From e2cb7bd2d6e3da2e31f8495fc1b107fe3e118a03 Mon Sep 17 00:00:00 2001
From: Jason <83615043+JJassonn69@users.noreply.github.com>
Date: Tue, 5 May 2026 12:12:04 +0545
Subject: [PATCH] =?UTF-8?q?fix(fpga):=20PR-X.1=20F-7.6=20=E2=80=94=20corre?=
 =?UTF-8?q?ct=20offset-binary=20mid-scale=20conversion=20in=20DDC?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Stage-7 ADC chain audit. ddc_400m.v line 273-274 converted AD9484
offset-binary 8-bit codes to MIXER_WIDTH-bit signed via
`(adc << 9) - (8'hFF << 8)`. The right operand evaluates to
`0xFF / 2 = 0x7F` (integer divide drops 0.5), so the formula
subtracted 65280 instead of the correct mid-scale offset 65536:
  - code 0x80 (0V analog, mid-scale) → +256 (expected 0)
  - code 0x00 (-fs)                  → -65280 (expected -65536)
  - code 0xFF (+fs - 1 LSB)          → +65280 (expected +65024)

A 0.5-LSB DC bias on the AD9484 stream is functionally invisible
in production: after the 120 MHz NCO mixer it lands at ±120 MHz,
which the CIC + FIR LPF (~50 MHz rolloff) rejects. But the offbin
and twoc paths disagreed on the same analog input, and any reuse
of `adc_signed_offbin` outside the mixer chain would carry the
bias.

Replace with an MSB-flip + sign-extend + 9-zero pad that mirrors
the twoc branch: `{~msb, ~msb, low7, 9'b0}`. Mid-scale code 0x80
now produces 0 exactly, matching the twoc path bit-for-bit.

Local FPGA regression 36/1/6 unchanged (1 FAIL = pre-existing
T-6 drift, deferred PR-M.4).
---
 9_Firmware/9_2_FPGA/ddc_400m.v | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/9_Firmware/9_2_FPGA/ddc_400m.v b/9_Firmware/9_2_FPGA/ddc_400m.v
index 1abbf5d..3793310 100644
--- a/9_Firmware/9_2_FPGA/ddc_400m.v
+++ b/9_Firmware/9_2_FPGA/ddc_400m.v
@@ -270,8 +270,18 @@ end
 
 wire signed [MIXER_WIDTH-1:0] adc_signed_offbin;
 wire signed [MIXER_WIDTH-1:0] adc_signed_twoc;
-assign adc_signed_offbin = {1'b0, adc_data, {(MIXER_WIDTH-ADC_WIDTH-1){1'b0}}} -
-                           {1'b0, {ADC_WIDTH{1'b1}}, {(MIXER_WIDTH-ADC_WIDTH-1){1'b0}}} / 2;
+// Audit F-7.6: AD9484 offset-binary → signed two's-complement is just an
+// MSB flip. Place the converted 8-bit signed value into the top of the
+// MIXER_WIDTH-bit field with one leading sign-extension copy and the
+// usual 9-zero LSB pad — exactly mirrors the twoc branch below. The
+// previous formula `(adc<<9) - (8'hFF<<8)` used `0xFF/2 = 0x7F` (integer
+// divide drops 0.5 LSB), which produced +256 at mid-scale code 0x80
+// instead of 0 and made the offbin and twoc paths disagree on the same
+// analog input. Functionally invisible in production (120 MHz IF mix +
+// LPF rejects the 0.5-LSB DC tone) but mathematically wrong.
+assign adc_signed_offbin = {~adc_data[ADC_WIDTH-1], ~adc_data[ADC_WIDTH-1],
+                            adc_data[ADC_WIDTH-2:0],
+                            {(MIXER_WIDTH-ADC_WIDTH-1){1'b0}}};
 assign adc_signed_twoc   = {adc_data[ADC_WIDTH-1], adc_data, {(MIXER_WIDTH-ADC_WIDTH-1){1'b0}}};
 
 // Combinational ADC sign conversion (no register — DSP48E1 AREG handles it)