Unidentified Emulex card from 1991

On Sat, Mar 17, 2012 at 10:17 PM, Mouse <mouse at rodents-montreal.org> wrote: Yes, and on most/all 68K Macs. It has no FIFO. It came from an era before serial chips routinely had FIFOs, but I've used it at 56K and faster with no problems - your CPU just has to keep up, which I know was a problem in the DOS era with 8250 and 16450 serial chips. That is true. It _is_ possible to stick it on a 68000 bus, but some glue logic, including, possibly wait-states or careful driver crafting may be required. It is full of registers. You don't just fiddle one thing one time, then slam data in and out. It's a very complex dual-port USART that will do Async, Bisync, SNA, and more. We used them on the later models of COMBOARDs, though to a fraction of their capability. Older models had a COM5025 sync serial chip (because we modeled that section after some DEC sync serial cards), so when we switched to the Z8530, we only ever used the secondary port for async debugging. One thing that complicated _our_ design was that we started using the Z8530 very early in its evolution, so we were hanging either a 4MHz or 6MHz part off of an 8MHz 68000. We had to inject wait states (via registered PALs and /DTACK, IIRC) _and_ we had to ensure we didn't jam up the inner workings of the Z8530 by hitting the registers too fast. We had NOPs in the driver between selecting and using the registers. Perhaps 8MHz parts would be easier to use now. Once we worked out what was needed, it was a solid, reliable chip, but I will admit that our embedded application guys did have a few fits trying to port front-end code from the older models that expected a minimal-option simple chip like the COM5025. -ethan