Thanks for that. It seems to make sense against the PCB I'm investigating... I never thought to look in Amstrad service manuals (and I have quite a number of them). They sed the SED9420 quite a lot (in the CPC, PCW, etc), but never with an external VCO. Are you sure the SED9421 is not a complete data seperator? The PCB I am looking at (from an Epson PF10 drive unit) doesn't seem to have much else around the 765 disk controller chip. -tony

Looking at am Amstrad manual, it would appear that the pins on the 9421 aee a subset of those on the 9420 (and I don't think that's because you link them together). And looking at the PF10 board, it must be the complete data sepearator, there simply isn't enough logic on the board otherwise. The PF10 is, of course, a 3.5" 40 cylinder drive for the PX8. Inside there's one PCB on top of the chassis half of which appears to be a ridiculously complicated PSU, the rest of which consists of : A 63A03X microcnotroller + 2764 EPROM + 6116 RAM A 765 disk controller + 9421 data seperator Some RS232 buffers (there are 2 8 pin mini-DIN serial ports on the back, one to link to the PX8, the other to link to another PF10. From what I can determine, they work at RS232 levels) An HA16642 read/write circuit A couple of motor driver chips to drive the head stepper Some glue logic. A few gates, a decoder (which seems to be an address decoder). No obvious sequential stuff. There is no seperate logic board on the drive mechanism. The heads, track 0 sensor and stepper plug into this logic board. There's a 6 pin plug that fits in to a socket on the spindle motor PCB under the chassis (the signals are helpfully labelled on that PCB, and are ground, +5V, Motor On, Index, Write Protect, and Disk Inserted). Nothing to speak of. I've certainly never had real problems that were traced to, say, a 9216 design. But then I've even had few problems with that infernal built-in sepeator of the 1771... -tony