12 Aug
2019
12 Aug
'19
8:13 a.m.
On 08/11/2019 08:00 PM, Douglas Taylor via cctech wrote:
This is where the electrical engineer could help. How do
you determine how long a cable the 74LS240 can drive?
Well, there are several considerations. First, it takes
some current to charge up the cable capacitance. More
current charges the capacitance faster, but also creates
faster edges which cause more crosstalk. Then, the data
rate needs to be considered. Mag tape data rates are not
that high. So, for 1600 BPI at 45 IPS, the data rate is 72
K bytes/second, or about 14 us per byte.
Twisted-pair cable should have a little less capacitance,
and it is supposed to reduce crosstalk, so should work better.
The most serious problem is when many data lines switch at
the same time, it may contaminate the clock pulses and cause
bytes to be dropped or added.
With the low data rates involved, proper delays to allow
ringing to settle on the data lines and prevent short
crosstalk pulses from affecting the clocks should make the
system very tolerant of cable issues. But, maybe some
engineers didn't really optimize their logic for these problems.
Jon