the common problems you will find when entering into the field of data recovery is that clean room work is only a small scope of the work you do. The rest can range from simple logical recovery, ATA password removal, firmware corruption, and PCB failures. Today I’m going to inform you a little about the latter.

Let me first say that if you are not familiar with the workings of a hard drive, or a computer for that matter; I do not recommend performing these procedures yourself. Even attempting something such as a PCB swap can put your valuable data at risk.

Before you completely swap a PCB, you must determine whether or not the PCB is the cause of the problem. To do this we must diagnose! If you have professional HD diagnostic tools on hand; this makes the job much easier. We are going to pretend we don’t; because not everyone can afford these tools as they can come at a steep price.

First, you are going to want the drive able to be powered on and in a place you can reach it. Once you have done this; power on the drive put your ear to it. For beginners, it may be hard to determine if the drive is making abnormal sounds. If the drive is making a repetitive clicking noise; this is most likely a problem more advanced then a PCB failure; and we will cover this another time.

If the clicking is present there is still a very small chance it may be the drive’s PCB; but let’s assume the drive had no clicking and continue.

Another scenario is that the drive does not power on at all; this is actually one of the most common signs of PCB failure. If your drive in question does not power on whatsoever; the following course of action needs to be taken:

First, you must obtain an identical drive, or as close as you can. Finding an identical match is different for every drive manufacturer, and model. Usually, you would want to match things like: model number, P/N (product number), Firmware revision, country of manufacturer, and site code. The site code can be difficult to locate, and understand, but with a little time and Google; you can find your answer.

Once you have located a match; it’s time to go to work. The first thing you will want to do is remove the PCB from the “bad drive” and put the suspected bad board on to your new “good drive” and observe if the same symptoms occur. If the drive seemingly starts up fine; the PCB may not be the problem and further diagnosis is required (which we will cover another time). If the drive exhibits the same behavior as the “bad drive” then it’s almost a sure shot that the PCB is the culprit.

Now we know that the PCB has failed. Most of the time, it has something to do with the fuses or some part of the power system of the chip. Sometimes, if you identify that the problem is a bad fuse you may bypass the fuse with a small piece of wire (electrically conductive). I don’t typically recommend this unless you absolutely cannot find a donor PCB; just because the fuse is what protects the important chips from damage. This brings us to the chip/chips that we DO need.

There are many scenarios of chips depending again on manufacturer, model, and the year that the drive was manufactured. The first chip is the EEPROM (Electrically erasable programmable read-only memory) which; besides being a mouthful to say, is essentially a form of flash memory that acts as the HDD’s BIOS it can contain important calibration data that is specific to that exact drive. Without going in to too much detail; every single hard drive is different, whether it be slightly misaligned, or had a couple bad sectors from the factory; the drive takes this into consideration and makes adjustments and saves them into this memory. This is why a board swap without swapping this chip typically does not work.

Another chip that may appear is the NVRAM (Non-Volatile random access memory). This is typically found on Hitachi drives; and contains extra data similar to the Calibration data found on the EEPROM. If you have to swap chips from Hitachi drives; the EEPROM and NVRAM will need to be swapped. Some drives may seem like they do not have an EEPROM; this usually means that the EEPROM is embedded within the CPU of the PCB; in this case you will just swap the CPU.

Swapping the chips may require slightly modified methods depending on the “form factor” of the drive. Some PCB may have “BGA” (Ball grid array) chip. This means that the solder is under the chip; and while you can remove these chips fairly easily, putting them back on requires a special “infrared” rework station. Expect to see BGA chips on smaller drives; 1inch and 1.8inch drives.

Assuming that the chips are not BGA chips, you will want to obtain a Heat gun capable of outputting ~260c (500f). You may want to pick up a variety of tips; particularly finer point tips so that you can be more precise where you are directing the heat. You will also want a good pair of tweezers for grabbing the chips you need to get off; as using your fingers would be a bad idea on many levels. Before you begin heating the chip be aware of surrounding objects such as capacitors and any other pieces of the chip that are soldered onto the board and will be under the scope of the heat gun; as they will be movable once heated and you must avoid them. Also make sure that the heat gun is not hitting the plastic where the interface connector is. It will in fact, melt. You must grab the chip while the gun is still running, because solder solidifies quickly and you will not have enough time to remove it if you simply heat it up and then attempt to grab it right after you turn the heat off.

Like this

or this

OK, so you have gotten all the chips needed off of the bad board; and prepared the good board by removing its chips. Now we must solder the EEPROM and any other chips needed on to the good board. For the best results you are going to want a soldering station that can output ~320c (603f). This will allow for the smoothest solder. As far as the tip of the soldering iron goes; find one that works best for you. Remember, you are working with very small chips so you need maximum precision; you don’t want to accidentally cook the top of the chip because you slipped. You will also need some sort of FLUX (used to remove oxidation) typically a Rosin flux pen will work just fine; just make sure it is not activated, because this can cause major damage to the board. Rosin-core Solder will be sufficient, or any solder labeled “for PC work”

example (not actually soldering here, just showing placement)

You won’t become a professional solderer overnight. If you have some old PCB you can practice on I would highly recommend this.

So that’s basically it. You removed the EEPROM and any other valuable chips off of the “bad pcb” and then soldered them onto the “good pcb”. Now all you have to do is install the soldered good pcb onto the “bad drive” and plug it in! If the pcb was the culprit; you will now be able to access your data. Keep in mind this is for data recovery purposes. I do not recommend using this technique as a “repair”. While the drive may seem rock solid; and it may be, it may also be very unstable and die at any time. I only recommend this technique for getting it up and running to extract the data to a safe source.

-HDD Doctor