diff --git a/docs/developer/newport_sastt_evaluation.md b/docs/developer/newport_sastt_evaluation.md
new file mode 100644
index 00000000..dcda931e
--- /dev/null
+++ b/docs/developer/newport_sastt_evaluation.md
@@ -0,0 +1,59 @@
+# Newport XPS-RLD + ILS300LM-S for SAXS-TT fly scans: evaluation
+
+Status 2026-11-09. Question: is the Newport cart (XPS-RLD4, XPS-DRV02, XPS-DRV11, M-ILS300LM-S, M-VP-5ZA,
+XPS-DK22/DK26) a good replacement for the Aerotech A3200 + ALS25020 for continuous X scans? **Verdict: yes,
+for light payloads.** It is a better acquisition system; it is not a replacement for the heavy Aerotech stack.
+
+## Controller: XPS-RLD vs A3200
+
+| Need | XPS-RLD (10 kHz servo) | A3200 today |
+|---|---|---|
+| Position per trigger | Hardware latch on sync input 1, < 50 ns, max. 1 trigger per servo cycle (~10 kHz), 1 M entries | Start position only, inside a capture window |
+| Continuous stream | Rolling gathering at 10 kHz (time, position, following error, `GPIO1.DI` gate bits), read by BEC during the scan | not available |
+| Trigger out | Position compare (PCO) 5 ns, < 700 ns latency, up to 5 MHz, distance or time spaced (axes 1-2 only) | PSO |
+| Rising + falling edge | One input: DDG/PandA sends a pulse at each gate edge (2 kHz at 1 kHz exposures) | - |
+
+Mean/std per exposure: computed in BEC from the 10 kHz stream (DI edge known to ~100 µs, i.e. v × 100 µs) or
+exact edges from the hardware latch. **Kilohertz dwell: yes**, limit ~10 kHz latching, 5 MHz PCO.
+
+## Stage: ILS300LM-S vs ALS25020
+
+| | ILS300LM-S | ALS25020 |
+|---|---|---|
+| Accuracy typ. (guar.) / bidir. repeatability | ±1.2 (±2.5) µm / ±0.12 (±0.30) µm | ±8 µm (LT std.) / ±0.5 µm |
+| Pitch, yaw typ. (guar.) | ±40 (±150), ±25 (±75) µrad | 8 arcsec (~39 µrad) |
+| Force cont./peak, load | 15 / 30 N, 250 N | 109-168 / 673 N, 70 kg |
+| Stiffness | roll 15, pitch 10, yaw 10 µrad/Nm | not published |
+| Encoder | incremental, 20 µm 1 Vpp, home mark at mid travel, commutation init at start | sine, incremental |
+
+## Payload setups (horizontal ILS)
+
+- **A:** VP-5ZA (2.4 kg, 4.8 mm, ±0.25 µm, encoder) + SmarGon (~1 kg, 50 g payload): ~4.5 kg moving,
+  ~3.3 m/s² continuous (today 0.5 m/s²). Preferred over OWIS HVM stepper stages (< 6 µm repeatability).
+- **B:** Steinmeyer PMT160-SM-200/300 (5.2/6.5 kg, stepper, no encoder, ±2.4 µm) on a right-angle bracket:
+  ~7-8.5 kg, ~1.6-1.9 m/s². Load OK; the issue is sample height: pitch × height gives ±6 µm (typ) at 150 mm,
+  ±12 µm at 300 mm, plus bracket resonances. Keep the working range low, bracket stiff.
+
+## Resonances, tuning, metrology
+
+- SGamma jerk time T_j ≈ 1/f_res puts a spectral zero near the resonance (standard). Command-side notch
+  filters (`PositionerCompensationPreFeedForward...NotchFilterSet`) and dual loop are **[Extended] firmware**.
+  2 servo-loop notch filters are standard.
+- Per-setup tuning profiles (PID, notches, jerk time) in the BEC device config, applied via
+  `GroupMotionDisable` → set → `GroupMotionEnable` (encoder keeps running, no re-homing). Re-apply after each
+  init; `ScalingAcceleration` is stages.ini only.
+- Abbe error and residual vibration: interferometer/capacitive sensor at the sample, captured by the PandA on
+  the same gates (value, mean, min/max per gate). XPS gathering stays as diagnostic stream.
+
+## Open questions for Newport
+
+1. Does our XPS-RLD4 include the Extended firmware (notch compensation, dual loop)? 2. Which edge latches on
+sync input 1? 3. DRV11 support for OWIS/Steinmeyer steppers? 4. Side mounting of ILS (if reconsidered).
+
+## Vendor documents
+
+Not in git (copyright, ~54 MB): `~/Downloads/newport_xps_cart/` on the workstation of M. Holler. ILS-LM-S and
+VP-5ZA data sheets and user manuals, XPS-RLD datasheet and Start-Up Manual (EDH0420), XPS-D Features (EDH0407),
+Configuration (EDH0406) and User Interface (EDH0405) manuals, XPS Unified Programmer's Manual, PSO multi-axis
+guide, trigger synchronization tech note, driver cards data sheet, cable kits manual (EDH0404). Source:
+[newport.com/p/XPS-RLD4](https://www.newport.com/p/XPS-RLD4) and the product pages of each item.
diff --git a/docs/developer/newport_sastt_evaluation.pdf b/docs/developer/newport_sastt_evaluation.pdf
new file mode 100644
index 00000000..eec9216f
Binary files /dev/null and b/docs/developer/newport_sastt_evaluation.pdf differ
diff --git a/zensical.toml b/zensical.toml
index 1d80efc8..f22597ab 100644
--- a/zensical.toml
+++ b/zensical.toml
@@ -50,6 +50,7 @@ nav = [
     { "PandaBox free-running setup" = "developer/panda_box_free_running_setup.md" },
     { "PandaBox data schema" = "developer/panda_box_free_running_data_schema.md" },
     { "Aerotech stages: design and triggering" = "developer/aerotech_stages.md" },
+    { "Newport XPS + ILS evaluation (SAXS-TT)" = "developer/newport_sastt_evaluation.md" },
   ] },
 ]