8 Jan
2012
8 Jan
'12
7:52 p.m.
On 7 Jan 2012 at 20:21, Tony Duell wrote:
For a suitable definition of 'anybody', which seems to exclude almost all
the memebrs of the group I was working in at the time. But anyway.
And yes, a decoder with the outputs ORed together (In the case of an
open-collector inverted output decoder, wire-ANDIng them is equivalent of
course) is equivalent to a multiplexer wit hthe data inputs tied high or
low. 'Anybody' who knows what a multiplexer consists of will realise
that. It's also equivlaent to a 1-bit wide PROM. Many programmable logic
devices consist of an 'AND array' which combines the input varialbees and
their inverses followed by an 'OR array' which combines the outputs of
the ANDs to produce the otuptus of the device. Some of the arly PLE chips
(the 82S100 springs to midn) had both arrays prgramamble. A PROM (or a
multiplexer, or an open-collector decoder with the otuputs tied toghter
as here) is a fixed AND array geenrating all the miniterms of the inputs
(a complete decoder) follwed by a prograable OR array (you get to specify
how they are ORed together to form the otuput). The standard PAL/GAL
architecture is a programmable AND array follwed by a fixed OR array
But that wasn't part of the problem Had one
of the signals been ECL,
I'd have specified adding a 10124 or 10125 as appropriate. The problem
wasn't to get it down to one chip. It was to come up with somethign
that could be put on the PCB, so that the artwork could be finished
and the boards made. If it has been 2 ICs, that would have been fine.
So I wanted to do something non-obvious. Anyone knows that a
MUX/decoder implments a truth table of its inputs, so the problem you posed was obvious. I proposed a workable solution
that no one else
Actually, I dispute that. Yes, the design works perfectly on paper. But a
deisng is only workable if you can get the devices needed ot build it.
It's no good deisnging with devices you wish would exist.
And I've been through a lot of databooks and catalogues and I can find no
mention of the '156 in 74F, 74ALS or 74AS (and actually not 74S either).
So a suitable high-speed decoder doesn't exist. FWIW, the '559 (4-16
line, open-colelctor version fo the '154) doesn't exist in the
high-speed families either.
Do you have evidence it exists? If so, I'll be happy grant you have a
workable solution
had. You didn't specify what the mystery device
was driving--if it
had been a line that required an OC output, the shoe would be on the
other foot.
Of course. I did say 'TTL level signal' I suppsoed that could have meant
an line shared iwth other open-collector deivces, but the default
interpretation of 'driving a TTL signal' would imply to me a totem-pole
driver. It never hurts to check these things, I agree.
-tony