On 3/19/19 8:26 AM, Jon Elson via cctalk wrote: > On 03/18/2019 10:35 PM, W2HX via cctalk wrote: >> 4. Anyone ever make a pertec to USB project? Might be interesting >> to get access to old tapes on more modern computers? >> >> > I have some CDC (Laser Magnetic Storage) 92185 drives (known as > Keystone, this is the 1600/6250 version, buffered Pertec Interface). I > made a horrible hack of an FPGA board I make to read tapes through a > PC's parallel port. it works, but is real slow.
I discovered after doing an interface with some CPLDs that you really need buffer memory to keep things going. My ISA-interface Computer Logics 16-bit card with 256K of memory does a pretty good job with my Fuji. I generally don't like SCSI interface tapes because they must buffer ahead to keep the tape going, so when an error occurs, you have only a general idea of exactly where it is. With a Pertec interface, you can usually pin things down (controller permitting) a bit better. I've got an on-again-off-again project using an STM32F29 MCU running at 180MHz with 256K of SRAM on-chip and 8MB of RAM off-chip to do Pertec interface work. My point is not to interface to any particular computer, but to provide for transfer between SD card flash and the tape drive. Data and control will be USB, but only after a complete tape has been read. I do a similar thing with an HP7970 and an STM32F407 MCU for 7 and 9 track NRZI tapes. Several years ago, I made a simple Pertec interface ISA card using 3 8255 PPI chips. While it worked, it made the need for buffering very apparent. --- On the subject of tapes, I was reminded about how "more is less" in terms of software. The job was to transfer some tar files from DDS tapes. With Linux running on a quad core 3 GHz 64 bit setup, I'd get tape "scrubbing" periodically (the drive was a Sony SDT5000) using dd. On a 450 Mhz K6 system running MS-DOS, the tape streams without interruption using a simple ASPI-interface program. The controller in both cases is an Adaptec PCI UW SCSI controller. --Chuck
