The Commodore 1581 drive was the C64 3.5 floppy released in 1987 to extend the usable life of its 8-bit line of computers. It was an 800K floppy disk drive using the then-new 3.5-inch disk format. It stored almost as much data as five 5.25-inch disks formatted by a Commodore 1541, and was the closest thing to affordable mass storage available for the C-64 at the time.
The Commodore 1581 drive provided more storage in a smaller and faster package than the much more common 1541 drive. Unlike the early model 1541, Commodore provided DIP switches to change the unit’s device number. This was important as nobody expected the 3.5-inch format to take over as the C-64’s primary storage medium. It was going to function as a secondary mass-storage device.
Born from Amiga
The 1581 drive was a departure from earlier Commodore disk drives for its 8-bit computers. It used the same 3.5-inch double sided, double density drive mechanism as the Amiga series of computers and it inherited the Amiga’s MFM disk format rather than using the GCR format. The GCR format used in Commodore’s other drives was more efficient, but also more costly. The Amiga floppy drive was very similar to DSDD floppy drives for IBM-compatible PCs, and the same modifications that allow certain PC floppy drives to work in Amigas also work in the 1581.
Internally, the 1581 contains a 6502 CPU with its own RAM and ROM, a WD1770 or WD1772 chip to communicate with the 3.5″ floppy mechanism and an 8520 I/O chip to provide an IEC compatible interface to communicate with a C-64 or other Commodore 8-bit computers.
Commodore announced the 1581 at the January 1987 Consumer Electronics Show in Las Vegas, at an introductory price of $399. But like many Commodore peripherals, it was heavily discounted and the price fell relatively quickly. By September of that year, a few mail order houses were advertising it for $199. And by November, some were advertising it for as little as $179, and claiming a suggested retail price of $249.
By comparison, in 1987 a 20-megabyte hard drive for a Commodore 64 cost $900.
The 1581 disk format consists of 80 logical tracks and ten 512 byte sectors per track, used as 20 logical sectors of 256 bytes each. The disk format isn’t compatible with other MFM formats like the IBM PC, Amiga and Atari ST. But with software, it’s possible for a PC to read a 1581 disk. Software was also available for the C-64 and 128 to transfer files to and from a PC 3.5″ floppy disk using a 1581. This made it possible, to some degree, for someone with a PC at the office to take work home and work from a Commodore 64.
The 1581 differed enough from the 1541 and 1571 that older Fast Load cartridges weren’t compatible with it. Later fast load products had to be revised specifically to work with the 1581.
Like other Commodore disk drives, the 1581 connected to the IEC bus the same way. The DIP switches to set its device number were the same as the later 1541-II. Unlike the Amiga, the 64 and 128 never got anything resembling Plug and Play.
Life with the 1581
Commodore claimed the 1581 was three times as fast as a 1541. The limitations of Commodore’s IEC bus slowed the drive down, but it was indeed slightly faster than a 1541, even when you used it with a stock C-64. The faster seek times helped. When you used it with a C-128, it was noticeably faster than a 1571.
It gave 3160 blocks free compared to 664 blocks free with the 1541. You couldn’t necessarily just dump five disks onto a single 1581 disk, however, because it only had about twice as many directory entries as a 1541. When I tried to dump multi-disk games onto the 1581, sometimes I ran out of directory entries. That was a shame. Large games like Pool of Radiance would have been a lot nicer to play on the 1581, with faster disk access and no disk swapping.
Some hobbyists have modified multi-disk games to work on 1581 disks, but a lot of that work happened in recent years, or at least only surfaced in recent years.
The Commodore 1581 drive and GEOS
The 1581 disk drive worked very nicely with GEOS, the graphical operating system for the 64 and 128. GEOS allowed you to copy multiple applications onto the drive and run them from it, eliminating disk swaps. And since GEOS was very disk-intensive, the faster disk access was nice.
The Commodore 1581 drive and BBSing
The 1581 was great for running a BBS because of its large capacity, but it had some limitations. Putting a message base on the 1581 sometimes resulted in files getting mixed. And putting a user database on it sometimes resulted in weird behavior. It was fine for storing file downloads, but I remember one local BBS operator being a very outspoken critic of the 1581 because of its limitations. I worked with the author of an obscure BBS program to try to overcome these limitations on a 1581. We weren’t very successful.
I think dividing the 1581’s physical tracks and sectors into smaller logical sectors may have had something to do with the odd behavior.
That said, a lot of BBS operators got by with a C-64, a 1670 modem, a 1541 and a 1581. That combination of hardware cost between $500 and $600 and provided a nearly megabyte of storage. It made for a modest BBS but big enough for people to take seriously. If the BBS caught on and the operator came into some money, they’d add another drive for more storage.
Although Commodore sold millions of 8-bit computers, the Commodore 1581 was a niche product. By the time a power user added a 1581, a RAM expansion unit, and a hard drive, the cost started to approach that of an IBM compatible PC with 512K or 640K of memory, both types of floppy drives, and a hard drive.
Commodore users tended to be really loyal, so some would start down that slippery slope. But not all. Many stuck with a single-drive setup with a 1541 until they were ready to upgrade to a PC or possibly an Amiga.
That’s why the 1581 is much rarer today than a 1541 or even 1571 floppy disk drive. Today, when one turns up on Ebay, it may very well sell for as much or more than it cost 30 years ago. While not as rare as a VIC-1540, it’s not common.
By using the same methodology the Allies used to figure out how many tanks the Germans could produce in WWII, Youtuber Robin Harbron estimated approximately 60,000 1581 drives were produced. Due to the likelihood that one of the series of serial numbers started at 10,000 rather than 0, and the margin of error of about 12 percent due to the sample size, the actual number produced could be as low as 46,000 or as high as 70,000. But based on the data he collected, we can be about 95% confident there were no more than 70,000 1581s produced.
As Commodore’s fortunes waned and C-64 sales slowed, Commodore found itself with more 1581 drives than it really needed. Some number of unsold 1581s sacrificed their drive mechanisms so they could end up in Amiga 500s. These incomplete 1581s were sold after Commodore went bankrupt, and hobbyists made them work by installing their own drive mechanisms.
Unfortunately it’s difficult to identify an Amiga 500 containing one of these mechanisms. The date code on the mechanism won’t match the rest of the machine, but you have no way of knowing if it’s because it’s a 1581 mechanism installed at the factory, or if someone swapped a mechanism to repair the machine at some point.