robertb@june.cs.washington.edu (Robert Bedichek) (07/18/90)
We are designing a NuBUS card here at UW to write 3-D splines into a stereo Tektronix frame buffer. We have NuBUS wire-wrap proto boards and we have the Texas Instruments NuBUS interface chip set. We were wondering if any enterprise company out there has done the obvious and made a proto board with the TI chips already on the board and wired up to the edge connector. Has anyone used the TI chips in bus master mode? Did you have any problems with this? Thanks for whatever help you can lend, Rob
teener@apple.com (Michael Teener) (07/19/90)
In article <12573@june.cs.washington.edu> robertb@june.cs.washington.edu (Robert Bedichek) writes: > Has anyone used the TI chips in bus master mode? Did you have any problems > with this? We have used the 2420 and 2440 in prototype boards and had good results. The biggest difficulty has been the complexity of the 2440 (controller) interface ... I think they made it *too* general-purpose. The 2420 (transceiver), on the other hand, is an excellent part. I just wish it had built-in FIFO's for burst transfer support. (If you want this feature, and a simplified controller, call your local TI person ... they are evaluating new NuBus designs now.) TI also has implemented a two-chip MCP (Macintosh Coprocessor Platform) interface, which provides all the logic to support a 68k-based bus master on NuBus. Using these chips, a 68k system with basic ROM and RAM and NuBus master/slave interface only takes up 1/6th of a card. Very slick. It directly supports our AROSE real-time kernal. ---- Michael Teener -- 408-974-3521 ---------------------------------+ ---- Internet teener@apple.com, AppleLink TEENER | ---- Apple may know my opinions, but *I* am responsible for them | ---------------------------------------------------------------------+ Transportation by Cheetah N9900U, a loyal beast for the past 7 years.