Hi everyone,

I am a student currently working on a personal project for educational purposes. I've recovered an old Seagate Barracuda ST3000DM001 (P/N: 1E6166-570) that is not working, and I am using this as an opportunity to learn basic electronic diagnostics.

To be completely transparent, I have no formal electronics training. I am working with an old multimeter and an old oscilloscope, and I am trying to learn how to troubleshoot a PCB. I have attached a few photos of the board.

Here is what I have observed so far:

Power: 12V and 5V are correctly reaching the PCB via the SATA connector.

Motor: The motor does not spin at all (no sound). When I probe the 3 motor pins (downstream from the SMOOTH chip) with my oscilloscope, the signal remains flat while the PCB is powered.

Components: I tried checking the 12V and 5V TVS diodes with my multimeter. They appear "OK" based on my basic readings, but I am not entirely confident in my methodology. Is there a better way to test them, perhaps using an oscilloscope?

What are the "must-check" points on this specific model?
Are there any common failure points known for this board that I should inspect with my multimeter?
Should I be actively hunting for shorts?

Any guidance or tips to help me progress in this diagnostic would be greatly appreciated. Thank you in advance for your help and for your indulgence with a complete beginner.

Best regards,
Zebananos

What voltages do you measure across each of the two TVS diodes?

_________________
A backup a day keeps DR away.

Next to the SMOOTH chip, there are large gray coils with two terminals. They should show something, about a volt or more. If all coils show 0 volts, it's worth checking whether the voltage is reaching the TVS diodes. Repairing such boards usually ends with removing the diodes shorted by voltage surges and replacing the burnt-out resistors near the power connector with jumper wires. If the drive still refuses to spin up, the board is usually thrown in the trash and replaced with a working one from a donor drive. No one thoroughly examines these boards, so you'll have to be the first.

There's also an 8-pin chip on the board; this is a ROM. One of its contacts should show 2.5 or 1.8 volts. If the contents of this chip are erased or damaged, the drive also won't spin up.

Judging by the obvious mechanical damage to the board (one of the inductors is broken), this board will never work again - perhaps the contacts under the smooch chip were broken by impact.

I'd also like to point out that these boards have very few components. You can check the capacitors for shorts and measure the resistors (ideally, using another similar, working board as a reference). However, they rarely fail. If the SMOOTH chip had accessible pins, you could measure their resistance relative to a similar chip on another, working board. But that's not the case.

I beg to differ. The board will still work if the coil is not open circuit.

The group of 3 coils are associated with buck converters for Vcore and Vio. The 6R8 coil and the associated diode (E0/U4) constitute the negative supply (-5v) for the preamp.

_________________
A backup a day keeps DR away.

Hi fzabkar and others,

Thank you for the guidance — here is my full set of measurements. I am using an old Metrix MX1 multimeter and a 2000s-era oscilloscope, so please treat these values as approximate.

Coils continuity (unpowered, diode/ohm mode): All 4 black coils beep, indicating continuity on all of them, including the visibly damaged one. Electrically, the windings appear intact (not open).

Coils voltage (powered): 0V on the two unmarked coils and on the "6R8" coil; only ~300mV on the "3R3" coil.

TVS diodes (unpowered): OL in one direction, ~200Ω in the other, identical for both TVS (see attached TVS_1 / TVS_2). When powered, I measure approximately 5.02V and 12.4V across them respectively.

ROM (8-pin, powered): Two pins around 600mV, one at approximately 400mV, the rest at 0V (with noise). None show the expected 1.8V or 2.5V.

Short circuits: No continuity beep between ground and 12V, or ground and 5V.

E0/U4 diodes (unpowered): Open circuit in one direction, ~50Ω in the other (see attached E0_U4_1 / E0_U4_2).
Note: My multimeter beeps in diode mode for these E0/U4 diodes, but does not beep for the TVS diodes. Is this difference normal or does it indicate a specific issue?

What do you think of these results? I hope it's not the SMOOTH chip that's malfunctioning. I've attached a close-up photo of the SMOOTH chip and coils.

Furthermore, I have the impression that the varnish has overheated on the trace connecting the SMOOTH chip to the 3R3 coil. Does this seem suspicious to you?

Thank you in advance for your time and expertise.

Best regards,
Zebananos

It does look like a SMOOTH failure. The most expedient solution would be to move the 8-pin ROM IC to a donor PCB. However, be aware that if you damage this IC, data recovery will become extremely expensive, if it is possible at all. PCB vendors such as hdd-parts.com include a free firmware transfer in the price.

_________________
A backup a day keeps DR away.

Overheated varnish looks different - it takes on a dark yellow or brown tint, depending on how hot it was. I can search my archive for photos of what this looks like if you're really interested, but what you see in the photo isn't evidence of overheating, but corrosion - moisture has gotten under the varnish. Scrape it off with something sharp and metallic, and the varnish will easily peel off, revealing oxidized copper underneath.

I think the board is completely destroyed, beyond repair. Why is that? The ROM chip should have 1.8 volts on one of its contacts. Why exactly that much? Because next to the ROM chip, there's another rectangular chip without pins, made by Samsung. This is DDR2 RAM. It also operates at 1.8 volts and must be powered; without working RAM, the microprocessor won't work. To reduce production costs, chips were selected that operate at a single voltage, which is standard practice for such boards.

One of the required voltages is missing. More precisely, there are no necessary voltages on the board at all, except for those coming from the power connector. The SMOOTH chip is responsible for all the voltages and all the hard drive components that need to be moved. That's why it's suspected of being faulty. It's possible that it's not the chip that's faulty, but rather a component nearby, but for further diagnosis, we need to measure what's feeding this chip.

This chip doesn't have any pins accessible for measurement. There are small yellow elements near the chip. On some of them, at one of the contact pins, there are 5 volts and 12 volts coming from the power connector. You can also take measurements at the yellow dots with drilled holes if you cannot find the voltage on the capacitors. Considering that there was a tiny voltage on the ROM chip, these voltages were likely coming into the chip.

And now the real problems begin.

This is an internal measurement, so it's difficult to tell whether the diodes are intact or not, but common sense dictates that these diodes don't usually fail, and certainly not two at a time.

The real repair begins with a visual inspection. We've already seen the damage in a place where it shouldn't have been, so the procedure begins by checking all tantalum capacitors for short circuits and resistors for deviations from their nominal resistance.

Since, again, all these measurements are performed internally, some resistors will have unusual resistance values, just like capacitors, even though they shouldn't conduct current. The problem is compounded by the fact that an analog meter isn't very suitable for these measurements. Of course, you can diagnose a board with it, but you'll have to use the same measuring device to take measurements on a working board. I probably have a working board with similar components, but if I measure it with my digital multimeter, the readings will differ from the analog one because these devices supply different voltages for measurement, which will lead you down the wrong path.

In a real repair, the quick and correct step is to replace the SMOOTH chip with a working one, or install the suspected faulty chip on a working board. Further actions depend on the outcome - either the faulty chip was the problem, or the board itself. If the board is the problem, then all components and contacts are checked against the working one, and the discrepancy will sooner or later be found, and with it, the source of the problem.

What do we have in the end? We can't take measurements on the board because we don't have the right equipment, there's no donor board, and there's no way to quickly diagnose the board by replacing the suspected chip. I'm missing the point that this particular chip on this board is difficult to replace because it's soldered onto lead balls underneath.

As a last resort, it would be worth looking at the back of the board for damaged traces.

This was meant to be a learning project, not a case of data recovery or disk operation (yeah, ST3000DM001, the old legendary Grenadas that failed en masse due to low-quality manufacturing), and judging by the condition of the drive and the fact that the sticker where the screw of the magnetic head unit is hidden was torn off, the hard drive had already been taken from a trash heap.

As a final, easy-to-fix cause of the malfunction, you can check the integrity of this trace at the locations marked in the image. Also, look for the hole on the back of the board and measure whether there's contact between the two sides of the board. Test your luck; the problem might simply be a slightly rotted trace or a corroded via.

Thanks for your replies. I was hopeful after cleaning that via—which was indeed full of oxidation—but unfortunately, it didn't change anything.

I'm attaching a screenshot of the measurements I retook a little while ago, though I'm not sure they provide much extra information.

Incidentally, the broken coil is starting to come loose. I initially didn't think it was the root cause of the problem since it still passes the continuity test (it beeps). Still, could a damaged coil be the source of the whole issue despite the continuity?

I'm not really sure how to clean the board any further. That via was the only visibly oxidized spot, and the rest seems minor. At this point, it must indeed be a deeper issue with the SMOOTH board, and I think I've reached the limit of what I can test on my end...

Thanks again for the help!

PS: The points marked with "?" correspond to voltages that dropped from 12V to 0V during my measurement (capacitors discharging?).

Not really. The SMOOTH chip has the necessary voltages near it, but it doesn't output anything. You could also try using an oscilloscope to measure the coils on the contacts facing the chip. If pulses appear briefly there immediately after power is applied to the board, it's still trying to function.

You can try household chemicals, such as a rust and water stain remover (the best solution available for rotted circuit boards), and then wash it with dishwashing liquid. Shake out any remaining water (it will be hidden under the chips) and let the board dry. Considering the board is already completely destroyed, you can freely experiment with it.

Maybe yes, maybe no. Considering the board was rotted and damaged, it was clear without any diagnostics that it was all over. As a bonus, I found an identical working board and measured all the resistances and voltages on it. Keep in mind that my multimeter had a dead battery and the probes were broken, so the readings may be slightly different, especially the voltages. Where the image shows 2.19 volts, it's actually 1.8 volts. And where it's slightly more than a volt, it's actually a volt or slightly lower. You can compare it with your board if you're willing and have the time.

I accidentally missed the answer to this in my post. Yes, it could be the source of the problem, because it's not just a component with copper wire wound into a coil for aesthetic purposes. However, when the gray plastic simply falls off the inductor, the hard drive can usually function in this condition. To rule it out as a possible culprit, it simply needs to be replaced. You can, of course, also check that the coil hasn't fallen off the board by measuring the connection from the capacitors on one side to the three-pin component on the other, but that's definitely not the cause.

It's ferrite, not plastic. The inductance of the broken coil will be less than a new one, but the effect will not be too serious.

https://en.wikipedia.org/wiki/Ferrite_core

_________________
A backup a day keeps DR away.

Hi,

Quick update. Unfortunately my current multimeter (Metrix MX1) isn't precise/reliable enough to properly measure the resistor values shown in ora's reference photos — the readings are too inconsistent to trust for a meaningful comparison. I'll get a more recent/accurate multimeter before attempting that comparison properly.
In the meantime, I ran the oscilloscope test on the 4 coils, using single-sequence trigger mode on my Tektronix TDS 210, triggering on the power rail rising edge to catch the power-up moment. I numbered the coils as in the attached diagram, screenshots attached for each:

Coils 1 and 2: similar signal, present only right at power-up, nothing afterward.
Coil 3: the only one showing square-wave-like pulses and varying waveforms that actually persist even after the power-up moment, not just a brief transient.
Coil 4: 1 or 2 pulses at power-up, different shape from coils 1 and 2, but nothing afterward either (unlike coil 3).

What do you think this tells us? Does coil 3's persisting activity suggest that particular converter stage is still switching/trying to regulate, while the others (1, 2, 4) only got a brief kick at power-on and then gave up?

While waiting for your thoughts and a better multimeter, I'm going to do a full clean of the board with white vinegar as suggested.
Thanks again for your help,
Zebananos

Here are the attachments.

I should clarify that the measurements were taken between ground and the pin of each coil on the SMOOTH chip side, and that for coils 1 and 2, the signal is below 25 mV, as I have to lower the trigger below that value to get the measurement to trigger.

Hi,
Quick update. Unfortunately my current multimeter (Metrix MX1) isn't precise/reliable enough to properly measure the resistor values shown in ora's reference photos — the readings are too inconsistent to trust for a meaningful comparison. I'll get a more recent/accurate multimeter before attempting that comparison properly.
In the meantime, I ran the oscilloscope test on the 4 coils, using single-sequence trigger mode on my Tektronix TDS 210, triggering on the power rail rising edge to catch the power-up moment. I numbered the coils as in the attached diagram, screenshots attached for each:

Coils 1 and 2: similar signal, present only right at power-up, nothing afterward.
Coil 3: the only one showing square-wave-like pulses and varying waveforms that actually persist even after the power-up moment, not just a brief transient.
Coil 4: 1 or 2 pulses at power-up, different shape from coils 1 and 2, but nothing afterward either (unlike coil 3).

What do you think this tells us? Does coil 3's persisting activity suggest that particular converter stage is still switching/trying to regulate, while the others (1, 2, 4) only got a brief kick at power-on and then gave up?
While waiting for your thoughts and a better multimeter, I'm going to do a full clean of the board with white vinegar as suggested.
Thanks again for your help,
Zebananos

Low cost multimeters like the one I use that came from fleabay. I found it to be reasonable given its rather affordable price. The recent models are decent and seem to be accurate enough for laptop regulator tests etc.

Many models out there so find one that has the functions you need.

_________________
USB 3.5" enclosures are $19 USD postage paid from China, I own several

you seem to beg to be banned....

_________________
Adatmentés - Data recovery

The pulses coming to coil 3 afterward turn into those strange 0.3 volts on the ROM chip. Unfortunately, this is also nothing new. I don't have a good understanding of how these converters work, but I do know that the pulses should be much higher in amplitude, since this is a step-down converter, not a boost converter. Fzabkar understands this much better than I do and can tell you what you should see on an oscilloscope. Unfortunately, I don't have one, so I can't help with measurements on a working board.

Since this problem is observed in the circuit where the via has rotted, I would simply resolder it using highly active soldering flux. The oxidation was previously copper plating, and it's possible that a feedback signal is connected from that location somewhere in the inner layer of the board. Or perhaps it's just a via for convenient testing of newly assembled boards at the factory, and it's not connected anywhere.

The only things you can check in this circuit are the capacitors and resistors nearby, which may be part of the voltage monitoring circuit. Other than these, the only other possible source of the fault could be the SMOOTH chip.

It's highly likely that the drive not spinning wasn't the initial cause of its malfunction; this model almost always fails due to damaged heads. The problem that it's no longer spinning is due to it being thrown away and left lying around somewhere. I've encountered a similar problem, but with a very old Western Digital hard drive. In 2024, one of its heads failed, and the drive was put in storage in a shed. This year, it was retrieved and connected to a computer again, but it no longer spinned, and the board showed signs of oxidation. The main damage was near the motor control chip, and the drive had been copied using a board from a similar one.

This problem becomes very difficult to diagnose because this SMOOTH chip isn't a simple voltage converter that starts working immediately after power is applied. It's controlled via a digital bus by the hard drive's microcontroller. Furthermore, there's no documentation for such chips; they're proprietary and, in our case, custom-made for Seagate. Therefore, it is difficult to say which voltages should appear, in what sequence, and under what conditions, without having a similar board and a large amount of professional diagnostic equipment.

Tutorial - Linear and Switchmode Regulators used in HDDs:
https://www.hddoracle.com/viewtopic.php?f=59&t=231

Coils 1,2,3 appear to be associated with buck converters (step-down topology). These provide Vcore and Vio for the MCU, SDRAM and ROM.

Coil 4 appears to be part of an inverting buck-boost topology. This provides -5V for the preamp.

I'm thinking that the PMIC section in the motor controller needs to provide Vcore and Vio supplies for the other ICs. After the MCU unpacks the ROM into SDRAM, it then commands the motor controller to switch on the negative supply.

_________________
A backup a day keeps DR away.