With my earlier mysterious loss of photos, today, I’ve gotten around to re-taking them. I’ve already covered the whole set once before with my camera phone, but I wanted to do it properly. I recently re-covered the cartridge and its internals, now it’s time for the drive. It has worked hard, and has recovered all four SyQuest 88Mb cartridges we could find, identically twice over.
As you can see, it’s not much excitement when it’s sitting inside its case. Yesterday, I had to take it apart because an internal plastic loading feeler had snapped, causing the drive to be inoperative. I since super-glued that back together, and it’s operational once more. One thing we discovered upon rotating the case around was the label with a manufacture date.
Popping off the lid, we can see this rather well-built external drive enclosure has a caged power supply, and a simple SCSI ribbon connection for data and for ID (after all, internal SCSI was the same as external SCSI with different connectors). The manufacturer lovingly magic-taped the three SCSI termination resistor packs to the top of the drive so that in case of removal and use internally, they could be refitted to enable termination. How thoughtful! It looks very plain at the top, but has a yellow-golden coloured metal lid, rather than the more “regular” grey of plated steel. On it is a circular “warranty void” sticker – that’s already been broken from my last repair. You can also see a position on the drive for a grounding lug attachment, in case the chassis didn’t provide good grounding.
A view from the exposed side gives some cryptic codes, probably related to manufacturing.
Removing the drive from the enclosure for a better view – the rear shows the “older” wisdom of mounting the molex connector “upside down” – people had to be careful to make sure it’s the right way up because they could be inadvertently forced in the wrong way if enough force is used. Modern drives tended to have the chamfers on the edges at the top rather than the bottom. The power requirements are fair for 12v, and modest for 5v. SCSI ID was set with the standard A, B, C bit jumpers in binary.
If we rotate the drive around, we can see the logo sticker and the serial number. This sticker claims it was Made in Singapore.
The underside of the drive has the PCB which runs the full length. The PCB is protected by a thin protective plastic sheet – a flimsy way to protect the PCB, but better than nothing. As it is somewhat inset from the mounting points, it is probably unnecessary, but just to prevent massive damage from shorting out if someone stacks the drives on top of another.
Visible is a cut-out for the spindle motor rotor which is branded Nidec – a respected spindle motor manufacturer for many hard disks over time! Removing the plastic cover gives us access to a shiny green solder resist multi-layer PCB. The stickers on it suggest that the PCB was assembled in Hong Kong and is loaded with the CFA revision ROM. It was a SyQuest revision A PCB from their own fab, or so it seemed.
We can go one step further and remove the board from the chassis by undoing two more screws. We have to be careful as there is a two-wire connector on the right side of the board which has to be removed before flipping the board over. The flex on the left side of the board needs to be treated with care as it may be a bit brittle.
And voila. That’s the brains behind the SyQuest drive. Visible are several notable things:
- Socketed Flash chip with the CFA sticker (which implies revision CFA ROM is loaded on it, unless updated in the field which is unlikely).
- Adaptec Inc. AIC-6110HQ – likely to be a SCSI driver/interface chip.
- Hyundai HY62256ALJ-70 chips which are RAM probably used for a cache buffer.
- Siemens SAB 80C535-16-N chip which is copyright Intel – it’s a 8051 CPU that runs the drive.
- Syquest branded STmicroelectronics made chip – likely to be an ASIC.
- A large Hitachi package at the bottom edge – HA13441 three-phase brushless DC motor driver which uses the frame of the drive as a heatsink.
- Kyocera 20Mhz crystal oscillator.
- And a smattering of other assorted logic chips.
On this side of the board, it says it’s Made in Taiwan, copyright 1991. Definitely a multi-national drive. There is a set of 14 gold pin headers which are used to communicate to the flex inside the head-area of the drive.
There is a yellow wire which connects from the CPU to an IC, and another two from an IC to a precariously placed resistor. The technique seems to be used to correct errors in PCBs or commit required modifications.
Also interesting is along the left edge, there are two green glass pods which are magnetic reed switches similar to the ones used in surveillance systems for door open monitoring. These are used as “sealed”, high reliability switches – which couple with these plastic sensing levers with magnets that contactlessly signals the status of those feelers.
The two plastic feelers can be seen, along with the gold pin header interface into the head area and the spindle motor in the corner.
Turning it over, lets take a look at the head area itself …
The cartridge is loaded from the left side. There are a lot of mechanics, which are probably best shown in a video (see later). The right side is the head, in its parked position – the bottom right area is the voice coil motor, the top right is responsible for head parking and locking, as well as head signal amplification (by virtue of an IC soldered to the ribbon).
Here you can see the head arm with flex, leading to the back where the head pre-amp is situated. The flex continues across the solenoid, all the way to the golden pin headers we saw earlier. The large yellow solenoid is responsible for the locking of the heads, whereas the clear cylindrical chamber is responsible for creating smooth back-pressure to gently load the heads onto the disk. There is a black plastic pair of arms which keep the heads apart while the heads are parked to prevent head damage.
Here is the head voice coil motor. It’s a hefty thing. A large enameled copper wire rectangular coil straddles a large magnet assembly to provide the torque to move the heads across the disk.
This drive has two heads, one for each surface of the single-plattered cartridges. The heads themselves are very primitive. The head consists of a large aerodynamic slider surface, of which only the tiny part is the head. You can see the coil wrapped around the corner of the slider? That’s the head – just the corner part. The whole slider is exceptionally dirty from use and age – this may be the reason for poor drive performance, as the aerodynamic effects may be disturbed by dirt causing flying height/azimuth issues.
Here’s another one focusing on the heads – the coils of wire. Wow. Looks like that one coil is responsible for read, write and erase possibly with several separate turns – but I can’t be sure.
This is what caused me to open the drive in the first place. This secluded feeler arm with the tang at the end is responsible for hooking onto the rotating sliding door of the cartridge and levering it open as the cartridge is being pushed into the drive. The tang snapped off the end of the arm – being only made of plastic. Luckily my application of some cyanoacrylate (aka super-glue) did the trick!
Here’s a video I made of the mechanism in operation –
Pretty neat! In the video you can see just how the solenoid unlocks the head, and it’s carefully loaded on against the force of that cylinder-thing. The system then positions the heads and calibrates it for the disk’s servo information before sitting idle and ready. Then I hit the eject button, the drive spindles down, and the unload lever pushes the heads back onto their “arm” where they’re held parked ready for the next disk. The unmistakable whine of the old drives … ahh!