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

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
09/12/11 11:59
Read: times


 
#183714 - Another Suggestion....
Responding to: ???'s previous message
If the number of I/Os is relatively small and your host MCU actually has 5V tolerant I/Os you may want to consider another approach. You could use a low pin count MCU such as the Atmel AT89LP4052 as your programmable transceiver. In the 20-pin package this part has 15 I/Os that can each be configured as Input Only, Quasi-BiDirectional, Output PushPull and Output OpenDrain. You could run this part at 5V to service the ribbon cable with standard 5V drive for 6 or 7 I/Os and then have the other side drive to your MCU using the bi-directional I/Os mode where you would utilize the 5V tolerance of the host MCU. This MCU uses a VIH spec of 0.65*VCC which at 5V operation translates to 3.25V. You would just get by with being able to drive to the transceiver MCU from the hosting MCU if you can get its outputs fully up to 3.3V. Another trick is to try running your hosting MCU to the high side of its VCC range to gain say another 100mV of output high drive.

The programming in the AT89LP4052 would be a tight loop that copies the port bits on one side to the other and back based on the desired cable device configuration. This part us a "single clocker" at 20MHz and I'll bet that the a tight loop programmed in assembly may be able to achieve an I/O reflection through the part at 2 or 3 MHz.

You also have the interesting possibility to add some additional programming to the MCU to perform some special handing of the device interface. This could end up impacting the reflection rate through the part depending upon how the programming was done and the special functions being performed.

Michael Karas


List of 40 messages in thread
TopicAuthorDate
Bit-configurable transceiver chips            01/01/70 00:00      
   Suggesting....again...            01/01/70 00:00      
      Alas 5V needed and ribbon cables are a bit "rough"            01/01/70 00:00      
   Another Suggestion....            01/01/70 00:00      
   level translator            01/01/70 00:00      
      Alas, "weak output drive" and no input hysterese            01/01/70 00:00      
      TI sn74gtl2010            01/01/70 00:00      
         or NXP GTL2010,GTL2000            01/01/70 00:00      
            looks promising            01/01/70 00:00      
            NXP GTL20xx -> NVT20xx            01/01/70 00:00      
         Need to read more to understand them            01/01/70 00:00      
            looks like the cat's miauw            01/01/70 00:00      
   NXP has ....            01/01/70 00:00      
      I2C or SPI just can't get even close to the huge bandwidth            01/01/70 00:00      
         nope, no I²C clocks            01/01/70 00:00      
            extender, not expander            01/01/70 00:00      
   I have been wondering this myself            01/01/70 00:00      
      Supported capacitance seems to be the snag            01/01/70 00:00      
   Differential SPI            01/01/70 00:00      
      Serial -> buffers on adapter boards is a potential solution            01/01/70 00:00      
   Have you considered programmable logic?            01/01/70 00:00      
      I had suggested this as well            01/01/70 00:00      
         Yes ... I remember that ...            01/01/70 00:00      
            Long life            01/01/70 00:00      
               They seem to live a long time ...            01/01/70 00:00      
                  Adapters            01/01/70 00:00      
                     These aren't necessarily so "huge"            01/01/70 00:00      
                        Not huge in size            01/01/70 00:00      
                           Some of them can handle that.            01/01/70 00:00      
                              Yes and no            01/01/70 00:00      
                                 There are ways ...            01/01/70 00:00      
                                    Body diodes            01/01/70 00:00      
                                       serial termination ...            01/01/70 00:00      
                                          Yes, current- and bandwidth-limiting components used            01/01/70 00:00      
                                             you youing whippersnappers, pay attention            01/01/70 00:00      
            pedantry, again            01/01/70 00:00      
               What would YOU suggest?            01/01/70 00:00      
   Here's a thought ...            01/01/70 00:00      
      Probably            01/01/70 00:00      
         I'd sugest you consider older CPLD's            01/01/70 00:00      

Back to Subject List