Separate the two fuel circuits
The AXX 2.0 FSI Turbo example in Volkswagen SSP 337, pages 17–21 has direct injectors, a low-pressure electric supply and an engine-driven high-pressure pump. Its system overview identifies low-pressure sender G410, rail-pressure sender G247, pump controller J538 and pressure-control valve N276. Low supply pressure and inadequate rail pressure are different symptoms, even though a supply problem can affect the high-pressure side. Confirm the components fitted to the vehicle being tested.
The AXX training material describes homogeneous running and two injection events during its cold-start catalyst-heating phase. That is a direct-injection operating strategy; it does not establish a separate port-injector system. Keep temperature and operating state with lambda and injector data.
A measured-block example
Ross-Tech's 2.0 TFSI AXX/BGB/BPJ/BPY/BWA fuel-pressure procedure gives the following VCDS measuring-block examples. Their stated prerequisites include engine idling, coolant at least 80 °C and system voltage at least 11 V. Verify the engine, ECU, channel labels and units in the actual diagnostic tool; these are examples, not universal channel assignments.
| Measuring block | Fields in the Ross-Tech procedure | Use |
|---|---|---|
| 103 | Field 1: low fuel pressure. | Establish whether the supply side is behaving plausibly. Ross-Tech notes that this group does not apply to the later chain-driven examples it identifies without G410. |
| 230 | Field 1: specified high rail pressure; field 2: actual high rail pressure; field 3: deviation. | Compare request and actual in the documented procedure and under the conditions supported by the equipment. |
The limits and conditions in the Ross-Tech test belong to that procedure; do not turn its idle or speed-check values into universal full-load tuning targets. When logging, record low pressure, requested and actual rail pressure, lambda request and actual, engine speed, load, temperature and any relevant fault or intervention state. Sampling speed and channel availability can limit what a log shows.
When rail pressure falls short
- Keep the original faults and freeze-frame data. If low-pressure supply faults are present, diagnose the supply, filter, lines and pump control using the appropriate service information before assuming the high-pressure pump is defective.
- Check whether the pressure shortfall is persistent, load-dependent or accompanied by a sensor, regulator or supply fault; compare lambda and injection data. Do not correct a mechanical fuel-delivery shortfall by simply changing the pressure request.
- On the particular BPY vehicles covered by Volkswagen bulletin V151401 / 2015153—2006–07 Eos, GTI and Passat, and 2005–07 Jetta A5—the manufacturer discusses cam-lobe and pump-follower wear associated with specific rail-pressure faults. The bulletin directs low-pressure faults to be diagnosed first and describes inspection under its service procedure. It is not evidence that every MED9 engine has the same failure or an inspection interval.
Use current vehicle-specific workshop instructions for fuel-system work. For the relationship between torque requests, boost and ignition, see MED9.1 torque, boost and ignition tuning. The cited VW training and service documents are hosted by independent sites and NHTSA; the guidance above has not been vehicle-tested by this wiki.
Upgraded HPFP: separate geometry from pressure request
Record the fitted pump and internal-kit part numbers before selecting calibration data. APR's EA113 HPFP information explicitly calls for pump-specific software to maximise its performance and notes that worn cam/follower hardware can prevent full delivery. A claimed flow increase is not evidence for a particular rail-pressure target.
In the Bosch MED9.1 function document for 41W038_PQ35, 20 August 2004, AMSV 8.70.0, pages 1686–1689, RQUAHDPK means piston radius squared, not piston area. SKHDPMX limits compression stroke; HDRPSOL 10.60.0, page 1724 defines VHDP as swept volume.
For a cylindrical displacement calculation, radius squared is (diameter / 2)2, area is π × radius squared, and swept volume is area × stroke. Keep mm, mm2 and mm3 distinct. For example, Autotech specifies a 9.8 mm piston for kit 10-127-100K; that diameter gives 24.01 mm2 for radius squared. This dimensional example does not establish the fitted stroke or a complete software calibration.
The December 2013–January 2014 HPFP discussion and February–May 2014 follow-ups discuss radius/volume changes, differing stroke assumptions and apparently inconsistent values in other engine files. Keep those reports with their hardware and software context; do not copy a two-scalar recipe across engines.
Check how the rail-pressure request is formed
Bosch HDRPSOL, pages 1723–1725, distinguishes mode-selected map output from the processed rail-pressure request and includes gradient limitation. Its KFPRSOLHOM example uses rpm and indicated driver-request torque, not a generic measured-load axis. With verified logging, compare prsollkf_w, prsoll_w and prist_w: an upstream request that is being limited is a different investigation from actual pressure failing to follow the final request. Names, addresses, units and availability remain software-specific.
LPFP upgrades: read the later outcome
The opening 2013 LPFP discussion contains conflicting experience: RaraK reports controller overheating under sustained demand, while hammersword reports using a stock controller with recalibration. The thread does not establish one compatible pump/controller/calibration combination for all MED9.1 vehicles.
In a later case, notoriousg first reported controller overheating and then on 28 March 2026 found a displaced blanking plug in the tank after repeated calibration attempts. The 4 May 2026 follow-up revises the earlier settings after regulation faults persisted. These are the author's reported findings; the thread does not provide sufficient complete software identity and validation to publish a transferable conversion. Check assembly, supply faults and the later outcome before attributing a pressure problem to maps.
For an upgraded system, document pump/controller identity, voltage, temperatures and duration alongside low-pressure supply, rail request/actual, lambda and faults. Include starts, idle and part load as well as controlled load testing. A short successful pull cannot establish sustained controller behaviour. Keep regulation diagnostics active and investigate unexpected pressure excursions before adding load. Additional sources accessed 16 November 2026.
For injector flow correction, pressure-dependent offset, short pulses and event phasing after the pressure checks, see MED9.1 injector calibration and injection windows.