So I have two DSD440s and they both needed work. One I got from David G. in trade for my TU10 that he restored. That one worked when he gave it to me. The other was bought off ebay. It was a disaster when it came (but that rehabilitation project is for another time. Needless to say, it was a labor of love, but it works now.)
In general, they are very nice machines. As Jack mentioned, they can format RX01 disks without needing to have a computer attached. They are also only half as tall as an RX01 in your rack. They use the venerable SA800 for the disk drives. They have a very useful set of internal troubleshooting diagnostics. The controller board is covered with jumpers, not for configuring the board, but for disabling sections or applying test signals for doing repairs, which are also very helpful.
The biggest problem though is how the controller is mounted inside the enclosure and what it does to the power busbar across the bottom of the board.
http://www.vintage-computer.com/vcforum/album.php?albumid=63&attachmentid=6798
See that the board is as wide as the enclosure, but is only held in place at the very outer edges. There is a stiff power busbar along the bottom edge that carries ground and +5 to the columns of chips. Small leads come from the busbars to go to the column traces on the PCB. As the PCB flexes from being pushed on by the tight fitting cables, the small leads crack and columns of chips loose power.
http://www.vintage-computer.com/vcforum/album.php?albumid=63&attachmentid=6797 Of course the middle column looses power first because this is the highest stress area.
So, people start jumpering the power over from adjacent columns that still have power. Both boards in both of my units show this "rework". One of my units got so bad though, that I just removed the busbars completely and replaced them with wire. The wire must be heavy enough, because there is negligible voltage drop at the end of the wire bus.
http://www.vintage-computer.com/vcforum/album.php?albumid=63&attachmentid=6802 (ugly wiring from adjacent columns still left in place, perhaps I should have removed it.)
I have had other random TTL IC failures in these before, but this time I had an 82S123 fuse programmable rom fail on one of the units. One of the outputs would only go down to no man's land (~2V) when it should have been <0.8V on a low output. I dumped the rom from the other controller and burned an EPROM and made an adapter to replace the bad rom. Of course I have an el-cheapo Needhams PB-10 and so had to make an adapter to read the 82S123. While making the adapter, I also gave it sockets to read all the other kinds of fuse programmable roms in this unit. If anyone else has this problem, I can dump any rom for them. The only place to put the adapter board in the DSD440 was behind the controller, since the front is so close to the SA800s.
http://www.vintage-computer.com/vcforum/album.php?albumid=63&attachmentid=6800
A few words about SA800s.... There were two versions of the boards on them. The "older" one has discrete transistors to drive the three pole head positioning stepper motor. Big beefy transistors in TO-66 cases. The newer boards got Sprague jam-four-darlington-pairs-into-a-14-pin-dip pole drivers. Of course, it cooked. I replaced it with a much beefier, but similar driver:
http://www.vintage-computer.com/vcforum/album.php?albumid=63&attachmentid=6801
Lastly, a track zero photogate failed on the older drive (the phototransistor. ) Fortunately I had an exact fit replacement GE H21A2 in the junk box. It must have come from a Tandon TM-100 which I parted out about a dozen of twenty years ago.
This recent round of repairs kept me busy for a while. Last time I worked on them I did full head alignments with an AAD. They all still seem to be in good alignment.
I know these units all too well now.
Lou