From 6943838652aeaba8b61b16f942800de0c5b7586e Mon Sep 17 00:00:00 2001 From: mrsqr Date: Fri, 21 Aug 2026 19:42:11 +0100 Subject: [PATCH] docs(bom): verify the 25 Hz claim against the data sheet Every alternative examined traded update rate against the number of constellations, and the M9N's own headline figure had been taken on trust from this document rather than checked. Applying that scrutiny only to the alternatives would have been careless. Checked, it holds without qualification. NEO-M9N-00B_DataSheet_UBX-19014285 gives the maximum PVT update rate per constellation configuration as 25 Hz across all five, including GPS + GLONASS + Galileo + BeiDou. The rate does not fall as constellations are added, which is what this project needs and what makes the part unusually well suited to it. Recorded alongside it that the newer u-blox parts are a downgrade for this application, because that is the substitution most likely to be proposed again. The M10 platform is optimised for low power and manages 10 Hz with four constellations, reaching 20 Hz only with three and a raised CPU clock. This device runs from vehicle USB, so the power saving buys nothing and the rate ceiling costs everything. Also recorded why two boards that are easier to buy are not substitutes. The MicroMod NEO-M9N is the same receiver in the wrong packaging: an M.2 edge connector that needs a carrier board it does not include, and a U.FL antenna connector that reintroduces the fragile joint selecting the SMA variant removed, in a vibrating car. The MAX-M10S breakout has the right connectors and half the required update rate. Sourcing is noted as it stands: Pimoroni no longer stock the part and The Pi Hut do not carry it, leaving DigiKey UK or SparkFun direct. Co-Authored-By: Claude Opus 5 (1M context) --- docs/HARDWARE_BOM.md | 55 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+) diff --git a/docs/HARDWARE_BOM.md b/docs/HARDWARE_BOM.md index 280ef51..254b228 100644 --- a/docs/HARDWARE_BOM.md +++ b/docs/HARDWARE_BOM.md @@ -100,6 +100,42 @@ Key characteristics relevant to this project: - receiver timepulse capability - onboard backup battery for warm/hot-start assistance +### The 25 Hz figure, verified against the data sheet + +Checked 2026-08-21 rather than taken on trust, because every alternative examined turned out +to trade update rate against the number of constellations, and it would have been careless to +apply that scrutiny only to the alternatives. + +`NEO-M9N-00B_DataSheet_UBX-19014285` gives the maximum navigation update rate per constellation +configuration: + +| GNSS configuration | Max PVT update rate | +|---|---:| +| **GPS + GLONASS + Galileo + BeiDou** | **25 Hz** | +| GPS + GLONASS + Galileo | 25 Hz | +| GPS + GLONASS | 25 Hz | +| GPS + BeiDou | 25 Hz | +| GPS + Galileo | 25 Hz | + +The rate does not fall as constellations are added. **25 Hz is available with all four +concurrently**, which is what this project needs and what makes the M9N unusually well suited +to it. + +### Why the newer u-blox parts are a downgrade here + +This is the substitution most likely to be proposed again, so it is recorded with figures. + +| Part | Four constellations | Best case | +|---|---:|---| +| **NEO-M9N** (M9) | **25 Hz** | 25 Hz | +| MAX-M10S (M10) | 10 Hz | 20 Hz, and only with three constellations and a raised CPU clock | +| ZED-F9P (F9) | ~9 Hz indicated | 25 Hz, and the table indicates that is a reduced configuration | + +The M10 platform is newer and worse for this application. It is optimised for low power, and +its default clock rate supports only 10 Hz with four constellations; higher rates require +reconfiguring the clock and dropping a constellation. This device runs from vehicle USB, so +the power saving buys nothing and the rate ceiling costs everything. + The firmware acceptance requirement remains **20 Hz minimum**. Development should establish reliable 20 Hz operation first, then evaluate 25 Hz with the display, SD logging and all normal tasks running concurrently. ## D. GNSS antenna @@ -154,6 +190,25 @@ The AA.170 covers all three bands explicitly. | SparkFun `GPS-14986` | 1575-1610 | ~£18 | Rejected: misses BeiDou | | SparkFun `GPS-15192` / u-blox ANN-MB-00 | 1559-1606 and 1197-1249 | £98.84 | Rejected: L2 band unusable on an L1 receiver, four times the price | +### Receiver packaging: why the MicroMod and MAX-M10S boards are not substitutes + +Both are stocked by The Pi Hut and are easier to buy than the selected part. Neither works here. + +**SparkFun MicroMod GNSS Function Board, NEO-M9N, £41.30.** Same receiver, wrong packaging. It +terminates in an **M.2 edge connector** and its listing states it does not include a MicroMod +Main Board, so it cannot reach the Waveshare board without buying a carrier whose purpose is to +host a MicroMod processor. Its antenna connector is **U.FL**, which reintroduces exactly what +selecting the SMA variant removed: a fragile joint, rated for few mating cycles, in a +vibrating car, plus a U.FL-to-SMA pigtail to reach the selected antenna. With a carrier and a +pigtail it costs about the same as `GPS-17285` and adds two failure points. + +**SparkFun MAX-M10S breakout.** Has the SMA connector and Qwiic, and fails on rate: **10 Hz +with four constellations**, half the minimum. See the table above. + +**Availability.** Pimoroni no longer stock `GPS-17285`, and The Pi Hut do not carry it. DigiKey +UK remains the source, with SparkFun direct as the alternative. + + ### Alternatives considered and rejected, 2026-08-21 Re-examined against the possibility of better reception. The M9N was retained. Recorded here