perm filename ILL.L2[1,VDS] blob sn#117762 filedate 1974-08-25 generic text, type C, neo UTF8
C00001 00001
C00002 00002	\\M1BDR25\M2SIGN57\M3NGR25\M4NGR20\F2\CVICARM
C00015 ENDMK
\F4\←L\-R\/'7;\+R\→.\→S   Telephone:
\F1\CAugust 24, 1974

Prof. David Waltz
Coordinated Sciences Laboratory
University of Illinois
Urbana, Ill.61801

Dear Dave:

\J	Enclosed are the drawings of  the proposed Stanford PDP-11/45
interface and the arm driver electronics.  As I have mentioned to you
several times, not all of this  is complete, and none of the  various
circuits have been tested on the PDP-11.  

	The arm power amplifiers, the  brake drivers, and some of the
circuitry  associated with  the actual drive  of the  arm has already
been implemented on one of our arms conected to the PDP-6 and PDP-10.
This works fine, and is  being duplicated for our older arm which has
switching power amplifiers which create bothersome noise spikes.  The
big difference is the interface itself.  

	The  fellow  who  his  handling  the  design  of  the  PDP-11
interface  at Stanford  is  Tom Gafford  (tel.   415-497-4971).   The
enclosed drawings signed  TAG were done by  him. They are  reasonably
complete,   but  no  doubt there  will be  some  changes made  before
everything  works  properly.   The entire  interface  in going  to be
placed out near the arm.  The Unibus from the  PDP-11 will be brought
all the way out to this  small rack which is about 24 inches high and
the standard panel width of 19  inches. Mounted on the front of  this
rack are two DEC racks containing a total of 10 single density blocks
and 6 double density DEC blocks.  The pin side faces foward,  and all
the electronics, cables  to the  arm and other  peripherals, and  the
Unibus  cable  plug into  these  blocks  via  DEC  cards.   The  only
electronics not on this part of the rack is the forced convection and
natural convection heat  sink amplifiers.   These are mounted on  the
bottom and back  of the rack unit with a  muffin fan blowing directly
on the forced convection  heat sink amplifiers.   The main arm  power
supply  used  at  Stanford consistes  of  two  Wanlass  "Brute"  type
supplies with  20 or so amp capability at 28-30  volts.  Two are used
as we use a bipolar supply.  These supplies are mounted on their  own
chassis, and  are separate  from the  rack unit.   The  A-D reference
supply, the op-amp and logic supplies are mounted in the rack.  

	The reasoning  for bringing the Unibus out to the arm is that
the longer distance we run analog signals, the more  noise and errors
we are apt  to get.  Thus, bring the Unibus  out.  Gafford feels that
he can do  better than  a DEC  DR-11-C or DR-11-B  card,   so he  has
designed his own way of connecting the Unibus to the interface.

	In parallel with all this, I am having  my own simple and low
cost interface developed for running either a Stanford or M.I.T. type
arm from a  PDP-11, or  a NOVA, etc.  This interface is  simply a  16
channel 12  bit A-D,  an  8 channel 9  bit DAC, and 7  brake drivers.
All this  is mounted on three double height DEC cards (about 75 ic's,
and interfaces  to a  PDP-11 via  a DR-11-C  card (from  DEC).   This
simple interface does  not have sample and holds (  I dont think they
are necessary), nor does  it have DMA capability,  as I dont feel  it
has to operate very fast.  For the encoder  version of the arm (which
you  have ordered),  three quadrature decoder  and counter  cards are
required for the three encoded joints.

	I am having Fred Dhrenkenhan of  the MIT- AI lab make up  the
quadrature decoder  and counter boards  for me.   I will  include the
boards with the kit, so that problem will be solved for you.

	As  for the Vicarm version of  the arm driver electronics, we
are using our MIT arm electronics package with a larger power supply,
and beefier  amplifiers.  The amplifiers are  similar to what Gafford
is using at  Stanford, and so  are most of  the other drive  details.
The difference is that I am making these amplifiers and brake drivers
as a single unit with pc boards instead of DEC cards and blocks.  The
power supply  is  a much  less  expensive transformer  and  capacitor
filtering  type which  is  unregulated (I  dont  think regulation  is
required  for  the power  amplifiers).   This  arm driver  package is
similar to what I supplied  to SRI, only it uses built  up amplifiers
instead of RCA HC2000H power  hybrid amplifiers which are hard to get
these days.  

	I don't know whether I mentioned it or not, but I am building
a Stanford arm  for the National  Bureau of Standards.   Their driver
unit  will be similar  to what  you want.   We could make  a complete
driver unit (minus  the interface, but  including a hand  controller)
for about  $2000.   I know that  this sounds like  a sales  pitch- it
isn't,   as that  price is about  cost. Regardless of  whether you do
things yourself or buy I will give you details as we generate them.
	I suggest  you have  some people in  the know  look over  the
Stanford arm  interface plans,   and maybe  even call up  Tom Gafford
(best in the afternooons or evening here, as he's usually at Stanford
2 pm  to 2 am).   The decision  is between a  full blown, able  to do
almost anything interface  versus one tailored just to the arm,  with
no regard to operating super fast, etc.

	Now to the arm  machining details.  In a  day or so you  will
recieve some of the  bearings for the arm.  Also will  be some of the
corrected  drawings.   In a couple  of cases I  said wait.   That was
because the wiring  details were being changed.   We have worked  out
what appears  to be an alternative  solution, and will  be testing it
out shortly, and will let you  know what hole patterns are  necessary
to mount the connector blocks when we know ourselves.   This is not a
big thing,   and I hope  you will be patient on  the wiring slots and
holes in the  square tubes.  I might also  add that we  intend to  do
another thing  to help  you out.   We will be  supplying some  of the
assemblies complete- including some of the parts you were supposed to
machine.  This will be on  an exchange sort of basis.  The  reason is
that I feel it wise to have us assemble the joint 1,2,3,4,and 5 motor
assemblies as the  brush and  tachometer assemblies  are fragile  and
some careful fitting is required.  As I also mentioned, all the parts
listed on the parts list will be supplied, and they will be machined
to the  drawings.    This  will eliminate  the  possibility  of  your
accidentally scrapping  a part,  and having  to reorder or  whatever.
Most  of the parts have  arrived at Vicarm, and  we will be machining
them within the next two weeks.   When a reasonable number are  done,
we will send them to you. We will probably have to send another batch
later, as some items are back ordered. In any event, I feel sure that
we will have everything for you by October 10 or so.

	Dave, I  hope this letter  clears  up some  of your  concerns.
Call me should you have  any other furthur questions.  I admit to not
being too prompt  on my promises,  but I'm  trying to get  everything
right the first time, and its taking a lot longer than expected.\.

Yours sincerely,

Victor Scheinman