Buck converter failures
in [Design]
|
Prev: TeeVee over Fibre
Next: Low noise guitar pre-amp / ADC
From: D Yuniskis on 12 Apr 2010 18:39 Hi, I've got an external SATA-USB cradle that gave up the ghost. 12VDC supply (used by the spindle motor) which is then dropped down to 5V (for the logic) via a buck converter (I am guessing this based on visually examining the foil traces -- I'll have to drag out the stereoscope to read the markings on the device). Note that the design is prone to the load being removed or applied at will -- can't expect folks to power the cradle down while inserting/removing (I mention this only in that it might be a contributing factor to the failure). Converter (SOIC8) gets nice and toasty! Voltmeter shows recirculating diode is shorted. This would be consistent with the overheating as the output of the converter would just be shorted through the diode! Is this a common failure (or am I making some bad assumptions)? I'll have to check the markings on the diode, as well. But, I assume it should be sized to see ~12V and ~whatever the current draw by the SATA drive is (note "~" as approx). So, roughly an amp @12V (without any derating). (of course, ripple current will have some impact on the *actual* current through the device) Any other issues? I assume this should be a reasonably fast diode, as well?
From: Joerg on 12 Apr 2010 19:04 D Yuniskis wrote: > Hi, > > I've got an external SATA-USB cradle that gave up > the ghost. > > 12VDC supply (used by the spindle motor) which is > then dropped down to 5V (for the logic) via a > buck converter (I am guessing this based on visually > examining the foil traces -- I'll have to drag out > the stereoscope to read the markings on the device). > > Note that the design is prone to the load being > removed or applied at will -- can't expect folks > to power the cradle down while inserting/removing > (I mention this only in that it might be a contributing > factor to the failure). > > Converter (SOIC8) gets nice and toasty! > > Voltmeter shows recirculating diode is shorted. > This would be consistent with the overheating as > the output of the converter would just be shorted > through the diode! > > Is this a common failure (or am I making some bad > assumptions)? > I haven't seen many of those diodes fail. > I'll have to check the markings on the diode, as well. > But, I assume it should be sized to see ~12V and ~whatever > the current draw by the SATA drive is (note "~" as approx). > > So, roughly an amp @12V (without any derating). > Huh? Did the 12V fail or the 5V? I can't imagine the logic needing a whole amp. > (of course, ripple current will have some impact on the > *actual* current through the device) > > Any other issues? I assume this should be a reasonably fast > diode, as well? Usually a Schottky, plenty fast. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: D Yuniskis on 12 Apr 2010 19:36 Hi Joerg, Joerg wrote: > D Yuniskis wrote: >> Hi, >> >> I've got an external SATA-USB cradle that gave up >> the ghost. >> >> 12VDC supply (used by the spindle motor) which is >> then dropped down to 5V (for the logic) via a >> buck converter (I am guessing this based on visually >> examining the foil traces -- I'll have to drag out >> the stereoscope to read the markings on the device). >> >> Voltmeter shows recirculating diode is shorted. >> This would be consistent with the overheating as >> the output of the converter would just be shorted >> through the diode! >> >> Is this a common failure (or am I making some bad >> assumptions)? > > I haven't seen many of those diodes fail. Well, here's one! :> Some research has turned up lots of failures -- though I can't see *how* those failures were caused. >> I'll have to check the markings on the diode, as well. >> But, I assume it should be sized to see ~12V and ~whatever >> the current draw by the SATA drive is (note "~" as approx). >> >> So, roughly an amp @12V (without any derating). > > Huh? Did the 12V fail or the 5V? I can't imagine the logic needing a > whole amp. 12V supply from wall wart. My comment was that the diode would see at least 12V across it. Current through it would (average) be the 5V load of the disk drive *plus* the logic in the USB-SATA converter itself. It seems pretty commonplace for a drive to use a significant fraction of an amp at 5V (like 0.75A). Even if the SATA bridge draws 50mA, you're still close to an amp. >> (of course, ripple current will have some impact on the >> *actual* current through the device) >> >> Any other issues? I assume this should be a reasonably fast >> diode, as well? > > Usually a Schottky, plenty fast. Existing diode is 3A 40V. *Seems* plenty big enough which leaves me wondering if there is some other "usage issue" that causes the failure. (?)
From: Joerg on 12 Apr 2010 20:54 D Yuniskis wrote: > Hi Joerg, > Joerg wrote: >> D Yuniskis wrote: >>> Hi, >>> >>> I've got an external SATA-USB cradle that gave up >>> the ghost. >>> >>> 12VDC supply (used by the spindle motor) which is >>> then dropped down to 5V (for the logic) via a >>> buck converter (I am guessing this based on visually >>> examining the foil traces -- I'll have to drag out >>> the stereoscope to read the markings on the device). >>> >>> Voltmeter shows recirculating diode is shorted. >>> This would be consistent with the overheating as >>> the output of the converter would just be shorted >>> through the diode! >>> >>> Is this a common failure (or am I making some bad >>> assumptions)? >> >> I haven't seen many of those diodes fail. > > Well, here's one! :> Some research has turned up > lots of failures -- though I can't see *how* those failures > were caused. > >>> I'll have to check the markings on the diode, as well. >>> But, I assume it should be sized to see ~12V and ~whatever >>> the current draw by the SATA drive is (note "~" as approx). >>> >>> So, roughly an amp @12V (without any derating). >> >> Huh? Did the 12V fail or the 5V? I can't imagine the logic needing a >> whole amp. > > 12V supply from wall wart. My comment was that the diode would > see at least 12V across it. Current through it would (average) > be the 5V load of the disk drive *plus* the logic in the > USB-SATA converter itself. > > It seems pretty commonplace for a drive to use a significant > fraction of an amp at 5V (like 0.75A). Even if the SATA > bridge draws 50mA, you're still close to an amp. > >>> (of course, ripple current will have some impact on the >>> *actual* current through the device) >>> >>> Any other issues? I assume this should be a reasonably fast >>> diode, as well? >> >> Usually a Schottky, plenty fast. > > Existing diode is 3A 40V. *Seems* plenty big enough which > leaves me wondering if there is some other "usage issue" > that causes the failure. (?) > That ought to do. The only thing that can happen is hot-plugging the wallwart which causes a ringing. Worse if not much capacitance at the 12V side. But I assume at 40V the switcher chip would have died as well. What size package is that diode? Hoping the 3A is a continuous rating. My hunch is a voltage spike. Most of the diode that I've seen after over-current failures were open. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: D Yuniskis on 12 Apr 2010 23:19
Hi Joerg, Joerg wrote: > D Yuniskis wrote: >>>> I've got an external SATA-USB cradle that gave up >>>> the ghost. >>>> >>>> 12VDC supply (used by the spindle motor) which is >>>> then dropped down to 5V (for the logic) via a >>>> buck converter (I am guessing this based on visually >>>> examining the foil traces -- I'll have to drag out >>>> the stereoscope to read the markings on the device). >>>> >>>> Voltmeter shows recirculating diode is shorted. >>>> This would be consistent with the overheating as >>>> the output of the converter would just be shorted >>>> through the diode! >>>> >>>> Is this a common failure (or am I making some bad >>>> assumptions)? >> >>>> (of course, ripple current will have some impact on the >>>> *actual* current through the device) >>>> >>>> Any other issues? I assume this should be a reasonably fast >>>> diode, as well? >>> >>> Usually a Schottky, plenty fast. >> >> Existing diode is 3A 40V. *Seems* plenty big enough which >> leaves me wondering if there is some other "usage issue" >> that causes the failure. (?) > > That ought to do. The only thing that can happen is hot-plugging the > wallwart which causes a ringing. Worse if not much capacitance at the > 12V side. But I assume at 40V the switcher chip would have died as well. No idea what sort of abuse this saw before it ended up in my lap (broken). With the plethora of different wall warts out there -- many with the same barrel power connectors -- who knows *what* may have been plugged in, here! (though it seems like this device has a bad reputation so it could just be a sh*tty design!) > What size package is that diode? Hoping the 3A is a continuous rating. It's an SS34. Most of the data sheets show it as a DO214 package. But, this is *much* smaller. I'm guessing 2mm x 5mm? (I'll have to pull it to measure it. The PCB is recessed within the enclosure so hard to get at it even with a dial caliper). > My hunch is a voltage spike. Most of the diode that I've seen after > over-current failures were open. <shrug> I'll start looking for something a bit bigger (after measuring the pad locations and checking clearance around it). There's not much else that *could* go wrong (unless someone plugged in a faulty drive? I don;t think the little converter has over current protection (?)) (FP6101 -- I'll go dig up a datasheet on it) Thx, --don |