perm filename NOTES.1[1,VDS]2 blob
sn#099236 filedate 1974-04-24 generic text, type C, neo UTF8
COMMENT ā VALID 00004 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00002 00002 MOTORS
C00004 00003 ARM PROGRAMS
C00005 00004 ARM SERVO PROGRAM DETAILS
C00007 ENDMK
Cā;
MOTORS
Globe motor details.
Typical demagnetization results
As delivered - back emf equals 9.2
Take out magnets and replace- back emf goes to 8.2
Take out magnets and press like poles together until they touch-
back emf drops to 6.2 or so.
So we see the potential torque loss due to dissassembly.
Careful disassembly gives only a 157 torque loss, but fooling
around results in up to 50% torque loss.
A good test for fully charged globe motor magnets is to see if it cogs. If it
does, then the magnets are at least 75% charged. If it doesnt, then you
have weak magnets.
Placing the poles of a globe motor between the poles of a large alnico magnet
seems to properly recharge the magnets. As an example- recharging the motor above
in this manner yeilded a back emf of 9.6. Higher than the original number- but no
doubt, when put to use, the magnets can be expected to discharge a bit in the
presence of demagentizing armature current induced fields.
As a furthur note- testing the motor with the large charging magnet between the
poles gives an emf of over 15.
ARM PROGRAMS
TO create a macro type
DEFINE MACRONAME
THEN THE INSTRUCTIONS
TERMINATE WITH A BLANK LINE
TO EXECUTE, TYPE
MACRONAME
DO
TO SAVE A FILE TYPE
BEGIN FILENAME
MACRONAME
MACROCNAME
END
TO EXECUTE A FILE TYPE
DO FILENAME
ARM SERVO PROGRAM DETAILS
Lou's programs are under FAIL - Look under Arm Spacewar code-
Line numbers 1554 to 1563
Here the trajectory coefficients are taken from the list, which contains
5 coefs. (a4 to a0 , the total trajectory time in microseconds, and an
address of the next trajectory coef. list.)
ELL1: MOVE AC,A4(I) Loads AC with A4 coef. of Ith joint
MUL AC,MQ Multiplies A4 by time coef (<1), a 36 bit word)
ADD AC,A3(I) Adds the A3 coef. to AC (18 bits long)
MUL AC,MQ Multiplies AC by time coef.
All of these operations are fixed point operations.
ADD AC,A0(I) Adds last coef. to AC, so now we have
(((A4)t)+A3)t+A2)t+A1)t+A0 which is a normal quartic.
Lou says that most trajectories are now cubics which means
that the A4 term is zero. This is a fixed point equation.
XOR,AC,SC(1) Adjust the contents of the accumulator by the Sign Factor
Which means change the sign if required.
FSC AC,0 Convert contents of AC to floating point.
The above sequence is done for each joint, each jiffy. All that changes is the
joint number (I) and the time t, which is in microseconds, and counts down the
trajectory length time stored in menory. Thus, this is how a typical expansion
takes place.