USB-powered Zip Drive does not power on --which part to replace?

[andy3055]

Do the easy things first. As rsjsouza suggested, find a split USB cable and plug in to two ports and see if it will be recognized. If there is not enough current, it will not even start up.

[IO Sam]

If I understand correctly from what rsjsouza wrote, USB 1.1 could potentially supply as much as 500 mA. But a) I'm not sure that would be constant (and thus not sure how many Zip Drives could be attached to a single USB 1.1 circuit) and b) I suspect a Zip drive could have spikes in power draw here and there, but I doubt it would go as high as 500 mA.

But then again, don't quote me on any of that...

I am going to quote you on that.

Clearly you have no clue yet pretend to have one.

Oh, I absolutely have no clue and I have no pretention to have one! And that is precisely why I'm here asking so many questions to people who probably have a clue... If I had a clue, I wouldn't be here asking questions now, would I?

If you read the thread, you'll see the person who said USB 1.1 goes up to 500 mA (the same as USB 2.0) was rsjsouza, not me (I didn't even know it could go higher than 1 unit load). So, my statement above was my assumption based on the new information provided by the other commenter and not me pretending to know anything (which was made very clear by me asking you to not quote me on that).

But then again, powering 5 Zip Drives from a single USB 1.1 port was never my question/statement here in this topic... It was yours.

[IO Sam]

Do the easy things first. As rsjsouza suggested, find a split USB cable and plug in to two ports and see if it will be recognized. If there is not enough current, it will not even start up.

It definitely doesn't hurt to try... I already have a split USB cable, but it is USB-A (the drive is USB-B), so I'll try to get an adapter to test this.

[Monkeh]

All good points! I do have some old school external 2.5 hard drives that came with these split USB cables, but that is definitely not the case with the USB powered Zip Drives. But it is important to know that USB 1.1 can go up to 5 unit loads.

Hmm, are you seriously suggesting that USB1.1 can power 5 Zip drives?

No, he's parroting the language used prior. USB power is allocated in 'unit loads' of 100mA increments, and negotiated power can be up to 5 unit loads, or 500mA.

Clearly you have no clue yet pretend to have one.

Perhaps you're describing yourself here. Or perhaps you should just remain out of threads you've no intention of usefully participating in.

Apologies for the side show.

[fzabkar]

Is there a fuse near the crystal that is being obscured by the USB connector?

The ISD200 chip (U4) is a USB-ATA bridge. This is the chip that communicates with the USB host.

https://www.rockbox.org/wiki/pub/Main/DataSheets/isd200.pdf

I would measure the voltages at the capacitors attached to the IC's VDD18 (core) and VDD33 (IO) pins.

There appears to be a serial 8-pin "ROM" at U12, probably a 24Cxx I2C EEPROM. If this EEPROM is present, then the device identifies itself to the USB host using the Vendor ID and Product ID of the OEM, in this case Iomega. However, if the EEPROM is absent, the bridge uses its own internal VID and PID, namely 0x5AB and 0x0030, respectively (page 22 of datasheet).

You can test whether the bridge is communicating with the USB host by invalidating the EEPROM while powering up the PCB. You could do this by shorting the EEPROM's data output pin to ground.

[IO Sam]

From what I could see around the USB connector housing (macro pictures below) I could not immediately identify a fuse (or a "000" cap), but then I don't have the required "trained eyes" that many of you guys here do for these things... Do you see anything that looks like a fuse in these images surrounding the USB connector housing?

As for your suggestions to measure the voltages at the capacitors attached to the IC's VDD18 and VDD33 I, I'll do that as soon as I get my new bench power supply that will allow me to test this board as suggested by a previous commenter.

As for shorting the EEPROM experiment, would I need to de-solder its data output pin from the board as well as bridge it to ground with a wire? Or could I just bridge it as it is? (Thank you!!!)

[IO Sam]

Alright, after macro-shooting the whole board I managed to find two "000" caps which could be the fuses here (although just one of them, S3-side-USB.jpg, is immediately close to the USB plug housing). I'll test both of these later and report back.

[fzabkar]

It should be OK to short the EEPROM's data pin to ground in-circuit. Data recovery guys do this all the time. I was originally reticent, but they convinced me it was OK. You only need to maintain the short until the PCB powers up, then you can release it. That's because the bridge IC only looks for the EEPROM immediately after power-on and then ignores it.

[IO Sam]

Nice! That sounds a lot easier! 

If bypassing the EEPROM should work, would that mean it's bad and should have its data pin permanently disconnected/grounded?

At any rate, I now have a full check list of things to try... Hopefully I'll find the culprit in one of these steps! Thanks again!!

[fzabkar]

If, after bypassing the EEPROM, the bridge IC appears in Device Manager, this will prove that the bridge IC is not brain dead, and it will verify the USB link. Then you will need to look elsewhere for a culprit.

I notice that there is a transistor at Q3 (near the USB connector) which is marked "PT". I'm betting that this is configured as a high side power switch. I'm guessing that it is a P-channel MOSFET with 4 drain pins, 1 source pin and 1 gate. If it is switched on, then the D and S pins should be sitting at +5V.

This device is probably equivalent to Q3:

http://www.aosmd.com/pdfs/datasheet/ao6405.pdf

I can't see what is under the sticker, but I expect that you will find a microcontroller with an ATA interface, a read/write channel IC ("B9"), a serial flash memory (AT27LV520), and a motor controller (U8).

[IO Sam]

Thanks again for all these tips! I removed the sticker and you can now see the markings on the ICs.

[Haenk]

I still would check the caps first.
If I remember correctly, these drives will do a medium check on powerup, i.e. starting the motor. USB power might not be enough for that (there is a reason, these split USB cables exist, USB powering was very close to the edge back then). So that cap might be needed for the spike in current draw.
Might be a stuck motor as well.

[IO Sam]

I'm getting a USB-A to USB-B adapter (to use with my existing split USB cable) this Monday. So I'll be able to plug the drive with the extra juice to test this hypothesis. And I am definitely testing all these caps with a multimeter this week as well.

As for a possible stuck motor, wouldn't the drive at least light up (and be seen by the OS) if that was the case?

[Whales]

+1 to the above

Plug in USB.  Measure voltages, see if the 5V is getting from the connector into the board.  Easy way is to keep the black probe of your multimeter clipped to the metal shell of a connector and then hunt with the red probe, starting from all of the pins on the USB connector.

Many USB devices will have a fuse OR a resistor acting as a fuse.  These might be blown.

I doubt the main caps will have shorted, I've never seen that failure mode (outside of "I exploded the whole board").  If the main caps have shorted then you'll be able to observe little to no voltage on the USB connector (the host USB device will/should limit the current if it's USB compliant).

So, in that case I could plug it in my laptop and then test it (I don't own a bench power supply yet, so if this is the case it would be convenient)?

Yes.  Your laptop should have polyfuses (or similar active protections) in the USB ports to prevent too much current exiting its ports.  Otherwise every faulty USB device would damage your laptop.

I don't know why you would bother with a bench power supply here.   If the device is drawing a reasonable amount of current then your laptop will supply it fine, if it's trying to draw too much then your laptop will refuse to supply it (eg voltage will drop OR it will get turned off, possible repeatedly cycled on/off/on/off).  A bench power supply could let you do the same (with more effort) or it could let you provide more current... but that's not a safe diagnostic method

As for the caps, I believe these are already polymer solid ones (this unit was made in 2001). So, as you said, I would also expect them to go out in quite a loud fashion instead of dying quietly of old age, right? [EDIT: after some research, I found out that these can indeed degrade and fail with time without any physical signs of damage]

Again I think this is unlikely.  There are hundreds of unlikely things I could list, I'm not sure where the focus on electrolytic caps came from.

Measure more, guess less.

If it was indeed a blown resistor, would I need the schematics of this PCB to be able to identify it (if I wanted to avoid testing them all one by one XD)? Or should I look for a "000" resistor (which I could assume to be the fuse)?

No need for schematics.  It will be wired very close to the USB port and you can use your multimeter on continuity buzz mode to find it (edit: whilst the unit is unplugged from power! otherwise the multimeter will be confused).   It will almost always be the first thing wired to the +5V pin of the port.

Yes 00, 000 or similar if it's a resistor.  If it's a fuse style package then it might be a different colour or have a letter written on it instead of a number.


Stop asking questions and go have a look   Plugging it into USB power and probing with a DMM is unlikely to cause any harm as long as you are careful to (1) keep the multimeter on volts (not amps) mode and (2) don't short nearby pins/pads out.

Once you have more information we may be able to help you more.  At the moment all we can do is guess.  Power supplies are always a good place to start looking.

[IO Sam]

Thank you for your commentary and tips here! As my collection of vintage hardware grows, I see the need to learn more about how to repair these things myself and that is why I'm considering acquiring some additional gear such as a benchtop power supply, hot air station, oscilloscope, etc. But it is great to know (from your experience) that my existing laptop can be used safely for this particular project.

My main concern when repairing old gear is safety, since some of these things are getting harder to find (and more and more expensive). So, one might understand my excessive care and precaution here.

As for questions, I asked enough of them already ... I have enough here to work with for days now! And, as a newbie in this forum, I have to say I'm blown away by the amount of helpful and useful tips and suggestions I got in this topic! I'm stoked to have found this place and it is pretty incredible to see the nice community that Dave managed to build here!

Thanks to everyone who tipped in, and I'll post the results of this little project here as soon as I have some updates.

[fzabkar]

Those capacitors of that era were prone to leakage, and this would often corrode the PCB tracks. That said, these failures would usually be associated with a fishy smell.

Here is the datasheet for the AT27LV520:

https://pdf.dzsc.com/V52/AT27LV520.pdf

If there is no Vcc, I would check whether Q3 is being turned on.

I suspect that the bridge IC controls power to all the other chips via Q3. If there is no activity on the USB link, I suspect that the bridge turns off Q3, thereby switching the drive into power saving mode. If Q3 is off, then this would point to the bridge as a suspect. But that's all just conjecture.

[IO Sam]

Those capacitors of that era were prone to leakage, and this would often corrode the PCB tracks. That said, these failures would usually be associated with a fishy smell.

I'm glad to report that there were no caps leakage here!

[tooki]

I'm getting a USB-A to USB-B adapter (to use with my existing split USB cable) this Monday. So I'll be able to plug the drive with the extra juice to test this hypothesis. And I am definitely testing all these caps with a multimeter this week as well.

As for a possible stuck motor, wouldn't the drive at least light up (and be seen by the OS) if that was the case?

I wouldn’t. Adapters like that aren't even supposed to exist.

Especially given the fact that we already know it doesn’t need two ports’ worth of current: it shipped with a standard single-headed cable. See the manual attached.

Some weird USB adapter simply runs the risk of frying something. Don’t do it.

[james_s]

If you're getting 5V to the circuit then my best guess is that the firmware has become corrupted, 20 years is about when eproms and flash, at least older flash, start to suffer problems from bit rot. This is especially true if it was exposed to elevated temperatures.

[IO Sam]

I'm getting a USB-A to USB-B adapter (to use with my existing split USB cable) this Monday. So I'll be able to plug the drive with the extra juice to test this hypothesis. And I am definitely testing all these caps with a multimeter this week as well.

As for a possible stuck motor, wouldn't the drive at least light up (and be seen by the OS) if that was the case?

I wouldn’t. Adapters like that aren't even supposed to exist.

Especially given the fact that we already know it doesn’t need two ports’ worth of current: it shipped with a standard single-headed cable. See the manual attached.

Some weird USB adapter simply runs the risk of frying something. Don’t do it.

You are correct that the drive came with a simple (non-split) USB cable in the box. I wonder if the other commenter was suggesting the use of a split cable just for testing purposes (I have no idea if the extra power from a dual lead cable could somehow help to recharge these old capacitors or something of the sort).

[Monkeh]

Take a look at Y1 from a different angle.

[IO Sam]

If you're getting 5V to the circuit then my best guess is that the firmware has become corrupted, 20 years is about when eproms and flash, at least older flash, start to suffer problems from bit rot. This is especially true if it was exposed to elevated temperatures.

As I test the PCB with my multimeter this coming week, this is something I intend to find out (even bypassing the EEPROM as another commenter suggested). I have Zip Drives in use for decades now without a hitch, but the fact that this one was never used for over 20 years indeed puts this unit in a different set of variables as far as firmware bit rot is concerned.

[IO Sam]

Take a look at Y1 from a different angle.

Y1 macros galore below:

[MikeK]

Two pins on the right side of U4 appear to be solder-bridged.

[ Specified attachment is not available ]

[Monkeh]

I saw those, they're fine - one is Vss and one is TEST3 which needs to be tied low. That bridge appears to be in the classic position for a trace linking two pads with no mask. I'd probably try removing it, but I'd also remove the chip to see where those pads are supposed to be going first - probably a bit advanced for this case..