62500 bench: analog clean, lamps lose ~6.5% — RIO speed ladder complete

NAK=69/~1060 lamp commands explains the observed stuck-bright/missed lamps (fire-and-forget writes turn each corrupt 4-byte burst into a visible state error) while 2-byte traffic is pristine (analog 88.1%, counters flat, zero wedges, worst gap 0.14s). The HC11 RX ISR's worst path sits right at ~320 E-cycles. Ladder: 9600 production, 31250v2 CLEAN (FastRIO), 62500 needs host NAK-retransmit, 125000 CPU-walled. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-18 13:08:46 -05:00
parent e718ddebfd
commit 03bd5830ff
2 changed files with 53 additions and 0 deletions
@@ -318,3 +318,26 @@ failed. Direction asymmetry nails the cause: board->PC was near-perfect
 estimate predicted. Verdict: 125000 NOT viable. Sweet-spot candidate:
 `RIOv4_2_patched_62500.bin` (320 cycles/byte, MAX232 in-spec, ACK-wait
 80 ticks = 12.8ms), sha256 below, awaiting bench.
+
+### 62500 bench results (2026-07-18) — the ladder is complete
+
+`testlogs/riomash-patched-62500-20260718-130333.log` (132s): analog
+pristine (88.1% delivery = the structural drop-gate ratio; worst mid-run
+gap 0.14s; counters flat; zero wedges; status exchanges fine) but
+**NAK=69 of ~1060 lamp commands (~6.5%)** — 4-byte bursts still outrun
+the RX ISR occasionally, and fire-and-forget lamp writes turn each loss
+into a visible stuck-bright / missed-bright lamp. The ISR's worst-case
+path therefore sits right at ~320 E-cycles.
+
+**The RIO speed ladder (2 MHz E-clock, all with wedge fix):**
+| rate | bytes OK inbound | verdict |
+|---|---|---|
+| 9600 | all | production (native-game compatible) |
+| 31250 v2 | all (NAK 0.3%) | CLEAN — FastRIO recommended |
+| 62500 | 2-byte yes, 4-byte ~6.5% loss | usable ONLY with host NAK-retransmit |
+| 125000 | ~40-65% loss | not viable — CPU wall |
+
+Host-side NAK-driven retransmit (the protocol's intended use of ACK/NAK;
+RioSerialLink currently fire-and-forgets outbound) would make 62500
+lamp-solid (1 retry -> ~0.4% residual, 2 -> ~0.03%) and is good
+robustness at any rate. Until then, 31250 v2 is the honest FastRIO.
@@ -0,0 +1,30 @@
+== RIO mash test :: COM1 @ 62500 8N1, 300s, chip label 'patched-62500' ==
+   lamp echo ON (drives reply/lamp collisions), wedge threshold 2.0s, app auto-recovery DISABLED
+   log: C:\VWE\riojoy\riomash-patched-62500-20260718-130333.log
+   firmware: 4.2
+   counters before: RestartCount=0  AbandonCount=0  FullBufferCount=0  [board/lamp faults reported!]
+
+>>> MASH NOW: two hands, 8 lamp buttons, as fast as you can. <<<
+>>> Test runs 300s. A wedge alarm will beep + banner.   <<<
+>>> Ctrl+C ends the run early and still prints the summary.   <<<
+
+[  33.04s] progress: presses=156 (156/min) analog=521 maxGap=0.22s wedges=0
+[  63.12s] progress: presses=299 (284/min) analog=998 maxGap=0.22s wedges=0
+[  93.15s] progress: presses=396 (255/min) analog=1478 maxGap=0.22s wedges=0
+[ 123.17s] progress: presses=505 (246/min) analog=1954 maxGap=0.22s wedges=0
+[ 132.35s] operator stop (Ctrl+C) — ending run, snapshotting counters...
+
+== MASH SUMMARY [patched-62500] ==
+run              : 132s on COM1, lamps on, wedge threshold 2.0s
+firmware         : 4.2
+presses/releases : 530 / 530  (238 presses/min)
+analog replies   : 2115 (~2400 poll slots; 88.1%), sentinels 0
+framing / NAK    : 24 / 69
+gap histogram    : <100ms:2103  100-250ms:11  250-500ms:0  0.5-1s:0  1-2s:0  2-5s:0  >5s:0
+longest gaps     : 0.25s@132.3s, 0.22s@1.8s, 0.14s@54.9s, 0.14s@117.2s, 0.13s@43.6s, 0.13s@14.8s, 0.13s@43.1s, 0.13s@115.0s, 0.12s@124.0s, 0.12s@94.9s
+wedge events     : 0
+counters before  : RestartCount=0  AbandonCount=0  FullBufferCount=0  [board/lamp faults reported!]
+counters after   : RestartCount=0  AbandonCount=0  FullBufferCount=0  [board/lamp faults reported!]
+counter delta    : RestartCount=+0  AbandonCount=+0  FullBufferCount=+0
+verdict          : PASS — no wedge; compare gap histogram + counter delta to baseline
+log file         : C:\VWE\riojoy\riomash-patched-62500-20260718-130333.log