Title: Intellicon-4 Multi-port Serial I/O Rev. B vs. Rev. D Changes
Keywords: Intellicon, Rev. B Rev. D, changes
Date: January 13, 2006
At the time when the Intellicon 4 was revised from Rev. B to Rev. D, there were a few subtle changes incorporated. From inspection of the two revisions, most changes appear to be minor is nature, such as a resistor, or other discrete device. There is one change that is not so noticeable that is worthy of mention.
Nature of the Problem:
Close inspection of the Rev. D board will reveal that Pins 1,2,3 and 14 of the 75154 chips are connected to +5V while on Rev Bs they are NC. This means that Rev D boards are running the receivers in “Normal Mode” and the previous rev was running in “Fail Safe Mode”. For those of us not familiar with these, Normal mode requires the inputs to be driven + or – to be put in a known state, while Fail Safe will set the input into the same as being driven – if left unterminated. So with a Rev B board DCD, CTS or RxD if left unconnected will return a known stable state of inactive. While a Rev D will not set the state reliably. Depending on a number of things it will set these states into a basically unknown state and the signals could change depending on noise.
a) Appropriate course(s) of action?
It is suggested that the customer verify that CTS and DCD are being driven somehow. If they are not and the software makes some sort of assumption about the state of these signals, this is likely the problem.
b) Possible solution(s)?
- Change the cabling so that all inputs are properly terminated. A possible answer might be to have the customer’s cable wrap the RTS to CTS and DCD to DTR The best cabling option will depend on the actual cabling being used by the customer.
- Change the user’s software so it has no dependency on the state of unused signals. This solution may still cause problems because the inputs ( DCD and CTS ) may still toggle at random, which will degrade performance.
- Change the 75154 threshold pins so that the chip is in Fail Safe Mode. Because of the way the board is routed, we cannot cut the traces directly, however, pins 1,2,3 and 14 could be cut from above with small side cutters and the remaining pin could be twisted until it breaks away from the board with some fine needle nose pliers. This will violate the warranty.
End of KDB-206