fouts@lemming. (Marty Fouts) (10/16/88)
Err, Moto wanting a bit of bucks for the chip is only the tip of the iceberg. I do a small amount of 'garage design' but no implementation because: 1) You've got to use semicustom to be competitive now, and I can't affort the NRE (Non Recurring Engineering) charges, or to have a foundry blast enough chips to justify (to them) bothering. 2) Many-pin packages like current CPUs require multilayer boards, which require expensive board building gear or jobbing out to board building shops. Also, more expensive handling in terms of inserting/removing parts (Or surface mount!) etc. 3) All of the above means using sophisticated CADware to aid in logic design and board layout, which means a powerful workstation an expensive proprietary CAD software. 4) The frequencies these things run at lead to microwave frequency behavior on the boards. Microwave design is not for amateurs. 5) Moto needs to do business with volume buyers to make money, they can't afford to sell me 1-5 quantity of chips without that huge markup. It isn't really feasible to try the discrete component wirerap prototype first stage because you can't get all the discrete's on one board so you end up debugging different problems. I've talked this over with a number of engineers here in the valley, and we tend to agree that a lot of money is needed to design an entire system from the ground up. What's really needed here is something like a chip-design coop along the lines of auto repair coops where the coop eats the high overhead costs and shares the resources. (Hey, maybe I just thought of something marketable?) Anyway, like physics has become big time, computer building has become big time, because its too expensive to do in the backyard. The days of hocking the Microbus to build a micro bus are over. Marty -- +-+-+-+ I don't know who I am, why should you? +-+-+-+ | fouts@lemming.nas.nasa.gov | | ...!ames!orville!fouts | | Never attribute to malice what can be | +-+-+-+ explained by incompetence. +-+-+-+