| ??? 07/26/05 13:30 Read: times Msg Score: +1 +1 Good Answer/Helpful |
#98161 - Not so sure Responding to: ???'s previous message |
Jan Waclawek said:
You stated it yourself - the difference lies in the complexity, not the method. However, software protocol allows (and usually also requires) much higher complexity than the "simple shift register", so the probability of error seems to be higher. Even if the hardware is just a simple shift register, there still has to be logic regarding what to do with an ISP command once it is shifted in. I'm not convinced that a UART approach has a higher probability of error than a shift register--and a UART has the added advantage that we shouldn't have these problems with strange cables that sometimes have restrictions such as "don't exceed 1 foot cable length", undocumented or incorrectly documented timing in datasheets, and guessing the size of the resistor we should place in-line in the cable. In short, it seems that a UART-based ISP has to be debugged and verified once by the manufacturer before they ship their part whereas the non-UART SPI approaches need to be debugged and verified by each person that attempts to build a cable. I'd rather it be debugged by the manufacturer and then I can just ISP the device with the RS-232 cable that I already have and which I know works. Regards, Craig Steiner |
