Hitachi HDS723020BLA642 always different sounds - diagnosis?

[michael chiklis]

Hello, i've this Hitachi HDS723020BLA642 which isn't detect anymore since a power surge.
PCB seems to be ok, i don't see any burned component and any shorted diode with multimiter check.

PCB: 110 0A90302 01
MCU: 0J11389
Motor Chip: 0A73080

Drive spins, sometimes does 5 clicking sounds cand spins down, sometimes 1 clicking and spins down and sometimes 4 clicking with NO SPIN DOWN.
When heads do clicking sounds is because motor loses rotation speed half second before!!

Can you help me to identify the problem?
Is this a head, motor or pcb issue?

Does someone know what tension value to check on pcb?

[fzabkar]

Does someone know what tension value to check on pcb?

It doesn't sound like a PCB issue but here are the test points AISI.

Attachments

HDS723020BLA642_Vneg.jpg (106.57 KiB) Viewed 18663 times

HDS723020BLA642_Vcore_Vio.jpg (89.62 KiB) Viewed 18663 times

[fzabkar]

I think there may be a second Vio rail. Let's call it Vio2.

HDS723020BLA642_Vio2.jpg (43.22 KiB) Viewed 18658 times

Here are a couple of datasheets that should help identify some of the other components.

K4T56163QN, Samsung, DDR2 SDRAM, 256Mbit, 1.8V:
http://www.mt-system.ru/sites/default/f ... rev103.pdf

TC75W60FU, Toshiba, CMOS Dual Operational Amplifier, marking 5W60:
http://www.mouser.com/catalog/specsheets/TC75W60FK.pdf

Therefore one of the Vio supplies must be providing +1.8V for the SDRAM. The other Vio supply must be powering the NVRAM, possibly +2.5V or +3.3V.

[michael chiklis]

This are the value i detect with voltmeter

[fzabkar]

I got the two 100R coils in the Cuk converter the wrong way round, but -3V seems reasonable for the Vneg supply.

+1.86V seems about right for the SDRAM. I expect that the Vio2 supply would be +2.5V or +3.3V.

Vcore seems low, but you would need to compare it against a working PCB. It's most likely OK. In any case, if it were low, I would think that the motor controller wouldn't release the POR signal, in which case the PCB would be brain-dead.

[michael chiklis]

MCU an motor chip get pretty hot without pcb connected to the hdd, is it normal?

[fzabkar]

It appears that I was wrong about Vio2. I don't know what to make of that section of the circuit, unless it's some kind of charge pump for the motor and VCM MOSFET drivers.

Can you tell us the markings on the other two 8-pin ICs near the "5W60" chip?

As for the motor controller and MCU getting hot, I wouldn't think this would be normal.

[michael chiklis]

[michael chiklis]

Microscope screenshots

it seems that the inductor is burned, what do you think?

What do you think?

[fzabkar]

The flash memory is a 1.8V part.

LE25FS406, Sanyo, 1.65V - 2.1V, SPI flash memory:
http://panasonic.co.jp/sanyo/news/2010/02/16-3.pdf

The other chip appears to be an analog part, possibly a low power, rail-to-rail variant of the LM2903 dual comparator.

If I'm right about the charge pump, then AIUI the boost voltage should be around 18V. However, it could be that the motor controller may not generate this voltage unless commanded to do so by the MCU, and the MCU may not be commanding it to do so if it thinks that something is wrong, eg if the preamp is not being detected.

But that's just wild speculation. You would need to compare your readings against a working drive. Otherwise look for the four "12.6V" test points on the other side of the PCB and measure the charge pump boost voltage (Vcp) while the drive is spinning.

[fzabkar]

it seems that the inductor is burned

It's hard to tell. If it has shorted turns, then that could explain the heating and the absence of the boost voltage. However, a multimeter resistance test won't help. You would need to test it with an inductance (LC) meter.

Your photos show the ferrite core, however you should be examining the wire turns for any evidence of blistering in the insulation.

BTW, if you're concerned about the white spot, I believe it is just a manufacturing mark.

Last edited by fzabkar on August 10th, 2013, 20:48, edited 1 time in total.

[michael chiklis]

Unfortunately i don't have another pcb like this, i should buy it.
Do you know where can i buy it without spending $ 50?
DriveStar has completed the escort of pcb 110 0A90302 01.

I have an ESR meter, do you know if it's possible to check if inductor is shorted with it?
ESR meter can misure from 0 to 10 ohms

[michael chiklis]

ESR meter indicates the inductor is 1,65 ohm.
Multimiter ohmeter of course indicates 0 ohm

[fzabkar]

I suspect that these inductors would have an ESR of 100 milliohms or so. That said, I have Bob Parker's ESR meter. It is designed to test the ESR of capacitors. When it is applied to inductors, it still gives a good indication of whether they are shorted. In fact it can even detect the inductance of "curly cord" test leads. Your ESR meter may be different, though.

Why not backup the NVRAM (remember it's a 1.8V part, not 3.3V), and then measure the Vcp voltage on the drive?

[fzabkar]

ESR meter indicates the inductor is 1,65 ohm.
Multimiter ohmeter of course indicates 0 ohm

Seems OK to me.

[michael chiklis]

My ESR meter is a cheap one, i bought it 2 years ago on ebay (from Bulgaria).
I bought it to misure ESR on electrolytic capacitors, it's important to have a low ESR capacitors on high frequency switching circuits.
Sometimes i use it on inductors just to find if they're shorted.
This pcb inductor seem to not be shorted, or maybe has only few shorted coils, i can't tell it.

I have Willem 5.0 eeprom, can i use it to backup the NVRAM of this pcb or it will damage it?
As i know willem 5.0 generates 3.3V for SPI roms, i use it to backup and write seagate roms (example 25P40)

[fzabkar]

I wouldn't use a 3.3V programming voltage, especially not in-circuit. The 1.8V supply also powers the SDRAM and MCU.

BTW, I have re-examined the Vcp supply and I think it may work like this:

[michael chiklis]

What programmer do you use to power NVRAM in-circuit at 1.8V ?

[fzabkar]

Code:

+12V O--- )()()( ---+---|>|----+---O Vcp (+18V ?)
           coil     |  diode   |
                    |         _|_
                 ||-+         ___ capacitor
                 ||<- MOSFET   |
   PWM pulses ---||-+          |
                    |          |
GND  O--------------+----------+---O Ground

The MOSFET would be internal to the motor controller.

[fzabkar]

What programmer do you use to power NVRAM in-circuit at 1.8V ?

I don't have anything like that. Mine is an old Expro-60 Sunshine programmer. Sorry. :-(

In any case, do you really think you have a PCB fault?