SCSI to IDE

On 1 Dec 2010 at 19:34, Tony Duell wrote: No, but there's at least one in a PLCC package, sockets for which have contacts on 0.100" centers, so in my mind, a PLCC is about the same as a DIL socket regarding ease of use. If I'm really dead-set on a through-hole package (perhaps I want to wire-wrap the design), SMT adapters are inexpensive are easy to use. What's more of a consideration for me is the shift to 3.3V logic and using mixed signal levels. While many uCs have 5V tolerant inputs, many don't and adding either a resistor-clamp-diode network to each 5V output, or resorting to a made-for the-purpose IC to shift levels is a bit of a bother. It's actually surprising that a number of manufacturers have modern MCUs in DIL configurations. Even the TI MSP430 has at least one family member in a 14 pin DIL package. If you're making things for your own use, fine--use whatever works for you. But don't expect others to use galena-and-catwhisker diodes or ammonium sulphate and aluminum wet rectifiers. Consider a project to convert a PC AT-type keyboard to a PC XT one. I could do this with a Z80, an SIO chip, some ROM and some RAM--and the "glue" to interface it all, or I could do it with an 8 pin PIC microcontroller that costs less than a dollar. The choice to me is a no-brainer--my design debugging time will be substantially less. As far as learning new instruction sets, that's a big shrug. After your first dozen or two, they all pretty much look the same. I've yet to be really surprised by any new one--hardly any learning at all- -and most new uCs can be programmed in C at any rate. Show me a new dataflow uC with 19 bit words that uses a Chen-Ho encoding and I'll be surprised. One big benefit of modern uCs is the number and variety of I/O pins. You simply can't find a 144-pin DIL package. --Chuck