Hi

I've been a lurker here for a while now, what a great forum! I have read and picked up quite a bit of knowledge from here so feel it is only fair to return some and share in my success in swapping SMOOTH IC and a shorted 10uF C12 cap.

I have a WD5000AAJS (taken out of a WD MyBook Premium external USB enclosure) that just suddenly stopped working without warning. It had been at the bottom of computer case for a while and not on for long periods (unlike other computers I have which are on 24/7), but probably long enough to stress the components. I only noticed it wasn't working when I came to copy some Adobe Video files off there. They were imported from MiniDV tapes, and whilst I still had the tapes, I would not have liked to have to re-import 25+ hour long tapes, especially considering the Canon XL1 MiniDV camcorder that originally created them had been sold a few years back.

After initial checks of just no spin-up, I set about looking for answers... Plenty of drives that had suffered overvoltage and blown components but this clearly didn't apply. This one had never been abused, although probably was running quite hot given the state of the foam (darkened and board was quite oxidised).

Tried shorting the jumpers to attempt PSV mode connecting to secondary SATA port but got no reaction from WDR Demo, clearly as the ST SMOOTH IC is responsible for providing power to the MCU (which had embedded ROM, no U12).

A had a FLIR C1 thermal imaging camera so thought it'd be interesting to see - it showed the SMOOTH L6284 3.0 getting warm (40deg C) but nothing else (at the time didn't know if this was normal or not).

On checking voltages (thanks to fzabkar and other posts around the internet) clearly showed no 3.3v or vcore, and no obvious shorts to ground on the outputs I had to conclude that the SMOOTH IC had failed internally in some fashion through overheating. A real shame there is no datasheet available for this IC.

I spent considerable time trying to locate a cheap matching PCB or donor drive but seems that they are all pretty rare now... Found an ebay listing for a drive exact same model, and a few days apart from the production date of mine, but the seller was overseas, asking far too much and wasn't responsive.

I could have spent £40+ on a PCB elsewhere but no guarantee of a direct swap and not having a dump of the ROM, nor the technology to read from the SA, so I decided to settle on looking for a drive with the same L6284 3.0 IC. That was easier said than done, but eventually found a seller with a WD800AAJS on eBay that was built in March 2007 so was fairly confident it would have roughly the same components in that area.

This really wasn't worth spending a lot on, as was a bit of a gamble, so this 80GB was a nice cheap donor drive!

£8 lighter and a day later the drive arrived, I first powered up to check that all worked and was fine, so removed the PCB and yayayayay it had the L6284 3.0 I needed, plus anciliary components should they be required (such as voltage regulators etc). A thermal image of just the board in PSV mode showed completely opposite to dead (marvel chip getting warm, smooth IC pretty cold).

Just needed to find the time to swap them over.

That time came a couple of hours ago, so having practiced reading the ROM off the donor with WDR Demo (to see what actually happens), I set about getting my preheating station, desoldering, soldering and hot air stations all going... Got the old board all nicely preheated and a bit of flux to help things along, got the hot air out and gently extracted the dead smooth IC.

In my earlier testing I had noticed there was 6V across C12 (18v one side, 12v the other, to ground) on the good board from the 80GB, the dead board this cap appeared to be short, but thought this short was in the smooth IC itself. However it turns out this capacitor was actually shorted, so that wouldn't have been good for anything either!

I lifted the IC + C12 from the donor board and whilst boards had these components off, took some some static resistance measurements to compare both boards and see if there was anything else obviously wrong, other than a slightly higher resistance to ground on the Vcore side (measured at R050 resistor) on the dead board everything else seemed ok...

Soldered the IC and C12 to the dead board, and a little cleanup and was ready to try it...

I put the jumpers on to enable 'PSV' mode and (HDA not attached) plugged the board into the computer with WDR ready and my thermal cam taking a peek, thoroughly expecting the worst and smoke to appear, plugged the power connector onto the board... saw Marvel MCU start to warm and unplugged again about 1 second later... glanced at WDR and saw status had changed, so plugged in again... Hit the ROYL option and got into FLASH to read the ROM (just in case anything did blow up, it may have been the only quick opportunity to grab the data). This was read OK!

So then a quick check of some voltages on the preamp connector (nothing scary, -5v and +5v seemed good), and attach the PCB back to the drive...

Had everything ready for a quick rescan in computer management and a folder ready to copy files to, and powered up the drive... Thermal camera watching proceedings, it spun up, initialised, and a rescan of drives showed everything there as normal!

An hour later (giving me time to write this) have now copied everything off this drive at about 65MB/s - it's working perfectly!

So to conclude, whilst a PCB swap and an adaptation service probably would have worked, I didn't really feel in a position to do this, and in this case, certainly a shorted C12 (and suspected L6284) was all that was needed to be swapped out to get it all going again.. I soldered the suspect L6284 back to the donor PCB and will find another 10uF smt cap some time just to see if it was JUST the capacitor that was short or indeed the L6284 had died too. It'd be interesting to know what purpose C12 has (with 18V at one side).

I hope this helps someone. I can close about 30 Firefox tabs now

Congrats; nice work!

_________________
Sabo Computer Repairs & Data Recovery

Very nice!


I believe C12 is the "boost" capacitor for the VCM H-bridge driver circuit. A boosted voltage is generated by a charge pump, and this voltage is used by the high side MOSFETs in the bridge.

See the datasheet for the L7250 SMOOTH controller (not the same chip, but similar):

http://pdf.datasheetcatalog.com/datashe ... Xyuswx.pdf

_________________
A backup a day keeps DR away.