Commit Graph

69 Commits

Author SHA1 Message Date
Mikolaj Wielgus 8a6a0ff7dc Allow inferred voltage/current sources to have a single float in Value
Update Opamp test to use this feature.
2022-08-24 06:19:38 +02:00
Mikolaj Wielgus b6f6d1ef81 Sim QA: Add legacy_opamp test to check legacy subckt fields
In particular, Spice_Node_Sequence needed some additional coverage.
2022-08-11 21:23:05 +02:00
Mikolaj Wielgus c669d55eb8 Sim QA: Use uopamp_lvl2 instead of uopamp_lvl1 in Opamp test 2022-08-11 18:32:20 +02:00
jean-pierre charras cc1e99ff5d QA test: annotate rlc.kicad_sch. only annotated schematic give reliable result. 2022-08-10 11:16:07 +02:00
Mikolaj Wielgus 128fedec1a Sim QA: Commit the missing rlc unit test project 2022-08-08 22:05:15 +02:00
Mikolaj Wielgus f6771ed789 Sim QA: Add rlc project to test RLC ideal model inference 2022-08-08 17:06:50 +02:00
Mikolaj Wielgus 9e7bc585ef Sim QA: Test only specific hardcoded points of the results
We don't want to assume Ngspice results to be very deterministic.
2022-08-04 21:07:43 +02:00
Mikolaj Wielgus 800b512fe8 Further cut off digits down to one after period in sim QA
We just want to make sure our netlist exporter works. Being
deterministic is the simulator's job, not ours, though perhaps we should
investigate that too eventually.
2022-08-03 16:25:00 +02:00
Mikolaj Wielgus 37209bb496 Try to fix Ngspice QA by resampling and cutting off value digits
`linearize` command resamples the data. Fourth and further digits from
decimal point are cut off by using `wrdata` command instead of `write`.
Oddly, "sources" unit test is not working (so it's still uncommented) --
some substantially different values are generated when generating the
reference with standalone Ngspice.
2022-08-03 14:28:01 +02:00
Mikolaj Wielgus d7e5254c17 Comment out failing sim QA tests
The tests will probably have to test against manually specified bounds
and inexact points.

Added a previously mistakenly uncommitted cmos_not.csv.
2022-07-30 09:19:20 +02:00
Mikolaj Wielgus 7cf5138c63 Sim: Bugfixes, mostly for MS Windows compilation errors
Unfortunately, Windows headers define a lot of macros for common words,
so we had to rename some enums to not collide.

We also fix some of the many bugs related to the new simulation
architecture and the Spice Model Editor dialog.
2022-07-30 02:25:34 +00:00
Mikolaj Wielgus 739b9255d9 Sim Model Editor improvements
- Tab-switching,
- Automatic expansion of categories on tab-switch,
- Various minor simulation improvements,
- Various new simulation-related bugfixes.
2022-07-30 02:25:34 +00:00
Mikolaj Wielgus 6984f63af8 Sim: Transmission line models
Implement transmission line models and perform some adjustments to
the current models. Add some QA tests.
2022-07-30 02:25:34 +00:00
Mikolaj Wielgus fe38c622a9 Sim: Improvements to model serialization
Don't serialize parameters in certain models for default values. Infer
models from Value field for some kinds of models. Resolve synonyms when
loading models from Spice libraries.
2022-07-30 02:25:34 +00:00
Mikolaj Wielgus 6450ec6b85 Sim: Spice netlist exporter rewrite
Rewrite the spice exporter to work with the new simulation model
architecture and data model, with many bugfixes related to the latter
two along the way.
2022-07-30 02:25:34 +00:00
Jeff Young 09c9bc3037 Fix net ordering in gold file. 2022-05-27 21:28:17 +01:00
Jeff Young e9e1878d96 Attempt to fix eeschema tests.
(Slashes in netnames must be escaped or else we think they're path
separators.)
2022-05-27 17:53:11 +01:00
Roberto Fernandez Bautista 55f22c526a Fix annotation of incomplete multi-unit symbols and re-annotation of duplicates
Fixes https://gitlab.com/kicad/code/kicad/-/issues/11496
2022-05-26 09:23:36 +00:00
Sylwester Kocjan cfb8fbc644 qa: move eeschema data to appropriate directory 2022-03-30 11:13:23 +00:00