toshiba mk2529 stiction or spindle symptoms, but NO sound

[mediaman]

Guys

this is a 1.8" micro sata drive. fusible link was blown and replaced. now disk will not spin up. There is no typical "servo strain" noise from the series of attempts a disk normally makes before giving up.

The current trace looks like something is going on.

Q. is it possible to be stiction or spindle with NO sound?

If no is the answer, I am guessing the PCB is bad.
Thanks for your time

Attachments

[Keatah]

Sounds like a bad PCB. Look there for the problem. If it was stiction, there'd be current spikes much greater then what is shown above.

[fzabkar]

I'm curious as to how your diagnostic tool can record a 244mA current draw from 12V (since the drive is rated for 700mA at 3.3V).

This leads me to ask, is the drive receiving power on its 3.3V SATA pins?
http://en.wikipedia.org/wiki/Sata#Micro_connector
http://www.computerupgradeking.com/ebay ... 529GSG.jpg

If you need help identifying the test points on your PCB, then post a detailed photo of the component side.

[mediaman]

Great question fzabkar. I made a voltage converter to power 3.3V out from my Atola Insight Box, it also passes 12v straight through, I would guess theses are parasitic signals. (Btw the converter is just a simple LM317T, works great).

I agree, the current spike would be larger with stiction or spindle. and they should stop after 5 or 6 tries as the respective controllers give up.

Any pointers where to look on the PCB, it all "looks" fine. I can also bet the adaptives are in the Marvel chip, so no easy swap.

Thanks!

[fzabkar]

I normally look for voltage test points in the vicinity of the transistors and coils/inductors. Coils are the landmarks for switchmode DC-DC converters, whereas NPN transistors are often used as pass transistors in linear regulators. There may also be three-terminal LDO regulators. I'd also be measuring the resistance at the output of each DC-DC converter. That should give you an idea whether the MCU and motor controller are alive. The latter IC also controls the DC-DC converters.

If you can find the current sense resistor(s) for the spindle motor, then the voltage across them should be proportional to the motor current. You can also locate the various onboard supplies by referring to the datasheets of known components. For example, pin #1 is usually the Vcc pin of an SDRAM, as is the bottom pin on the same side.

BTW, IMO Atola's developers should revisit their design. ISTM that any tool that measures 244mA into an open circuit has a problem with Ohm's Law.

[Dmitry Postrigan]

BTW, IMO Atola's developers should revisit their design. ISTM that any tool that measures 244mA into an open circuit has a problem with Ohm's Law.

That was most likely due to a spike from a capacitor charge on power on. There is a small cap after the current measuring sensor in our units.

[Vulcan]

@mediaman,

If, as I think you're saying, you use an LM317T to (try to) produce a +3.3V supply from a +5V supply from the Atola, then I suggest carefully measuring the +3.3V output. The LM317 isn't sold as an LDO regulator, and at 500mA, one manufacturer's version which I picked as an example (Fairchild) has approx 2V dropout (even greater dropout at higher current), and of course 5V (input) - 2V (dropout) does not equal 3.3V (output).

Therefore some of your problem might be an out of tolerance supply to the drive, due to a possible +3.0V (or lower!) supply voltage. To detect any dips in that voltage (especially while the drive tries to spin-up), you'd need a 'scope - a DVM wouldn't be sufficient IMHO, due to its slow updating of the display.

Or choose an LDO regulator instead of the LM317. Or use the +12V supply from the Atola as the input supply to the LM317 (if the Atola allows that, and doesn't require a minimum load on the +5V rail - if it does require that, you may need to provide it artificially).

I don't own an Atola; my background is electronics, so those are just a few thoughts, in case you're chasing a power problem as well as (or instead of) a drive problem. Good luck

[mediaman]

I really appreciate the feedback, thanks. It works perfectly as is, with the divider resisters that I had in my box the output is a clean 3.33V (most drives tolerances are much higher, but I do check each drives spec before trying).

It would be ideal if atola enabled a 3.3V output directly.

I understand what you mean about the 3.3 and 5.0 being too close together, I was tempted to feed 317 from the 12V line from atola. Maybe that will be a better idea. Then I can use the 5v straight through to the Microsata connector.

Incidentally: has anyone ever seen a drive that uses BOTH 3.3V and 5.0 as per the microsata interface standard?

[Vulcan]

the output is a clean 3.33V

As I explain above, based on your existing design and the specs of one typical LM317 manufacturer, that output voltage cannot be relied upon if the drive tries to spin-up. (Note my earlier comment about Vdo varying with current.) Of course YMMV depending on your specific LM317 manufacturer's specs. Anyway, I'll leave you to it.

[Dmitry Postrigan]

mediaman
You can try measuring 3.3V under load (300-500 mA) - this is where you will see the dropout.

If you move LM317 to the 12V rail - keep in mind that it will get VERY hot unless you have a good heatsink:
0.2 A * (12 V - 3.3 V) = 1.74 Watt of heat.

Consequently, should a spike in current draw occur - the LM317 will make fireworks (and will possibly pass 12V directly into 3.3V input of the drive).

[Vulcan]

@Dmitry,

You can try measuring 3.3V under load (300-500 mA) - this is where you will see the dropout.

Thanks, that's exactly what I tried to point out previously - Vdo increases with increasing current.

The specs I saw for this Toshiba drive say that during spin-up it takes 2.5W, so at 3.3V that means approx 750mA supply current (not for a long time, of course) which will cause an even greater Vdo, and therefore potentially an even lower supply voltage to the drive at that critical time. (Previously I gave the example Vdo at 500mA because the Fairchild datasheet shows Vdo at that specific current - the Vdo value would need to be extrapolated for 750mA, but will be greater than Vdo at 500mA).

Can you confirm that the Atola units do not require any minimum external load on their +5V or +12V supply rails for good regulation of the "other" supply voltage, so therefore it does not cause any problem if only one voltage rail is used? Thanks.

[Dmitry Postrigan]

Vulcan
+5 and +12 power lines in our units are completely independent. There is no minimum load required on one line for the other to operate. This means you could use only 12V rail to power the LM317 (or any other circuit).

[Vulcan]

@Dmitry,

That's great, many thanks for explaining.

[mediaman]

@Dmitry, please give recommended 3.3V output circuit. I would be happy to buy too if Atola wanted to make a nice microsata adapter (or maybe there is one already?)

[Dmitry Postrigan]

I would recommend any LDO regulator. Some good examples are MIC29150-3.3WU (or MIC29150-3.3WT) and LD29150DT33R (we have positive experience with both). Or, you could just buy an adapter from ebay: http://www.ebay.com/sch/i.html?_nkw=mic ... c&LH_BIN=1