Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
11/03/06 17:28
Read: times


 
#127333 - let's clear up some misconceptions ...
Responding to: ???'s previous message
First of all I think you believe I disagree with you on everything in this discussion. That's just not the case. I surely agree that it's silly to attach a device that provides three external ports if you have to give up three ports in order to use it if it doesn't provide any additional value. Further, I doubt that the auto-handshaking modes will be of great interest to many users. However, if the 8255 had decent output current and were fast enough that you didn't have to change external bus speed in order to use it, and this includes the case wherein you don't have any other devices externally attached, it would provide significant benefit.

The fact that it's old doesn't matter. What matters is whether or not it works.

If you have to give up three ports in order to gain three ports, even if they are buffered, well, it's probably not worth the expense. However, if you need 12 additional ports, you still only give up those three ports to make it possible. That might warrant some thought. Let's say you're driving an LED matrix ... that's 96 segments that you can turn on or off, depending, of course, on how you do it.

I agree that there are combinations of circumstances and MCU's that would make the use of an 8255, even if enhanced with much greater output "punch," unwarranted. However, I believe that there are circumstances where it just might be warranted, even if you have to "fiddle" with the bus speed before and after each and every time you have to "talk" to them, which you wouldn't have to do if there were no external memory that required you write to it "fast". After all, if you're writing to external data memory in the course of your code, you probably won't have to extend more than 1 in 10 cycles. It might be "worth" the performance sacrifice. Besides, you then can use a cheaper memory.

Yes, there are tradeoffs. However, if the "new, enhanced" 8255 had 24 mA outputs and would operate at speeds at or exceeding those of typical SRAMs (not the fastest ones), it can, indeed, prove useful. Lots of code has been written for the thing. Lots of applications "live" on it now. It still does what it always did, and, with the added output drive and higher speed, it would easily work its way into new applications by offering function at a good price.

If the price weren't good, and/or the enhancements weren't there, I wouldn't use it.

RE




List of 61 messages in thread
TopicAuthorDate
Any value to a fast 8255?            01/01/70 00:00      
   Fast 8255            01/01/70 00:00      
      8255 breakeven volume            01/01/70 00:00      
         there are a few things ...            01/01/70 00:00      
            8255 vs 6824            01/01/70 00:00      
               too many users failed            01/01/70 00:00      
            Reconfiguration            01/01/70 00:00      
               Yes, one had to know how the device behaved.            01/01/70 00:00      
               the 8255 was tough to learn, but in its crummy way            01/01/70 00:00      
   peaple would rather forget the 8255            01/01/70 00:00      
      Why do you suppose that is, Jez?            01/01/70 00:00      
   raison d'etre            01/01/70 00:00      
      Main reason, I think            01/01/70 00:00      
         There are other reasons, too.            01/01/70 00:00      
            Thanks for sharing your experience!            01/01/70 00:00      
      There's also a large replacement part market.            01/01/70 00:00      
         but for that you have to compste with Intersil            01/01/70 00:00      
            Intersil still only offers the \'80\'s technology            01/01/70 00:00      
               May be I am wrong, but...            01/01/70 00:00      
                  It hasn't got the same sort of ports as the 805x            01/01/70 00:00      
                     Yes, you are right            01/01/70 00:00      
                     8255            01/01/70 00:00      
                        Why would a novice build a PCB?            01/01/70 00:00      
                           Products and packages            01/01/70 00:00      
                              gee ... and I've never before heard of Tekmos ...            01/01/70 00:00      
   quick thoughts            01/01/70 00:00      
      You're on the right track, EXCEPT ...            01/01/70 00:00      
   on analogy...            01/01/70 00:00      
   8255 in India            01/01/70 00:00      
      It makes little sense to teach 8085, or does it?            01/01/70 00:00      
         I Agree            01/01/70 00:00      
            Do not attend            01/01/70 00:00      
               Every Collage            01/01/70 00:00      
               Not "College Think"            01/01/70 00:00      
      just had a thought - Lynn, Jez            01/01/70 00:00      
         Enhanced 8255            01/01/70 00:00      
            If all one could get is PLCC ...            01/01/70 00:00      
         well i have come up with a way.            01/01/70 00:00      
            There is a problem with FPGA\'s ...            01/01/70 00:00      
               nRD is needed in the 8255 as a clock            01/01/70 00:00      
                  Perhaps that\'s because they use latches            01/01/70 00:00      
                     on using undefined states            01/01/70 00:00      
                        grounding chip select is risky            01/01/70 00:00      
                           but what if it is grounded?            01/01/70 00:00      
               clinging to the past            01/01/70 00:00      
                  Don't forget the Maxim/Dallas parts, they use 5V0            01/01/70 00:00      
                     as usual apples and oranges            01/01/70 00:00      
                        you need to see it from Richard's viewpoint            01/01/70 00:00      
                           he has the right to his viewpoint ...            01/01/70 00:00      
                              There's room for disagreement on that ...            01/01/70 00:00      
                        "development" and "replacement" are different            01/01/70 00:00      
                           well, for the (in)famous "universal board"            01/01/70 00:00      
                              If there's a market, it would be for THE part            01/01/70 00:00      
                                 again mixing parametres            01/01/70 00:00      
                                    the value of an improved 8255 in legacy applicatio            01/01/70 00:00      
                                       I'm not so sure about that one ...            01/01/70 00:00      
                                          OK that was just one - bad - example...            01/01/70 00:00      
                                             My point and point of view, all along has been ...            01/01/70 00:00      
                                    let's clear up some misconceptions ...            01/01/70 00:00      
      8085?            01/01/70 00:00      
         RE:8085?            01/01/70 00:00      

Back to Subject List