edb024b7 BUGFIX: in lbm_is_number that would sometimes incorrectly state something is a number when it is not
a620e42c small tweak to REPL
f9911c0f terminate REPL upon problems, added error codes for traceability
c36f1e72 small tweak lbmref
04cad3d8 rearrange charts
86429f0f added images related to f32, f64 performance on various platforms
108791b9 update the reference manual
90dbbf63 added secs-since operation to repl
134c59a5 changes to perfomance plot python script
6a02f301 Update lbmref.md
b913f173 small note about usage of vesc hardware for data collection
0bbf246a update lbmref with some typo fixes
c62eec6d updates to lbmref
f6e188e0 adding some information about numerical operation performance
b8a55394 adding some information about numerical operation performance
ef58050d adding some information about numerical operation performance
e98d5b87 correct file 32bit
3002333a added perfomance examples
c74558cc adding to refman about numbers
ebca34cb type fix lbmref
3fde1dcb type fix lbmref
bfcd8247 update to lbmref
32b56fd3 added some info to lbmref
git-subtree-dir: lispBM/lispBM
git-subtree-split: edb024b75fe37c8f948fccb4e7d992cf783fd429
The 25MHz clock couples through the floating phase filter capacitor opendrain output on 1 phase. Which appears as a voltage offset and screws up flux linkage measurements at lower voltages. By oversampling, this offset is removed and flux linkage measurements now work properly at all voltages.
This update corrects the shunt resistance.
Adds generic calibration values for the current.
Adds 6 sample filtering for current (takes 6 current samples each pwm cycle and averages them to reduce noise)
Reduced max pwm in foc to 45khz
Disabled offset calibration on boot to speed up boot time. (Boots in under 2s now)
520fde1a fix promotion bug when converting i32 to u64
1ea55e8a fixes related to LBM BYTE type
fba7e9a9 bugfix related to subtraction of byte values
baab84c3 update lbmref with information about numbers
4a7ac41f addition of tests and changes to type promotion and flat values
7a7cf54a updates to type promotion and some changes to flat value encoding for 64 bit
0e30b53a remove some args from flatten_value_size
git-subtree-dir: lispBM/lispBM
git-subtree-split: 520fde1a9610ec05300d72a898dd477553c29624
610c2d2a added a silent mode to the REPL, --silent
79c69181 Fix bug in how errors happening inside of fundamentals are returned
de5c356b fix bug in addition that incorrectly promotes byte values to U type
64219f1b remove some printing in REPL
b24e4c39 fix bug in unflatten value related to 64bit int
baa5bfba fix bug in flatten_value related to the size of expressions that contain Byte sized data
5fc5a7df Merge branch 'master' of github.com:svenssonjoel/lispbm
52665236 minor changes to new repl
22070c29 update README
59b5144a update README
9e8c2a2a update README
582f2235 minor changes to new REPL
24d4b6e0 added storing of flattened result to the repl
412e6483 added a terminate option to the REPL
15af284f enabled loading and storing of environment files from the new REPL
b5319ebe fix bug in 64bit version, added a NULL check after readline in the REPL
05b3eb61 fix missing include repl
300213b9 preparing for loading and storing of global environment
fc3d20b0 sequence loading and evaluation of multiple source files when using multiple instances of -s flag on repl
1370102c loading source files
4a2f1c3d preparing for accepting source files from command line to the repl
9b45e4da small changes to new repl
ba47441e promoted the experimental_repl to repl status, it is still work in progress but at a much nicer state
fe16c290 cleanup some unused code in repl.c
e6705251 adding general purpose functions from vesc_express to the REPL
a0955b93 fix to state change logic
cfbb6114 started a revamping of experiment_repl now depends on readline, more to do
41c142a3 small fixes to vesc_express hil test scripts
ee763c72 work on automated vesc express fw build with vesc_express from main and lbm from master
3a2884b5 some maintenance and clearning, removal of unused function
ade15e9d removing some unused code
e16a8598 added loging of failed tests to a unique log file for each instance of run_tests.sh
ce9bdcdd fix potential out of bounds writes found by Marcos. Also removed a few dead-writes
5e1bad69 Small tweaks to eval. Changes to test script, store failing tests to test.log and print way less during testing.
93d14889 remove a blank line
04e1dc62 small refactorings
dbadb211 Trying to make the vesc_express hil script slightly more robust
99b6540d added a TTY available check before attempting to launch the tests
fb7fbee7 Added a slightly longer running vesc_express hil test as an experiment. The run_tests scripts for hil tests needs to be made more robust
0dc4b25c vesc_express hil tests now use the new vesc_tool feature of stdout redirection, no uart needed
78781da2 Added framework for Vesc Express HIL tests
git-subtree-dir: lispBM/lispBM
git-subtree-split: 610c2d2a3b0b4724a31929150789cda70b5aacfc
If m_speed_pid_set_rpm is allowed to exceed foc_openloop_rpm before index is found,
then there will be a jump in torque when index is found.
If m_speed_pid_set_rpm is allowed to exceed [l_min_erpm, l_max_erpm] then there is
a wind up effect since it will take time for m_speed_pid_set_rpm to return to valid
range due to ramp limit.
Needs to be rescaled back to int for storage, 8* higher resolution so the hw can use oversampling.
BLDC can use floats for these variables, since they get changed to floats later anyway. Might as well do it earlier to make the code simpler.
Removed a bunch of ifdefs by making is_second_motor = 0 when not using dual motors.
Fixed a bug in hw.h where it wrongly defined GET_CURRENT3 when inverted shunts was selected. GET_CURRENT3 will now always be 0 by default, removing the need for a few ifdefs in the foc code.
Hw_conf can define CURRENT_CAL1,2,3 to set individual calibrations for each current sensor.
By default no calibration is used. (x1.0)
For example this is used on the A50S by looping a calibrated current using a power supply through each phase. Then the correction is saved in external eeprom to persist across firmware updates.
Tested on FOC, DC and BLDC modes.
7ba933de lbm_lookup_extension_id now returns the actual extension symbol ID instead of the index into extension table
1c5c8da6 fix typo
0fb03637 defining different kinds of built-in symbols in some specific ranges. Cleaning out some old code.
b14dbe32 refactor of stack_push operations
4d1f5fd9 cont_closure_application_args refactor
aeb7d4d9 small tweak
8aafd407 expanding on the thuroughness of the ALWAYS_GC test mode
6eb4f614 clean out comments in cont_qq_expand_list
0d1e92b2 Bugfix in quasiquote expansion
c9bb4005 Added an ALWAYS_GC testmode and improved on the feedback from the tester
ccc3560a small refactor of cont_match and addition of 2 tests of match with guards
34609e1b tweaks and typo-fixes in lbmref.md
eb4e7def merror at lbm_lift_array if header cannot be allocated
git-subtree-dir: lispBM/lispBM
git-subtree-split: 7ba933de6261d5b33092f320adfc07050973b007
Benjamin Vedder
Euan
Carl Ljungholm