Should I Try Swapping the Printed Circuit Board on My Failed Hard Drive?

by admin

Many peoples’ first response to a drive which they can no longer access is to try and replace the printed circuit board (PCB). The logic is clear enough, they have no idea what is actually wrong but here is something that can be tried easily and relatively cheaply, simply purchase another hard drive with the same model (some go to further lengths and find a donor which matches firmware version and is close in terms of date of manufacture to the defective drive). Unfortunately for the vast majority of those who try it, the results will be disappointing, once in a while it will work (and it is of course these people who post their results on online forums, the large majority who have paid for a replacement PCB and still don’t have access to their data tend not to bother, this gives the misleading impression that there is a realistic chance of success).

The first problem is one of accurate diagnosis, or more accurately lack of it. While PCB problems do make up a significant amount of the jobs seen by any specialist data recovery lab it is no more than around 10% of the total. Therefore the first obstacle that you face in replacing the PCB to reacquire access to the data on your drive is that it probably isn’t the PCB at fault.

For the purposes of this article let’s assume that in this instance the PCB is indeed the problem, and the only problem. It’s worth noting that electrical fluctuations which damage PCBs will also routinely damage the read/write head pre-amplifier circuit which is mounted on the head assembly itself inside the hard drive. Let’s further assume that you then purchase an exact match donor hard drive, it’s your luck day, you find a drive for sale that has the identical model, firmware and even rolled off the manufacturer’s production line on the same day. You swap the PCBs but still no access. Depending upon the particular maker and model you will usually find that the drive spins-up but then will click repeatedly or will make no sound at all after spin-up, in either case there is very unlikely to be any access to the data.
You will be understandably frustrated, everything about the donor drive is the same as the original defective drive, why on earth doesn’t this work?

Why Doesn’t the PCB Swap Work?

To illustrate why, let’s look at the example of a Hitachi HTS545050 hard drive.
A hard drive requires some start-up and configuration information in order to get itself up and running and then to present itself as a storage area to your computer, this information is called firmware. This firmware can be thought of as boot program, it includes factory-set parameters specific to the individual drive such as drive-unique calibration values calculated and stored on the drive during testing in the factory. It also contains information which is built-up over the lifespan of the drive such as the re-allocation map of failed sectors (all hard drives have a limited pool of spare sectors for use when a sector fails, but a record is needed to map the replacement sectors to those that failed, again this is unique to each individual drive). A hard drive’s firmware is both unique to the individual hard drive and essential for data access.

This firmware is stored partly on the platters inside the drive (on areas of the drive which are not accessible without specialist knowledge and tools) and, critically, on the printed circuit board.
Sometimes the PCB portion of the firmware is stored within the main processor chip and sometimes in a separate memory chip. In the case of this Hitachi drive, it is stored in the so-called NVRAM (non-volatile random access memory) chip. This is essentially just a fancy name for a chip that can be programmed and that holds its contents even when power is removed.

The NVRAM chip is indicated by a red arrow


To give some idea of the information held in this NVRAM chip let’s look at the raw contents:

Contents of the NVRAM chip, the PCB firmware


The data underlined in red shows the firmware version installed on this drive. The data in green shows the serial number. The remainder defines (among other things):

-Where the start of the remainder of the firmware is located on the platters (inside the drive), this varies from drive to drive even for drives of identical model and firmware version, why?- you’d have to ask Hitachi, but if this doesn’t match the actual start of the firmware stored on the platters then the drive will never initialise.

-The head-map. Often drives will have unused heads. For example there might be two platters inside the drive, therefore 4 platter surfaces and potentially 4 heads if all surfaces are in use, however often only 3 of the 4 will be in use and so the hard drive needs to know which head is not fitted (any one of the four might be absent). Any donor drive PCB NVRAM must have the same head-map as the original defective drive.

Hopefully it can therefore be seen why a straight PCB swap is a much more complicated proposition than it might at first appear.

There are ways around this of course. The donor PCB can be used to regain access to the original defect hard drive but only after the PCB portion of the firmware has been transferred from the defective to the donor PCB. This will involve either the use of specialist equipment or some expert soldering to re-locate the chip that holds the firmware.

So, should you try swapping the PCB on your failed hard drive?

Most of the time you certainly won’t do any harm (the notable exception is an early incarnation of Seagate Barracuda drives, specifically the Barracuda II family, the firmware on the PCB can be spontaneously overwritten). However the likelihood that it will allow you to access the user data on your failed drive is very small.

{ 2 comments… read them below or add one }

Michael Martin September 7, 2012 at 11:28 pm

Having had to recover data from drives that were in the early stages of failing (i.e. clicking randomly or intermittent boot-up), I can relate to the article above. I have been asked by many customers about swapping the PCB as it’s been viewed as quick fix. But after explaining to customers it’s not that easy, they always view my statement as a way to make more money. So I always advise people “When any major changes have been made to your date, create a system image at least quarterly and backup you system weekly. If your data is time critical and/or sensitive then backup daily or every two days. I have a home server which manages all my home laptop and desktop systems.

chris February 4, 2013 at 2:59 pm

its possible to solder out the nvram chip, but do you think this will help (with regards to the situation above)?

Leave a Comment

Previous post:

Next post: