There I fixed it..but don't know why or how.
May 5th, 2012, 3:14
I recently had a a WD MyBook go bad, so I took it apart and soldered some stuff inside. And now it works. Living proof that DIY can and does work!
Re: There I fixed it..but don't know why or how.
May 5th, 2012, 6:35
This means you are going to offer this kind of "service" or "help" or what is the purpose of the thread ?
Unknown failure (if there was a failure, after all) , unknown procedure...
Unknown failure (if there was a failure, after all) , unknown procedure...
Re: There I fixed it..but don't know why or how.
May 5th, 2012, 7:42
Just to show that DIY can work. And that even strange unorthodox procedures can work too! I did the MacGyver approach to fixing up this drive using a Telephone Pickup coil ($5 from RadioShack) and a Zippo lighter and flux, solder, and vodka.
Symptom - drive not spinning up. No sounds of any kind.
I verified good power to the drive. Ok.
I used the TP coil to verify there was no stiction. Check ok.
I also used it to verify activity on the bridgeboard. That checked ok too.
I then used it to listen in on the motor and voicecoil. No activity there.
Finally I used it to check which chips were working on the controller. Here's where it got interesting. With the coil placed here and there at various spots, I could press and flex the board and the speakers would squeal and scream like a banshee. This was the speakers and the amplifier picking up EMF through the coil, not the hard drive mechanism. I've head them squeal too.
The marvel chip made the most sound, the cache-ram didn't do anything other than buzz, the motor speed controller wasn't jerking the spindle like it normally does. Most important pressing in a certain area would shut the whole board off, and silence fell upon the lab. I couldn't get any activity till I power cycled it again.
Ah hah!! Cracked solder joints Someplace. Instead of trying to trace it out and pour through datasheets and sit all day measuring things; I used my lighter to heat the underside of the board and re-flowed each of the chips. All at once, got every connection re-flowed nicely.
The serial rom, cache, marvel controller, smooth motor controller, and all those other little parts. Yep. I used plenty of flux and gave it a Svedka vodka bath after that. That and some plain water. I blasted it dry with the air compressor. I inspected everything with a magnifying glass and noted quite an improvement of the finish of the solder. It now had a glistening silver color to it instead of the flat grey finish. Good.
And the drive powered up just fine! This was strictly an electronics problem and had nothing to do with the HDA or any mechanicals. To verify the good repair I pressed here and there an all around with popsicle stick. The board didn't cut out or anything. And the marvel chip kept buzzing no matter how hard I pushed. Earlier it had stopped if I pushed in a certain area. Now it was cooking with gas fer'sure!
The smooth motor chip now gives a good blast of power and then the platters spool up smartly and smoothly. You can actually here booming and thunder through the amplifier if you put the coil on the chip or motor spindle. You gotta try the butterfly test or a skipping sequential test. Awersome sound effects. Sound effects that can tell you a lot about the drive without having to open things up.
Well now, the drive shows up in windows and all seemed well. This being an electronic problem, the data was never in danger to begin with. But I ran a read-only surface test just to be sure and that worked. I even checked the smart data and that was ok too.
You know how auto mechanics sometimes use steth-o-scopes? Well the TePu-C is exactly that except it works with EMF and radiation! Go on! Try it. Have a listen in on various parts of the disk. You can even tell when the heads retract & park, where they are positioned, if there is servo calibration going on. You can tell when they advance from track to track even. AND, the fun part is with the motor spinning up. Makes like a hammer sound to get the platters rolling and then Vrrrroooooom!! It speeds up like a jet. And the dynamic braking, you can tell if it is working or not too. It either makes a sound like a distant truck rumbling forever or a fast TTTSSSHHHHZZZZZzzzoooooommmbbbbbbb - like a deceleration sound. You have to have a listen because I can't spell these sounds all that well.
More astounding, you can tell if the bridgeboard's power saving and monitoring is working. It makes a distinct quiet warble-whistle like a machine going into idle after the drive spins down.
Finally, as the last experiment of the day. I ran a defrag on a regularly used drive. I turned on and off some caching software. And the difference is night and day. The amount of head movements are cut down to at least 1/3rd of normal. Think of the wear and tear being saved. And it goes faster too.
Symptom - drive not spinning up. No sounds of any kind.
I verified good power to the drive. Ok.
I used the TP coil to verify there was no stiction. Check ok.
I also used it to verify activity on the bridgeboard. That checked ok too.
I then used it to listen in on the motor and voicecoil. No activity there.
Finally I used it to check which chips were working on the controller. Here's where it got interesting. With the coil placed here and there at various spots, I could press and flex the board and the speakers would squeal and scream like a banshee. This was the speakers and the amplifier picking up EMF through the coil, not the hard drive mechanism. I've head them squeal too.
The marvel chip made the most sound, the cache-ram didn't do anything other than buzz, the motor speed controller wasn't jerking the spindle like it normally does. Most important pressing in a certain area would shut the whole board off, and silence fell upon the lab. I couldn't get any activity till I power cycled it again.
Ah hah!! Cracked solder joints Someplace. Instead of trying to trace it out and pour through datasheets and sit all day measuring things; I used my lighter to heat the underside of the board and re-flowed each of the chips. All at once, got every connection re-flowed nicely.
The serial rom, cache, marvel controller, smooth motor controller, and all those other little parts. Yep. I used plenty of flux and gave it a Svedka vodka bath after that. That and some plain water. I blasted it dry with the air compressor. I inspected everything with a magnifying glass and noted quite an improvement of the finish of the solder. It now had a glistening silver color to it instead of the flat grey finish. Good.
And the drive powered up just fine! This was strictly an electronics problem and had nothing to do with the HDA or any mechanicals. To verify the good repair I pressed here and there an all around with popsicle stick. The board didn't cut out or anything. And the marvel chip kept buzzing no matter how hard I pushed. Earlier it had stopped if I pushed in a certain area. Now it was cooking with gas fer'sure!
The smooth motor chip now gives a good blast of power and then the platters spool up smartly and smoothly. You can actually here booming and thunder through the amplifier if you put the coil on the chip or motor spindle. You gotta try the butterfly test or a skipping sequential test. Awersome sound effects. Sound effects that can tell you a lot about the drive without having to open things up.
Well now, the drive shows up in windows and all seemed well. This being an electronic problem, the data was never in danger to begin with. But I ran a read-only surface test just to be sure and that worked. I even checked the smart data and that was ok too.
You know how auto mechanics sometimes use steth-o-scopes? Well the TePu-C is exactly that except it works with EMF and radiation! Go on! Try it. Have a listen in on various parts of the disk. You can even tell when the heads retract & park, where they are positioned, if there is servo calibration going on. You can tell when they advance from track to track even. AND, the fun part is with the motor spinning up. Makes like a hammer sound to get the platters rolling and then Vrrrroooooom!! It speeds up like a jet. And the dynamic braking, you can tell if it is working or not too. It either makes a sound like a distant truck rumbling forever or a fast TTTSSSHHHHZZZZZzzzoooooommmbbbbbbb - like a deceleration sound. You have to have a listen because I can't spell these sounds all that well.
More astounding, you can tell if the bridgeboard's power saving and monitoring is working. It makes a distinct quiet warble-whistle like a machine going into idle after the drive spins down.
Finally, as the last experiment of the day. I ran a defrag on a regularly used drive. I turned on and off some caching software. And the difference is night and day. The amount of head movements are cut down to at least 1/3rd of normal. Think of the wear and tear being saved. And it goes faster too.
Re: There I fixed it..but don't know why or how.
May 5th, 2012, 9:44
This was not exactly a failure and on some part it seems more like of a fictitious story to me. But it's an opinion , this time let's assume "bona fide".
I still stick with professional instruments in any case , either if it is data to be recovered or another puppy to refurbish.
I still stick with professional instruments in any case , either if it is data to be recovered or another puppy to refurbish.
Re: There I fixed it..but don't know why or how.
May 5th, 2012, 18:57
This is absolutely true. If I was going to make up a story I'd say something like:
"I overclocked my system so fast it made a singularity and sent its computations into the future. And the system worked on the principle of reverse relativity 'catching-up' with the output, thereby using all the time in the future to give the impression of making trillions of calculations in the here and now instantly. If you overclock too fast, the system sucks itself through the space-time continuum into nothingness.."
Now, I do indeed fully appreciate the complexity and mechanical tolerances with which modern drives are built. Most certainly not like the first RAMAC350 of yore; where technicians would repair surfaces with sandpaper and brush on some new oxide. Today's drives have 500,000 tracks per inch compared to the 20tpi back then. To fix today's drives in that manner, you'd need a microscopic moonbuggy with another microscope on-board it and a set of fine-ballpoint pens, size being relative of course! All shrunken down to where a head would appear the size of the Boeing factory as compared against a car. Well.. Think about it.
I figured I'd work on this drive without "proper" tools because the worst that could happen would be loss of the drive itself, the data was stored elsewhere. So no need for a pro operation.
What I want to know is how many controller boards are failing in this manner? Having intermittent connections and solder joints.. How many controllers can be saved without the need for getting donors and swapping the serial roms? How many bad connections are made from different material expansion rates when exposed to temperature changes? E.g. the metal of the HDA housing vs. the fiberglass PCB material, vs. the solder itself.
I see that there is an endless supply of failures with USB connector problems, but what about this issue? How many controllers can be fixed by re-doing all the solder all at once?
When I used my microscope and pried at some of the IC pins, they more or less moved without a lot of force. The solder seemed to behave as if it were thick paste as opposed to a solid "piece" of metal. That's because of the nickel and tin and lead-free composition. RoHS regulations. Not to mention a low melting point for easier and efficient fabrication at the factory. In comparison, I had a sound card from the 1990's laying around. I poked at its connections and they were rock hard, not soft and flexible. The way it should be!
Sure I could have set up the Metcal station and air gun. But I wanted to try it this way, with stuff that could be purchased at Wal-Mart. Anyhow. I had no idea what component(s) needed the reflowing. So I did them all. But it worked! And each solder joint now has a silver-like sheen. The way it should be.
I've now tested the drive and external housing. I let it run all night with a surface test. The next day I copied files for a couple of hours. Made sure it got nice and hot too. All the way up to 52`C! Then I put it in the refrigerator for a while and ran it while it was at 6`C. Everything seemed to be fine.
And that is that!
"I overclocked my system so fast it made a singularity and sent its computations into the future. And the system worked on the principle of reverse relativity 'catching-up' with the output, thereby using all the time in the future to give the impression of making trillions of calculations in the here and now instantly. If you overclock too fast, the system sucks itself through the space-time continuum into nothingness.."
Now, I do indeed fully appreciate the complexity and mechanical tolerances with which modern drives are built. Most certainly not like the first RAMAC350 of yore; where technicians would repair surfaces with sandpaper and brush on some new oxide. Today's drives have 500,000 tracks per inch compared to the 20tpi back then. To fix today's drives in that manner, you'd need a microscopic moonbuggy with another microscope on-board it and a set of fine-ballpoint pens, size being relative of course! All shrunken down to where a head would appear the size of the Boeing factory as compared against a car. Well.. Think about it.
I figured I'd work on this drive without "proper" tools because the worst that could happen would be loss of the drive itself, the data was stored elsewhere. So no need for a pro operation.
What I want to know is how many controller boards are failing in this manner? Having intermittent connections and solder joints.. How many controllers can be saved without the need for getting donors and swapping the serial roms? How many bad connections are made from different material expansion rates when exposed to temperature changes? E.g. the metal of the HDA housing vs. the fiberglass PCB material, vs. the solder itself.
I see that there is an endless supply of failures with USB connector problems, but what about this issue? How many controllers can be fixed by re-doing all the solder all at once?
When I used my microscope and pried at some of the IC pins, they more or less moved without a lot of force. The solder seemed to behave as if it were thick paste as opposed to a solid "piece" of metal. That's because of the nickel and tin and lead-free composition. RoHS regulations. Not to mention a low melting point for easier and efficient fabrication at the factory. In comparison, I had a sound card from the 1990's laying around. I poked at its connections and they were rock hard, not soft and flexible. The way it should be!
Sure I could have set up the Metcal station and air gun. But I wanted to try it this way, with stuff that could be purchased at Wal-Mart. Anyhow. I had no idea what component(s) needed the reflowing. So I did them all. But it worked! And each solder joint now has a silver-like sheen. The way it should be.
I've now tested the drive and external housing. I let it run all night with a surface test. The next day I copied files for a couple of hours. Made sure it got nice and hot too. All the way up to 52`C! Then I put it in the refrigerator for a while and ran it while it was at 6`C. Everything seemed to be fine.
And that is that!
Re: There I fixed it..but don't know why or how.
May 6th, 2012, 1:56
Hilarious
Re: There I fixed it..but don't know why or how.
May 6th, 2012, 5:04
Keatah wrote:When I used my microscope and pried at some of the IC pins, they more or less moved without a lot of force. The solder seemed to behave as if it were thick paste as opposed to a solid "piece" of metal. That's because of the nickel and tin and lead-free composition. RoHS regulations. Not to mention a low melting point for easier and efficient fabrication at the factory. In comparison, I had a sound card from the 1990's laying around. I poked at its connections and they were rock hard, not soft and flexible. The way it should be!
I think you have it the wrong way round. Lead-free solder melts at a higher temperature than leaded solder. That's why many manufacturers are/were experiencing QC problems. In fact you can find many discussions relating to BGA soldering issues, for example. There have also been several discussions at sci.electronics.repair where technicians have observed poor soldering in RoHS affected equipment.
If you examine recent Seagate and WD PCBs, you will see substantial oxidisation, particularly around the preamp contacts. I believe this is a consequence of the materials changes mandated by RoHS.
Another potentially more serious problem related to lead-free solder is "tin whiskers".
See http://en.wikipedia.org/wiki/Whisker_(metallurgy)
"In computer disk drives [tin whiskers] can break off and cause head crashes or bearing failures."
Ironically, unsafe substances are allowed to be used in those applications where safety is paramount, namely medicine, defence, and aerospace.
"Tin whiskers ... caused the failure of the Galaxy IV satellite in 1998."
"Pacemakers have been recalled due to tin whiskers."
"Tin whiskers ... have been found upon examination of failed relays in nuclear power facilities."
Re: There I fixed it..but don't know why or how.
May 6th, 2012, 5:41
Keatah wrote:Just to show that DIY can work. And that even strange unorthodox procedures can work too! I did the MacGyver approach to fixing up this drive using a Telephone Pickup coil ($5 from RadioShack) and a Zippo lighter and flux, solder, and vodka.
Well what can you say to that except... TTTSSSHHHHZZZZZzzzoooooommmbbbbbbb!
S
Powered by phpBB © phpBB Group.