lf t55xx writetest: read each block three times and name how it differs

One read cannot see the fault this suite is most useful for. A block that comes
back silently wrong on some reads and right on others scores the same as one
that is simply wrong, and the same as one that is fine. Read each block three
times and mark disagreement between them as `unstable`.

Where a block is wrong the same way every time, name it: `rol1`, `ror5`,
`inverted`, `inv+ror2`, `shr1`. The kind matters. A rotation says the word
boundary moved, an inversion says psk picked the opposite phase, and a shift
says a demodulation opened a bit early and padded with zero - three different
faults that want three different fixes, and all of them previously reported as
"block 3,4 bad".

Measured on a T5577 over the 21 PSK1 configurations, this turns 56 fields of
"bad" into 56 marked unstable, and on a build that has since fixed part of it,
into 24 rotations whose size tracks the bit rate - ror5 at RF/32, ror4 at RF/40,
ror2 at RF/100 - which is a fixed sample offset rather than noise, and says
where to look.

Block 1 is 00000000, which equals all of its own rotations, so a rotated read of
it cannot be told from a correct one. Noted in the header rather than changed,
since the payload is the suite's own choice.

Also corrects the header, which described block 2 as ffffffff where the script
writes aa5500ff.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Matthew CarrollandClaude Opus 5 committed 2026-09-10 12:30:20 -07:00
1 parent 0884083530
commit bdc4828e0f
2 files changed
+87 -10

No files matched your search

+1
View File
@@ -3,6 +3,7 @@ All notable changes to this project will be documented in this file.
This project uses the changelog in accordance with [keepchangelog](http://keepachangelog.com/). Please use this to write notable changes, which is not the same as git commit log... This project uses the changelog in accordance with [keepchangelog](http://keepachangelog.com/). Please use this to write notable changes, which is not the same as git commit log...
## [unreleased][unreleased] ## [unreleased][unreleased]
- Changed `tests/lf_t55xx_writetest.lua` - reads each block three times and names how a bad read differs (`rol1`, `ror5`, `inverted`, `shr1`) instead of only pass/fail (@mfcarroll)
- Fixed `lf t55xx detect` - PSK1 now detects at every bit rate and subcarrier. (@iceman1001) - Fixed `lf t55xx detect` - PSK1 now detects at every bit rate and subcarrier. (@iceman1001)
- Fixed `lf t55xx detect` - FSK was skipped entirely when the field-clock pair measured as neither legal pair, losing FSK1 at RF/32 and RF/40 and every variant at RF/16 (@iceman1001) - Fixed `lf t55xx detect` - FSK was skipped entirely when the field-clock pair measured as neither legal pair, losing FSK1 at RF/32 and RF/40 and every variant at RF/16 (@iceman1001)
- Fixed `lf t55xx dump/read` - blocks were extracted at a bit offset cached from the last `detect` The offset is now anchored in graph samples (@iceman1001) - Fixed `lf t55xx dump/read` - blocks were extracted at a bit offset cached from the last `detect` The offset is now anchored in graph samples (@iceman1001)
+86 -10
View File
@@ -16,7 +16,7 @@ A strict T55x7 test suite:
lf t55xx wipe lf t55xx wipe
lf t55xx write -b 1 -d 00000000 lf t55xx write -b 1 -d 00000000
lf t55xx write -b 2 -d ffffffff lf t55xx write -b 2 -d aa5500ff
lf t55xx write -b 3 -d 80000000 lf t55xx write -b 3 -d 80000000
lf t55xx write -b 4 -d 00000001 lf t55xx write -b 4 -d 00000001
@@ -24,6 +24,14 @@ Each row names the configuration word, what it means, what `lf t55xx detect`
made of it and how many of the four data blocks read back intact. Markers are made of it and how many of the four data blocks read back intact. Markers are
also written into the session log ( `rem [ERR:...]` and `rem [SUMMARY:...]` ) also written into the session log ( `rem [ERR:...]` and `rem [SUMMARY:...]` )
Every block is read three times. A block that reads back differently between
those is marked `unstable`, which is the fault a single read cannot see: the
value is silently wrong rather than missing. A block that is wrong the same way
each time is named where it can be - `rol1`, `ror5`, `inverted`, `shr1` - because
a rotation points at the word boundary and an inversion at psk phase, and the two
want different fixes. Block 1 is `00000000`, which is equal to all of its own
rotations, so a rotated read of it cannot be told from a correct one.
Note: that the card is wiped before each modulation. Note: that the card is wiped before each modulation.
Needs a T5577 in the field. Needs a T5577 in the field.
@@ -161,30 +169,98 @@ local function prepare()
return true return true
end end
-- a block is read this many times, so that a value which is silently wrong on
-- only some reads is not mistaken for one that is right
local READS = 3
local function rol32(v, n)
n = n % 32
if n == 0 then return v end
return ((v << n) | (v >> (32 - n))) & 0xFFFFFFFF
end
--- ---
-- Read blocks 1-4 back. Returns how many matched and a short note naming the -- Name how `got` differs from `want`, when it can be named. A rotation says the
-- ones that did not. -- word boundary moved, an inversion says psk picked the opposite phase, and a
-- shift says a demodulation opened a bit early and padded with zero - different
-- faults wanting different fixes, so worth telling apart. A word equal to all of
-- its own rotations carries no boundary information and only gets 'differs'.
local function classify(want, got)
if got == want then return 'ok' end
if rol32(want, 1) == want then return 'differs' end
for n = 1, 31 do
if rol32(want, n) == got then
return (n <= 16) and ('rol' .. n) or ('ror' .. (32 - n))
end
end
if (want ~ 0xFFFFFFFF) == got then return 'inverted' end
for n = 1, 31 do
if (rol32(want, n) ~ 0xFFFFFFFF) == got then
return 'inv+' .. ((n <= 16) and ('rol' .. n) or ('ror' .. (32 - n)))
end
end
for n = 1, 8 do
if ((want >> n) & 0xFFFFFFFF) == got then return 'shr' .. n end
if ((want << n) & 0xFFFFFFFF) == got then return 'shl' .. n end
end
return 'differs'
end
---
-- Read blocks 1-4 back, READS times each. Returns how many matched on every
-- read and a short note naming what went wrong with the ones that did not.
local function read_back(config) local function read_back(config)
local good, bad = 0, {} local good, bad = 0, {}
for block, want in ipairs(DATA_BLOCKS) do for block, want in ipairs(DATA_BLOCKS) do
local data = core.t55xx_readblock(block, '0', '0', '') local wantv = tonumber(want, 16)
local got = data and ('%08X'):format(data) or '' local seen, first, stable = {}, nil, true
if got:lower() == want:lower() then for _ = 1, READS do
local data = core.t55xx_readblock(block, '0', '0', '')
local v = data and (data & 0xFFFFFFFF) or nil
table.insert(seen, v)
if first == nil then
first = v
elseif v ~= first then
stable = false
end
end
if stable and first == wantv then
good = good + 1 good = good + 1
else else
table.insert(bad, tostring(block)) local what
core.console(format('rem [ERR:READ:%s:%d] block %d: read %s instead of %s', if stable == false then
config, block, block, got, want), false, true) what = 'unstable'
else
what = classify(wantv, first)
end
table.insert(bad, format('b%d %s', block, what))
local reads = {}
for _, v in ipairs(seen) do
table.insert(reads, v and ('%08X'):format(v) or '----')
end
core.console(format('rem [ERR:READ:%s:%d] block %d: %s, wanted %s, read %s',
config, block, block, what, want,
table.concat(reads, ' ')), false, true)
end end
end end
local note = '' local note = ''
if #bad > 0 then if #bad > 0 then
note = 'block ' .. table.concat(bad, ',') .. ' bad' note = table.concat(bad, ' ')
end end
return good, note return good, note
end end