Another big thanks to Robert for this one! I’ve briefly covered ZIP drives to some degree (really only covering the USB bus-powered drive) in an earlier posting, but today I was given a SCSI drive. This drive had a bit of a rattle when I got it – and by the time I got home, the top window had broken off already.
I did plug it in, only to find that it seemed to be faulty. It would click endlessly like the infamous click of death, but the spindle motor speed seemed to be somewhat off. This warranted some investigating.
I do have a parallel port version of the ZIP 100 drive that looks exactly like this one. In fact, the differences are so subtle, I will not bother to photograph the parallel port drive. Having previously taken a peek inside that one (before this blog was even established), I knew my way into the drive.
One thing to note is the bottom label which varies. This one has the SCSI connection diagram – the parallel port version has a different label with the serial number also on the bottom. Here you can see that the drive has two DB-25 connectors, so the SCSI bus can pass through the unit. There are switches for termination on/off and also for setting SCSI ID to 5 or 6 only. A rather limited selection, but most other devices can have the full 0-7 set, so that shouldn’t be a big issue unless you wanted more than two of these on a single SCSI bus.
There’s also the famous emergency eject hole which you can use to release your disk without power – as these drives had an electronic ejection mechanism unlike most floppy disks. The parallel port version also features two DB-25 style connectors, just without the slide switches and with different symbols in the plastic.
Interestingly, the serial and the drive details are stuck on the side of the drive, where the power connector is. As this drive didn’t come with a power supply, I used the one from the parallel port drive, a linear transformer based supply rated at 5v 1000mA.
Cracking it Open
Unfortunately, these drives aren’t built with screws. They’re instead built with internal locking tabs. These tabs are located two on the left side along the left seam, and one on the right side, with a special cut-out pattern for the rear I/O plate which slots into the case.
It is pretty much impossible to break open the device, even apply pressure at the seam, without breaking one or more of these darned tabs. So if you want to keep everything in perfect pristine condition, do not attempt to open it! But if you want to try to get it working … a sacrifice is necessary.
Getting on with it, the drive itself with the top cover off looks like this:
Already you can see that the PCB sits underneath with a large capacitor for power filtering. The drive mechanism is quite special as it uses guide rails which anchor in the top of the case and the bottom of the case, sitting in specially made holes for the plus-shaped posts to go into. Without the top case, the mechanism wobbles from some to side a bit. There are some special posts in the top case which interact with the sled on rails (black plastic parts, responsible for cartridge handling) to unlock the head return spring and keep the sled just at the right loading position.
Removing the front panel, we can see the sled mechanism with the spindle motor, the feeler arm which opens the cartridge door and the heads safely ensconced in the protective area behind along with its voice coil motor system.
The two springs near the front help return the mechanism back to its resting place when the disk is ejected. The whole platen with the spindle motor actually rises up to engage with the disk when one is inserted which is rather interesting.
In the rear, we can see when the mechanism is loaded, a feeler arm sits right where the emergency eject hole ends – depressing this lever/arm causes the mechanism latch to be freed resulting in the springs at the front returning the drive back to the ejected state.
There is also a spring in the head area which helps return the heads back into the safe area in case of ejection.
When looking at the main PCB, it is evident that the drive was made in 30th November 1995 in Singapore. It has firmware D.06, and it has an adaptec SCSI interface (no surprises here!). There is an iomega ASIC with the codename Phaedrus and some Hynix flash/RAM in the corner. And there are also a few capacitors connected with greenwiring techniques – seems like that was the era for these kinds of “repairs”. On the board is also what appears to be an IR LED and phototransistor combination unit for the retroreflector in the disc which signals the capacity – the drive immediately ejects any disk inserted if it doesn’t sense that optical signal.
On the underside of the sled, we see the use of some aluminium shielding for the lower head, and the spindle motor (several designs spotted), this one being a Nidec (as seen on the orange flex right near the motor). Interestingly, the platen on the spindle motor is magnetic and grips onto the disk that way thus the hub of the ZIP disk doesn’t need a hole or slot like the 3.5″ floppies do.
Now time for a look at the heads –
You can see that the drive has accumulated some dust over the years – and the head is safely ensconced in its special safe spot. The head rests on some fibrous material – to keep the two heads separate and to keep them from damage. It is textured, and it seems to also perform the function of keeping the heads clean (although it is of limited effectiveness). One thing to note is that the tracks run perpendicular to the head as pictured here – we are actually looking at the side of the head, despite this being the side that faces the cartridge and outside.
Here’s a better shot showing it. Notice how the coil and head gap are at the rear and the direction of the head gal goes across the image – that’s the definition of the track right there. For those who have read my last post about the SyQuest drives, you can see the head technology is on a similar primitive level that the head coil is visible and physically big.
Getting an Even Closer Look
So, my mission was to fix this thing. And lets face it, I’ve already gotten way past the point of getting out the superglue and sticking that top window back together, so why not go a little further. I had a few ZIP disks that wouldn’t pass long format anymore – and a desire to get an even closer look at the operation of the drive. Given the unhealthy status of the drive initially, I figured I could get a little closer by cracking open a cartridge and getting the drive to accept it and watch it work.
But to do that, one must start defeating the design itself.
You’d think a photo like that would be simple to make – but you’d be mistaken. The drive has been very carefully designed to retract heads in case of power removal, and or ejection. I had to defeat the head-retract spring and carefully hold the head ribbon to prevent it from retracting back into its cave (ribbon instead of head as pressure on the head may cause it to be damaged or misaligned). Then the was the problem of getting the cartridge inserted and detected – if you insert it all the way, the drive complains of an error and does not start the spindle motor. This is because the top of the case contains cut-outs that hold the disk in the perfect position, which is not fully-fully inserted.
This took a while to realize and left me pondering and scratching my head thinking just how would I get to my intended destination. But I did it in the end.
The glorious view of the head, suspension, arm, wire, coil on top of the mirror-shiny dark-coloured flexible disk media. The bar there appears to be a locking bar possibly used to lock the heads into position, or maybe it’s a weight to keep pressure on the heads. I’m not entirely sure.
And here it is from a slightly different angle.
What was more interesting was that I managed to get a video of the drive working – the disk spinning, the head loading – and something I never realized ZIP drives did – they sweep their heads across the media continually when idling, likely to prevent excessive media wear due to contact, and possibly to clean and keep dust off all areas of the disk. Now that’s interesting!
Unfortunately, in the journey, I did break all of the clips, so the drive is now held together with electrical tape. But now the drive spins, reads and detects my good cartridges! A step forward indeed! Unfortunately, when asked to long format a good cartridge – it does complete but it turns my good cartridge with media status of good 99% to marginal 78%.
Ultimately, it seems, this drive may have had a hard life in its past and its alignment could be rather marginal. Or maybe all my drives have one particular alignment bias that this drive hasn’t – and all of them are bad. Who knows!? That’s the mystery of ZIP disks! Some drives slowly and subtly damage disks …