On Tue, Sep 21, 2021 at 12:15:12PM -0400, Paul Koning wrote: > > On Sep 21, 2021, at 12:00 PM, Jonathan Stone via cctalk > > <cctalk@classiccmp.org> wrote: > > > > The RD53s in both Microvax 2000s are dead. I'm reluctant to buy more 30+ > > year old rotating media. > > If I'm going to buy a disk emulator, I'd much prefer SCSI to MFM, for the > > obvious performance reasons. > > Emulators presumably act, performance wise, like SSDs. Given that, an MFM > emulator will significantly outperform a SCSI hard drive since it has no seek > delays. > > David Gesswein's MFM emulator works impressively well and is inexpensive, too. >
The MFM emulator can't transfer data faster than the 5 Mbit/second, 625k Byte/sec of the ST-506 standard. Seek time isn't zero due to overhead in moving the data in the Beaglebone to where the PRU processors can access it. Think its around 0.6ms/head. An entire cylinder is transfered on each seek. Additional time if cylinder was written to. Each track written needs to be moved also. Most ST-506 drives are slower. I'm faster than average seek for RD54 but it wins for single track seek at 4ms with 15 heads.
