Difference between revisions of "Todo"
From DIDEAS Wiki
(→hours) |
(→hours) |
||
Line 13: | Line 13: | ||
*get the KCA code working. | *get the KCA code working. | ||
− | =hours= | + | =Sep 09 hours= |
==Sun : Sep 13, 2009== | ==Sun : Sep 13, 2009== | ||
*15 mins - admin | *15 mins - admin | ||
Line 47: | Line 47: | ||
− | ==Wed Sep 9 | + | ==Wed Sep 9 == |
− | + | * 4:05pm - 1am (-1 hr dinner w/RC) + 1.5 hrs on phone at home== | |
− | + | ** discuss status of fixture with Rob (wanted to see if there were problems.) | |
+ | ** spoke w/Rick and Zhixiu on status then tune up the AK to full power. | ||
− | + | **Met with Kristin we experiment with lifefix to find cause of resonance | |
− | + | ** experiments with life fix and Q of resonance | |
− | + | ** clampped masses to end of drive ball screw, fixture, etc, nothing seemed to change resonance freq | |
− | + | **found that there was a angle dependence on the Q of resonance.; adjust MC current loop PI settings which had no effect on frequency, only the Q | |
− | + | **confirmed ZERO cogging in motor outside of lifefix | |
− | + | *measured LW thrust with motor leads shorted, and when driving 'constant' current; found significant ripple. | |
− | ==Tue Sep 8 | + | ==Tue Sep 8 == |
− | + | * 3:15 - 2:20aa | |
− | + | ** working on life fix, demo results, rick observed that lw thrust is very low; we use loadcell as a conformation and discover there's a problem | |
− | + | ** turns out that charge amp scale values were different from those on my PC viewer, and ankle code ( guess they had been reset.) | |
− | + | ** discuss with Zhixiu alignment of the datasets. figure out how a PVT works, | |
− | + | ** final turning of life fix and leave it running | |
− | + | **find that at high thrusts there are problems. | |
+ | **modified ReadLoadCell.cpp - adjusted sample rate, added time to filename, improved file write | ||
− | ==Mon Sep 7 | + | ==Mon Sep 7 == |
− | + | * 4pm - 4:15am | |
− | + | ** working on life fixtujre | |
− | + | ** scripted pfcmd in linux, got test jit working | |
+ | ** burned out rs232 cable due to groun problems? | ||
=Aug 09= | =Aug 09= |
Revision as of 17:23, 13 September 2009
- garden12 ; emg_2009 ; Enc_2010 ; AKENC ; Kknee1 ; Br46_v1 gps1 ; lifefix ; battery ; loadwasher ; elecdog ; Brushless_control ; ceb_misc_prj ; pf_top_nav ; T
- ceb_log todo BMdoc notes hours_1012 hours_1101 hours_1102 hours_1103 lifefix_log T
- PF Users Navigation:
- PCB; Lifefix_users; Pf_users; benchtest_users; CalibFix Users; Hardware; Assembly ; iochan ;
- PCA : PCB ; AKENC SC/MC 218 ; FET 217 ; IMU219 ; Swifi ; Rev200_mods ; PCA Inventory
- Special Commands to the Ankle (PFCMD) : State Controller Commands; Motor Controller Commands; Python Examples; IMU Commands ;PFCMD_PY; Pf_calb_table_py; Virtual spring test; PF EEPROM
- DOC: Pf_users; Powerfoot Keyboard User Interface; Steps for Manual Tuning; "Dashboard" Program For Assisting with Tuning
- NEW (CEB) WIKI
- Torque Feedback Controller Guide
- http://www.chdist.com/displayproductdetail.do?itemGroupOID=309&baseItemOID=500
- http://www.chdist.com/displayproductdetail.do?itemGroupOID=306&baseItemOID=495&skuSearch=Y
Contents
todo
- investigate apparent current overflows with large backdrive freq.
- figure out what is going on with motor resistance.
- recode PI loop to saturate Isum
- get the KCA code working.
Sep 09 hours
Sun : Sep 13, 2009
- 15 mins - admin
Sat : Sep 12, 2009
- hours 30 mins + 30 mins home depo
- discuss LF status with RC (nearly vibrated off table, ball nut and loadwasher out of AK)
- attempt to import large overnight text file into matlab. need to write a new converter.
- trip to HD to purchase blue SP and clamps to hold LF
- 4:05 - 9:10pm
- greatly improved Kistler charge amp class, mostly working
- 1 hr discussion / interview with Robot Designer
Fri : Sep 11, 2009
- hours : 3:10pm to 9:30pm
- create background process for move_ecat from python, rewrite the copley_control class
- explain motor PI loop to ZH, discover improper saturation of I term.
- w/ZH use sweep frequency to drive fixture under mechanical various conditions
- start life fixture
Thur Sep 10, 2009
- 12:45 - 4:20 pm; email re test fixture, phone with RC re lifefix; starting writing firmware to allow getting/setting MC variables by para_id, write python serial_port_helper; write program to test belt.
- Arrive IW 5pm - leave 3:55am (-2 hr for dinner)
- test min/max
- discuss with RC the torque ripple.
- replace broken belt
- write motor current function generator
- do belt resonance exprements, send plot to RC / ZH
- help a proper belt broken test
Wed Sep 9
- 4:05pm - 1am (-1 hr dinner w/RC) + 1.5 hrs on phone at home==
- discuss status of fixture with Rob (wanted to see if there were problems.)
- spoke w/Rick and Zhixiu on status then tune up the AK to full power.
- Met with Kristin we experiment with lifefix to find cause of resonance
- experiments with life fix and Q of resonance
- clampped masses to end of drive ball screw, fixture, etc, nothing seemed to change resonance freq
- found that there was a angle dependence on the Q of resonance.; adjust MC current loop PI settings which had no effect on frequency, only the Q
- confirmed ZERO cogging in motor outside of lifefix
- measured LW thrust with motor leads shorted, and when driving 'constant' current; found significant ripple.
Tue Sep 8
- 3:15 - 2:20aa
- working on life fix, demo results, rick observed that lw thrust is very low; we use loadcell as a conformation and discover there's a problem
- turns out that charge amp scale values were different from those on my PC viewer, and ankle code ( guess they had been reset.)
- discuss with Zhixiu alignment of the datasets. figure out how a PVT works,
- final turning of life fix and leave it running
- find that at high thrusts there are problems.
- modified ReadLoadCell.cpp - adjusted sample rate, added time to filename, improved file write
Mon Sep 7
- 4pm - 4:15am
- working on life fixtujre
- scripted pfcmd in linux, got test jit working
- burned out rs232 cable due to groun problems?
Aug 09
- car insurance - discuss metlife w/DWB
- MITFCU - call about fees for savings/ money market.
- Report expenses for Chicago trip
- Locate receipts for MIT video card purchases.
- contact WI re sales tax collection error
IW
- test RTOS
- PCB - isolated RS232, SWIFIs
7-10-09
- for monday - damping FF at zero degrees.
- send out loadwasher.
7-4-09
- force fields are loaded automatically, need to ENABLE
- (yesterday)(fixed issues of feature bits and runleve.
- SC - check for hall zero - but to correct, need the spline table.
7-3-09
- force fields are loaded automatically, need to ENABLE
- IMU automatically on (why is EEPROM ignored?) (believe causes by feature bits in walking_param)
- check for hall zero location
- chech for normal SG offset.
- PVF engages at zero vel crossing, allow for vel-offset.
7-2-09
- IMU signal processing
- accelerometer, direct output ; ratiometric, powered by 3.0V ref
- z axis gyro : powered by 5V, uses 3.0V ref on Vratio, 267k at Sumj; dynamic rage = 500 deg/sec
- x&y gyro : 3v power supply
- gyro and accl signals all have 1k, 0.1uF RC filter
6-30-09 todo
- zero encoder based on calibrated hall sensor.
- write python program to bump overload parameters, and monitor max
- PC programs to load/save defaults
- program to convert matlab calb files into python calb table
before next track test
- create a command that has a single parameter to specify the type of code to run
- wake into mode 0, sensors are operational (for test jig)
- wake into mode 0, but also do commutation (for test jig where ankle will be driven.)
- wake into walking mode, with impulse current = 0
- wake into walking mode, using EEPROM defaults, but impulse current = 0
- wake into walking mode, using EEPROM defaults, with impulse current
- DEMO mode : using EEPROM defaults, relaxed enabled, full impulse current
- ability to specify which EEPROM table is the default.
- ability to read out the SG offset log - seems to have an issue...
- DONE : use CRC with WIFI_FAST
- MORE WORK : load / save walking parameters
- IS USED FOR HALL SENSOR : use calibration table
- TODO : use motor current offset.
- VERIFY : use force fields
- TODO : glue down hall sensor.
statemachine
- detech motor faults - just to holding state.
- overload warning - detect and servo to safe position.
- integrate in the EEPROM code.
- VERIFY : during swing issue a test current into the motor to measure it's resistance (store somewhere)
- statemachine - look at MC offset, if there is much disagreement with defaults then signal error
- statemachine - log SG offsets
- add thermal model for motor and ACS712 to state controller
- code to make sure the motor is moving correctly when reseting the encoder.
- learn to use the EEPROM. log SG torque at boot, and during each step.
- capture initial SG offset into EEPROM
- capture SG offset during swing into EEPROM
PCB IW812a
- encoder breakout
- current sensor - isense1 (unipolar)
- faux SG board
- sg breakouts x 4
- pser1d - udp compat pser
- fup1 - gender changer
PCB todo
4 layer
- re-spin motor and state controllers? Knee PCB?
- send out PSER1D and other boards in 4layer nice.
- WIFI w/2Meg RAM
- electronic fuse, ballancer, ON/OFF switch - battery pack less the charger.
- FUP flash drive datalogger.
- if needed - FUP datalogger w/Wifi / BT and 2meg ram
- motor voltage circuit
- In VNC1 : change 3.3V regulator to MCP1700, 3.3V ; 2.2M resistor to 2.0Meg; voltage and resistor for the LT3470 regulator
old
- verify thermal models Thermal data
- need a way to verify that TVS is operational. need to see what effect T has on TVS voltage.
- MC code improvement
- verify Q15 operation of code on MC
- MC feedforward.
- motor controller safety and thermal model for motor and FETs - needs to stop motor....
- motor commands needs to allow readback of selected sensor (so we can cycle through lots of sensors)
- current measurement PCB (with Y voltage measure)
- encoder breakout boards
- SG breakout boards
- new WIFI wireless