Re: Black boxes, Open servo card

> BTW, so is EMC, when used for REAL servos. Anybody have an "open" stg
> card?

I modified EMC to talk to a D/A card I found in the trash at work,
used in conjunction with an LS7266 DRO board like Kulaga's for
feedback. Could probably graft something similar onto the design of
that DRO board for $50-$75 in additional parts for 1/2 channels.
There may be some fancy features on the STG card but they aren't used
by EMC. The hard part was figuring out what files I needed to make
new versions of and modifying the Makefile appropriately... supporting
the new board was fairly trivial once I reverse engineered it (they
didn't throw the manual out with it, but there wasn't much on it
besides the DACs and 74x00 series logic... and a shorted voltage
reference brick)

> I suspect that the reason most SBC's are not used "as is" for this is
> that they are not well suited to the job and more expensive than need
> be: (they have too much of "this", and not enough of "that", whilst
> duplicating functionality already present in the pc; i.e., why go to
> an sbc-based black box, if we're still dealing with granularity
> issues due to pc type timers?)

The PC timer is very primitive compared to that available in most
SBC's - plus the PC has only one (of 3) really available and you can't
reprogram it without also dividing it down in software to support
things that need to keep running at its standard rate.

I think the SBC stumbling block is that you have to make custom boards
or pick a board vendor and everybody has to buy from them and we all
get sunk if they go under - although one could pick a simple one and
only use its more easily duplicated features with the idea of fabbing
replacement if/when that became desireable or necessary.

Mariss's project looks like a possible solution to much of this.

> So whether we NEED a "black box" is a matter of an individaul
> developers' abilities, beliefs and focus.

I'd like to see the realtime core of programs like EMC and maybe
Master seperated out by a well documented interface so that one could
combine new front ends with their realtime cores, or replace the
realtime stuff with black box, or whatever. That in addition to being
able to substitute a black box step-rate generator for a
velocity-command servo, which seems like it would be very easy (same
task as supporting my trashbin d/a board)

Chris Stratton

PS - If someone wants one enough to 'fund' development, I am willing
to do PCB design for an 'open' D/A and encoder board for a 'fee' of an
ISA prototype card, one of the first run of pcb's, and a set of parts
to stuff it. Or I'm happy to correspond with anyone else who wants to
take on the project. Jon Elsons' project should be checked out for
suitability first though.

-------------------------------------------------
Christopher C. Stratton
Engineer, Instrument Maker, and Horn Player
ccs@... 617 628 1062
http://web.mit.edu/~stratton/www/brassbuild.html