AW: Re: Analog Sequencer Poll

[JDMcEachin]

At 07:38 PM 12/23/96 GMT, brad sanders wrote:

>As a newcomer to this list, I've been following this discussion with
>some interest. I can't resist asking any longer: is there any
>*particular* aversion to the use of micros amongst the members of this
>list? I mean, after all: a micro can do all a 4017 does - it's just a
>matter of "coding" in silicon vs. coding in solder.

These guys are really into solder fumes! ;-}

I take it to be another manifestation of the great hardware/software divide.
People who do one are wary of the other.  Very few of us do both.


At 06:25 PM 12/23/96 -0500, Tony Clark wrote:

>   Well I'd imagine that the number one reason that we don't want to go 
>this route simply would be that CPU's come and go.  What is there to 
>assure that the special chip you use will still be in production years 
>from now?  TTL/CMOS will be with us forever.  :)

Members of the 8051 family will be around at least as long as DIP TTL.
Hell, I'd be more worried about everything going surface mount!

>   Using only TTL or CMOS technology keeps the overall design simple.  
>Simple enough that everyone shouldn't have any problems working with the 
>circuitry, modifying it how ever they want to. amd troubleshooting it in 
>the event something does go wrong.

Actually, if the circuit has ANY complexity, such as a full featured
sequencer, or a polyphonic key decoder/voice assigner, a CPU-based solution
will be less complex.  If you know the code works, debugging is just a
matter of verifying that the CPU's inputs are correct.  CPU designs get hard
to debug when the chip count goes up, and you don't have a good idea of what
it's supposed to do.

>   Besides, I seem to recall the large debate that went on about DSP 
>chips.  Everyone has their preference.  Using a special chip might lead 
>to some unhappy camper.  :)  This way, no one will complain (unless you 
>simply botch up the design).

Real hackers can program ANYTHING.  My preferences are based on cost, ease
of hardware design, and accumulated code.  If you're familiar w/ the 8051,
you'd know it's not based on the instruction set.  :)

>> Alternatively, has anyone considered using a 74AHC283 instead of a
>> counter? This is a four bit ALU and "accumulator" on a chip. By
>> controlling the four data input lines and the five ALU control lines
>> (three for op sel, carry in, carry out, etc) you can pretty easily get
>> "random" steps, or "skips" or go up or down or....
>
>   The chip possibilities are limitless.  :)

As a peek inside an old video game cabinet will attest too.  But, there
comes a point where laying out, purchasing, soldering, and debugging a mess
of logic gets overwhelming.  The question is, will this circuit be that complex?

If you look at the SuperSeque schematic, it's pretty easy to just stick the
chips on some perfboard and start wiring.  If you start adding functions,
such as a quantizer, random, clock, skip, etc., then you would probably be
better off laying out a pcb.  At this point, the CPU looks tempting.

Another advantage of using a CPU would be that you could go ahead and design
the hardware and some simple looping code.  Then you could add functionality
in software w/o touching the hardware.  If someone gets an idea for a new
function, great, just code it and send everyone a new .bin file to program
their EPROMs with.  You may find yourself wiring in some more switches for
control, but it's a lot easier than trying to add logic to a CMOS-based circuit.

Having said all that, I suspect that what will happen is that one of the
hardware guys will have a design and a pcb long before the software guys
agree on which CPU to use.  :P

JDM