Datasets:
nickh007
/
sparam-conformance

Dataset card Data Studio Files
xet
 Community
1
expected_failures
list
expected_to_pass_all_laws
bool
file
string
n_ports
int64
name
string
note
string
physically_realizable
bool
sha256
string
tags
list
z0_ohm
float64
[]
true
passive_line.s2p
2
passive_line
Lossy 20 ps delay line, 0.5 dB insertion loss. The baseline sane case.
true
74d631c14f5c18e9e9cf06bde2579628ff15820fc4b652035cd09d2809f4166d
[ "passive", "2port" ]
50
[]
true
passive_resonator.s2p
2
passive_resonator
Shunt resonator, Q=30 at 20 GHz. Sharp phase slope near resonance -- the case where a group-delay check without phase unwrapping fails.
true
1a4b1e2f9f873fcdc03327dababdccce6fc3d6ab0a0968ac70b2f03a803c3f39
[ "passive", "2port", "sharp_phase" ]
50
[]
true
passive_attenuator.s2p
2
passive_attenuator
Ideal 10 dB matched attenuator.
true
ee4b89263c942825af357be7f0273716d62ea984da5d441bcaf204c70f59963d
[ "passive", "2port" ]
50
[]
true
matched_load.s2p
2
matched_load
All-zero S: perfectly matched, fully absorbing. A degenerate but legal network; checkers that divide by |S| must not blow up.
true
7a60bff097406dfcb7cab59bcc837935353727634161806c53eae99ba865470b
[ "passive", "2port", "degenerate" ]
50
[]
true
marginal_lossless.s2p
2
marginal_lossless
Lossless line with sigma_max = 1 - 1e-12. Sits on the passivity boundary; a checker with a too-tight tolerance false-alarms here.
true
2dc0808d30a664b78cc263f06a9a46b48f7920bb3e9a132d56dfe4364d9179b4
[ "passive", "2port", "boundary" ]
50
[]
true
passive_4port.s4p
4
passive_4port
Four-port with two independent thru paths. Exercises N>2 handling.
true
1c24a1157c72806fa771c2e592eb8751724c8ba5cae446dc7b798dd8b0302f17
[ "passive", "4port" ]
50
[ "energy_conservation", "passivity" ]
false
active_gain.s2p
2
active_gain
Delay line with 3x through-path gain. Creates energy: fails both the spectral-norm and the row-power tests.
false
3f78dfcd6598f6734633a65a37e30edc8c4486c89d0c2e29cc52720780f22c32
[ "nonphysical", "2port" ]
50
[ "energy_conservation", "passivity" ]
false
energy_row_violation.s2p
2
energy_row_violation
Row power > 1 when port 1 is driven.
false
05e30904110032dc75777b11b96dbde5398ae9e922ab64a89704f9675db7ae83
[ "nonphysical", "2port" ]
50
[ "energy_conservation", "passivity", "positive_real_z0" ]
false
negative_resistance.s2p
2
negative_resistance
|S11| > 1 gives Re(Z_in) < 0: negative resistance at the port. It unavoidably breaks energy conservation too -- a reflection coefficient above unity returns more power than arrives -- so this case cannot isolate a single law, and the label says so.
false
2eb111856d1bfcdfb53e93f873704073f21af7926cb2ed18f5b2ef5b3114fdcc
[ "nonphysical", "2port" ]
50
[ "group_delay_nonneg" ]
false
noncausal_advance.s2p
2
noncausal_advance
Phase advances with frequency: the output precedes the input. Passive and reciprocal, so ONLY the causality check should fire.
false
ede95a08f616205688b97eac8c7665b2491ce4a038767614ebc69ad0730ffe5c
[ "nonphysical", "2port", "isolates_one_law" ]
50
[ "reciprocity" ]
false
ferrite_isolator.s2p
2
ferrite_isolator
A ferrite isolator. NON-RECIPROCAL BY DESIGN and entirely realizable -- the medium is not reciprocal. The reciprocity check correctly fires, and that is a true positive for the law but NOT a defect in the device. Any tool reporting this must let the user say so.
true
574d727015c7c3bb436bf1f9723098ac12a73e0392d3e2045a1b356c53e57f30
[ "physical", "2port", "expected_law_failure" ]
50