perm filename BKPH.MSG[1,VDS] blob sn#131682
filedate 1974-11-19 generic text, type T, neo UTF8
Date: 19-NOV-74 9:57 PM
From: Vic Scheinman (VDS @ SU-AI)
To: BKPH @ MIT-AI
- - - -
This is a reply to your message of Nov. 18.
1)- Joint stops on outer joints. Yes, this has been a problem. To remedy the
situation, I have installed a type of cushioned stop on joint 4. On joint 5 I
glue some rubber to the case where the joint hits. This cushions things a bit. On
joint 6 I have added a slip clutch to the motor. As you probably know, intertia
is reflected by the square of the gear ratio. Thus with a gear ratio of 35 or so,
the motor inertia looks like over 1000 times its armature inertia at the output. The
slip clutch allows it to slip a bit, rather than put excessive forces on the gear
2) The wiring has been improved around joint 5. There is no more catching and
cutting. This has been done by using smaller, more flexible wire which is
teflon insulated for better abraision resistance and low slipping friction. I
am currently using 36 ga. stranded wire. Also, all the shafts, housings,projections
etc. inside have been smoothed to prevent the problems you mention.
3) Case deflection. The arm has been designed with the shell type structure in mind.
I plan to eventually try even thinner case material. The case is very weak withone
half removed. To make it strong and stiff both halves must be in place and the
edges securely fastend together with the strips and screws. On the next arm, you
will immediately note that this case fastening detail has been worked out much
more simply. Structurally, box structures are very strong and stiff, but open
channels are relatively weak, especially in torsion. In other words, no need to worry
about deflection, there is just about zero in the properly screwed together arm.
4) Size of wrist. I am sending you a photo of the new wrist I spent more than
$3000 trying to develop. You will note that it is MUCH shorter than the current
wrist. Well, I've given up on it for the time being as it has proved to expensive
to produce at this time. Also there are still some design problems. The arm you
will get will have a wrist similar to the one you currently have, only it will be
about a half inch shorter. I agree, a shorter wrist would be great, and I hope
to get this expensive prototype design working, within a year or so, depending on
You might like to know what is happening with your two arms on order. Well,
I am currently building 6 of these improved version arms. The first one was just
sent to Texas Instruments Corp. You are supposed to get the next two. Purdue gets
another one, and Dave Waltz keeps saying he wants two of them. The fact that
T.I. got there's ahead of you should make you glad, not mad- as many of the bugs
were ironed out in making that one. Also, since they are paying a couple grand
more for the arm, I figured it was financially advantageous to get an arm to them
first.... of course they may just decide to return it, if its not to their liking.
Also, the fact that I was sending an arm out to industry scared me a bit... into
making it a reliable as I could. Once again, I feel that your next arms will
be vast improvements over the first.
I also got you fowarded summary of arm problems as outlined by
Dick Waters. I will reply directly to him.
SUN@MIT-ML 11/19/74 16:50:39
1) EXCELLENT NEWS ABOUT PYTHON.
2) IN CONNECTION WITH YOUR SUGGESTION ABOUT A LISP IAP
COURSE, LISP MANUALS HAVE RUN OUT AND PROBABLY WON'T GET
CONSED UP TILL SOMETIME AROUND JANUARY OR SO, THOUGH
I DOUBT A COURSE WOULD NEED MORE THAN A COUPLE FOR