Tuning - Revision history

Nyet: Alpha-N moved to a separate page

2026-10-02T01:45:55Z

Alpha-N moved to a separate page

@@ Line 158: / Line 158: @@
 '''Note that KFLF's description as "lambda" is a misnomer. It is not lambda at all. It is an injector on time scaling table where larger numbers are longer injector time, smaller numbers are shorter injector time.'''
-Also, you should always compare msdk_w (alpha-n estimated air flow) vs mshfm_w (measured airflow). If they are different, the car will run poorly if the MAF fails or is unplugged. See the [[#Alpha-N calibration|alpha-n calibration]] section below for how to fix this.
+msdk_w is the alpha-n (throttle angle based) airflow estimate that the engine falls back to if the MAF fails or is unplugged. If it differs from mshfm_w (measured airflow), the car will run poorly without the MAF. Calibrating it is not needed for a normal tune; see [[Alpha-N]].
 Finally, periodically check long and short term fuel trims (LTFT/STFT) from VCDS group 32/33. They can also be logged directly via rkat/fra/fr/usvk:

Nyet: add me7logger connection mode guide

2026-10-02T01:36:01Z

add me7logger connection mode guide

← Older revision		Revision as of 01:36, 2 October 2026
Line 50:		Line 50:
	# [http://nyet.org/cars/ECUxPlot/ ECUxPlot] to graph logs. Also has a HP/TQ calculator and various other tools.		# [http://nyet.org/cars/ECUxPlot/ ECUxPlot] to graph logs. Also has a HP/TQ calculator and various other tools.
	# [https://github.com/nyetwurk/trim-heatmap Trim Heatmap tool] - Plot STFT heatmaps from your ME7Logs to assist tuning KFKHFM!		# [https://github.com/nyetwurk/trim-heatmap Trim Heatmap tool] - Plot STFT heatmaps from your ME7Logs to assist tuning KFKHFM!

			=== Connection modes ===
			ME7Logger's connection mode is the <code>Connect</code> line in the <code>[Communication]</code> section of the .ecu file. <code>SLOW-0xNN</code> is a 5-baud slow init to address <code>NN</code>; <code>FAST-0xNN</code> is a fast init. Each ECU answers slow init only on the addresses listed in a table in its firmware:

			{\| class="wikitable"
			! ECU !! Connect !! Notes
			\|-
			\| M-box, K-box, L-box and most other ME7.1 (8D0907551, 4B0907551, 4Z7907551, 4B0906018 except DA and DQ, 06A906032, 066906032E, 022906032, 4D0907558S, 4D0907559D, 8E0906018B, 8N0906018CJ) \|\| SLOW-0x11 \|\| 0x11 is KWP2000, with no engine speed check, so it connects with the engine running. 0x01 is KW1281 (the protocol VCDS uses).
			\|-
			\| ME7.5 1.8T (8E0909518), 4D1907558, 4B0906018DA, 4B0906018DQ, 4E0910559E, 006410010A0 \|\| SLOW-0x01 \|\| 0x01 is KWP2000. There is no 0x11 entry and no SLOW-0x00 sequence.
			\|}

			Fast init (a 25 ms wake-up pulse) reaches the same KWP2000 code as SLOW-0x11 on all of these ECUs, also with no engine speed check.

			SLOW-0x00 ("double slowinit", used with <code>DoubleDelay</code>) is for cars where the instrument cluster interferes with a normal connection. The logger opens a KW1281 session at 0x01, ends it, and inits 0x01 again within a time window; the ECU then answers with KWP2000. The ECU checks that engine speed (<code>nmot_w</code>) is zero at three steps of this sequence, so SLOW-0x00 only connects with the engine stopped: connect first, then start the engine. This applies to every ECU in the first row of the table, not only the K-box.

	=== RS4 K-box logging ===		=== RS4 K-box logging ===
	There are reports that the K-box will only accept an ME7Logger connection while the engine is stopped. ~~Its SLOW-0x00 connection sequence checks that engine speed (<code>nmot_w</code>)~~ is ~~zero at three points, so~~ SLOW-0x00 ~~only~~ ~~connects~~ ~~with the engine stopped~~. ~~Other~~ ~~boxes such as the M~~-box ~~have~~ the same ~~checks~~ and ~~are~~ ~~normally~~ ~~logged~~ ~~with~~ the ~~engine running using SLOW~~-~~0x11~~, so ~~try~~ SLOW-0x11 ~~on the K-box first~~.<ref>[http://nefariousmotorsports.com/forum/index.php?;topic=14119.0 Connect ME7Logger with engine running using SLOW-0x00]</ref> If SLOW-~~0x11~~ ~~does~~ ~~not~~ ~~work~~, these patches remove the checks ~~so that SLOW-0x00 connects with the engine running~~:		There are reports that the K-box will only accept an ME7Logger connection while the engine is stopped. That is the SLOW-0x00 behavior described above. The K-box firmware has the same address table and the same checks as the M-box, so use SLOW-0x11.<ref>[http://nefariousmotorsports.com/forum/index.php?;topic=14119.0 Connect ME7Logger with engine running using SLOW-0x00]</ref> To use SLOW-0x00 with the engine running, these patches remove the three checks:

	0x35FD2 - 3D 06 -> CC 00		0x35FD2 - 3D 06 -> CC 00

Nyet at 01:14, 2 October 2026

2026-10-02T01:14:57Z

Nyet at 01:08, 2 October 2026

2026-10-02T01:08:42Z

← Older revision		Revision as of 01:08, 2 October 2026
Line 365:		Line 365:
	* FKVA (0x1A4D5) - Constant conversion factor for consumption display<ref>[http://nefariousmotorsports.com/forum/index.php/topic,496.msg5693.html#msg5693 Nefmoto FKVA discussion]</ref>		* FKVA (0x1A4D5) - Constant conversion factor for consumption display<ref>[http://nefariousmotorsports.com/forum/index.php/topic,496.msg5693.html#msg5693 Nefmoto FKVA discussion]</ref>

			This is a multiplier for KVB, but only for the consumption counter sent to the cluster (<code>kvakbi_w</code>, which accumulates injection time × FKVA × KVB). It does not affect the pulse-width consumption output below.
	This is a multiplier for KVB.

	For reference, you may be able to derive the right value using this (although some find trial and error tweaking is much easier):		For reference, you may be able to derive the right value using this (although some find trial and error tweaking is much easier). This is the Bosch formula for the pulse-width consumption output (<code>tkvag_w</code>, in 0.0512 ms steps per main program loop), where Kd = 26 mm<sup>3</sup>/ms and TKVAR carries over the rounding remainder from the previous loop:

	<math>TKVAG = \frac{\sum_{t=0}^{TMAIN} (rk_w+ rkte_w) \times KRKTE [ms] ~~\times FKVA~~ \times KVB [cm^3/min] \times 1000}{Kd [mm^3/s] \times 0.0512 [ms] \times 60000}</math>		<math>TKVAG = \frac{\sum_{t=0}^{TMAIN} (rk_w+ rkte_w) \times KRKTE [ms] \times KVB [cm^3/min] \times 1000}{Kd [mm^3/ms] \times 0.0512 [ms] \times 60000} + TKVAR</math>

	=Boost=		=Boost=
Line 661:		Line 661:
	* LDRQ0DY - LDR PID Q0 in dynamic operation (proportional term) - likely leave this alone, log <code>ldptv</code>		* LDRQ0DY - LDR PID Q0 in dynamic operation (proportional term) - likely leave this alone, log <code>ldptv</code>
	* LDRQ0S - LDR PID Q0 in static operation (proportional term) - likely leave this alone		* LDRQ0S - LDR PID Q0 in static operation (proportional term) - likely leave this alone
			* LDRQ1DY - LDR PID Q1 in dynamic operation (integral term)
	* LDRQ1ST - LDR PID Q1 in static operation (integral term) - likely leave this alone, log <code>lditv_w</code>, <code>ldimx_w</code>		* LDRQ1ST - LDR PID Q1 in static operation (integral term) - likely leave this alone, log <code>lditv_w</code>, <code>ldimx_w</code>
	* KFLDRQ2 - LDR PID Q2 (differential term) - adjust this to compensate for overshoot when your boost ramp meets requested. If you have overshoot, try decreasing KFLDIMX first. If you have undershoot, try increasing KFLDIMX first. Only change Q2 as a last resort. Log <code>ldrdtv</code>		* KFLDRQ2 - LDR PID Q2 (differential term) - adjust this to compensate for overshoot when your boost ramp meets requested. If you have overshoot, try decreasing KFLDIMX first. If you have undershoot, try increasing KFLDIMX first. Only change Q2 as a last resort. Log <code>ldrdtv</code>

	All PID terms act on <code>lde</code> ("requested pressure" - "actual pressure"), so positive <code>lde</code> means underboost. Only KFLDRQ2 has <code>lde</code> as an axis (nmot × <code>lde</code>). LDRQ0DY and LDRQ1ST are curves over RPM, and LDRQ0S is a single constant.		All PID terms act on <code>lde</code> ("requested pressure" - "actual pressure"), so positive <code>lde</code> means underboost. Only KFLDRQ2 has <code>lde</code> as an axis (nmot × <code>lde</code>). LDRQ0DY, LDRQ1DY and LDRQ1ST are curves over RPM, and LDRQ0S is a single constant.

			<code>B_lddy</code> (dynamic operation) is set when <code>lde</code> rises above UMDYLDR (100 hPa on the M-box) and cleared when <code>lde</code> drops to 0 or below, i.e. when actual boost reaches requested. In dynamic operation the D gain is KFLDRQ2, the P gain is LDRQ0DY - KFLDRQ2 (not less than 0), and the I gain is LDRQ1DY (20 at all RPM on the M-box). In static operation the D gain is 0, the P gain is LDRQ0S (2.0), and the I gain is LDRQ1ST (0.3-0.5). Because the dynamic P gain is LDRQ0DY minus KFLDRQ2, raising KFLDRQ2 alone moves gain from P to D rather than adding D on top. The D term (<code>ldrdtv</code>) is KFLDRQ2 × the change in <code>lde</code> since the previous cycle.

	KFLDRL can also be used to get open-loop type behavior for operation past the [[MAP]] and requested boost limit by making the output duty cycle unresponsive (flat) to uncorrected duty cycle (from the PID) at various RPM/DC points. Again, if you do this, make sure to leave DSLOFS at the stock value. This way, requested boost will always be higher than measured boost, and you will stay in open loop control.		KFLDRL can also be used to get open-loop type behavior for operation past the [[MAP]] and requested boost limit by making the output duty cycle unresponsive (flat) to uncorrected duty cycle (from the PID) at various RPM/DC points. Again, if you do this, make sure to leave DSLOFS at the stock value. This way, requested boost will always be higher than measured boost, and you will stay in open loop control.
Line 825:		Line 828:
	* Ensure that <code>mibas</code> remains below <code>miszul</code> to avoid intervention (which you will see in <code>mizsolv</code>) by lowering KFMIOP in areas reachable by actual measured load. This mainly becomes a concern in the middle of KFMIOP.		* Ensure that <code>mibas</code> remains below <code>miszul</code> to avoid intervention (which you will see in <code>mizsolv</code>) by lowering KFMIOP in areas reachable by actual measured load. This mainly becomes a concern in the middle of KFMIOP.

	* <code>mibas</code> (from actual load) must not exceed the allowed torque limit (<code>miszul</code>) allowed for a given request from driver's pedal to avoid torque intervention. '''Simply raising KFMIZUFIL/KFMIZUOF to avoid level 1 torque intervention will only cause level 2 torque intervention, so it is critical to tune KFMIOP correctly.'''		* <code>mibas</code> (from actual load) must not exceed the allowed torque limit (<code>miszul</code>) allowed for a given request from driver's pedal to avoid torque intervention. '''Simply raising KFMIZUFIL/KFMIZUOF to avoid level 1 torque intervention will only cause level 2 torque intervention, so it is critical to tune KFMIOP correctly.''' Level 2 is the separate torque monitoring, with its own allowed-torque maps: KFMPED_UM (0x14169, pedal sensor voltage × RPM, the same values as KFMIZUFIL stock) plus the offset KFMDZOF_UM (0x14129, about 12% at most points where stock KFMIZUOF is 0). Bosch's documented reaction when level 2 trips is a safety fuel cut (SKA).

	To verify you have done things correctly, log <code>mifa</code> (or <code>misolv</code>), <code>mimax</code>, <code>mibas</code>, <code>miszul</code> and <code>mizsolv</code>.		To verify you have done things correctly, log <code>mifa</code> (or <code>misolv</code>), <code>mimax</code>, <code>mibas</code>, <code>miszul</code> and <code>mizsolv</code>.
Line 855:		Line 858:
	* SMK08MDSW (mkar_w axis) (0x1DF8E)		* SMK08MDSW (mkar_w axis) (0x1DF8E)
	* DMARMX (0x19B9D) - tmot axis @0x19B97 (maximum steady-state ARMD intervention; 0 everywhere stock)		* DMARMX (0x19B9D) - tmot axis @0x19B97 (maximum steady-state ARMD intervention; 0 everywhere stock)
			* KIFZGHG (0x1589A) - ARMD vehicle-model integrator gain per gear (<code>gangi</code>), used with a manual gearbox. Automatics use KIFZGAWG (0x1DF3C) when the torque converter clutch is locked, and a gain of 0 when it is not.
	* KIFZGHG

	'''It is also possible for a faulty or failing crank position sensor to cause RPM readings to fluctuate, which may result in (inappropriate) ARMD intervention.'''		'''It is also possible for a faulty or failing crank position sensor to cause RPM readings to fluctuate, which may result in (inappropriate) ARMD intervention.'''

Nyet: boost/pressure changes

2026-10-02T01:03:22Z

boost/pressure changes

← Older revision		Revision as of 01:03, 2 October 2026
Line 648:		Line 648:
	On a WOT pull, I is maxed out due to integrator windup (accumulated error) so it is simply following IMX.		On a WOT pull, I is maxed out due to integrator windup (accumulated error) so it is simply following IMX.

	* KFLDIMX - LDR I-Regulator limit. The x-axis input is relative requested pressure (<code>plsolr_w</code>) in hPa (same as mBar), which, in the M-box, is "requested absolute pressure (<code>plsol_w</code>)" - "~~ambient~~ pressure (<code>~~pu_w~~</code>)".		* KFLDIMX - LDR I-Regulator limit. The x-axis input is relative requested pressure (<code>plsolr_w</code>) in hPa (same as mBar), which, in the M-box, is "requested absolute pressure (<code>plsol_w</code>)" - "base boost pressure (<code>plgru_w</code>)". <code>plgru_w</code> = <code>pu_w</code> × KFVPLGU(nmot, <code>pssol</code>) × KFWPLGTA(nmot, IAT). Both maps are flat 1.0 stock, so <code>plsolr_w</code> is <code>plsol_w</code> - <code>pu_w</code> unless you change them.

	Along with KFTARX, there is another IAT correction that ME7.1 uses to allow the PID to add waste-gate duty cycle at elevated IATs. If you altered KFTARX you should compensate for those corrections here (but note that a completely flat KFTARX will still result in a requested boost that increases with IAT):		Along with KFTARX, there is another IAT correction that ME7.1 uses to allow the PID to add waste-gate duty cycle at elevated IATs. If you altered KFTARX you should compensate for those corrections here (but note that a completely flat KFTARX will still result in a requested boost that increases with IAT):
Line 664:		Line 664:
	* KFLDRQ2 - LDR PID Q2 (differential term) - adjust this to compensate for overshoot when your boost ramp meets requested. If you have overshoot, try decreasing KFLDIMX first. If you have undershoot, try increasing KFLDIMX first. Only change Q2 as a last resort. Log <code>ldrdtv</code>		* KFLDRQ2 - LDR PID Q2 (differential term) - adjust this to compensate for overshoot when your boost ramp meets requested. If you have overshoot, try decreasing KFLDIMX first. If you have undershoot, try increasing KFLDIMX first. Only change Q2 as a last resort. Log <code>ldrdtv</code>

	~~The x-axis pressure input to the~~ PID terms ~~are~~ ~~all~~ <code>lde</code> ("requested pressure" - "actual pressure"), so positive <code>lde</code> means underboost		All PID terms act on <code>lde</code> ("requested pressure" - "actual pressure"), so positive <code>lde</code> means underboost. Only KFLDRQ2 has <code>lde</code> as an axis (nmot × <code>lde</code>). LDRQ0DY and LDRQ1ST are curves over RPM, and LDRQ0S is a single constant.

	KFLDRL can also be used to get open-loop type behavior for operation past the [[MAP]] and requested boost limit by making the output duty cycle unresponsive (flat) to uncorrected duty cycle (from the PID) at various RPM/DC points. Again, if you do this, make sure to leave DSLOFS at the stock value. This way, requested boost will always be higher than measured boost, and you will stay in open loop control.		KFLDRL can also be used to get open-loop type behavior for operation past the [[MAP]] and requested boost limit by making the output duty cycle unresponsive (flat) to uncorrected duty cycle (from the PID) at various RPM/DC points. Again, if you do this, make sure to leave DSLOFS at the stock value. This way, requested boost will always be higher than measured boost, and you will stay in open loop control.
Line 1,386:		Line 1,386:
	You can also try		You can also try

			* FHOKH - set to FF FF. FHOKH is the minimum altitude factor for catalyst heating (stock 0.75): catalyst heating only starts if the altitude factor at engine start (<code>fhost_w</code>, from <code>fho_w</code>) is above it, so FF FF means it never starts.
	* FHOKH - set to FF FF

	=Additional Reading=		=Additional Reading=

Nyet: ZKLAMFAW elaboration

2026-10-02T00:58:08Z

ZKLAMFAW elaboration

← Older revision		Revision as of 00:58, 2 October 2026
Line 119:		Line 119:
	*MLOFS - MAF offset. For Bosch MAFs, it should be 200. For Hitachi, 0. Subtracted from the output of MLHFM.		*MLOFS - MAF offset. For Bosch MAFs, it should be 200. For Hitachi, 0. Subtracted from the output of MLHFM.

	Scale MLHFM (constant throughout scale vs stock!) to get your idle and partial short term fuel trims (STFT) near zero and your WOT [[AFR]] to roughly match your requested [[AFR]]. Another good ~~MLFHFM~~ sanity check is to log ps_w vs actual boost at WOT - ps_w should generally be slightly below actual boost. If it is over, your MAF is probably overscaled, and if you are running the stock MAP and not a [[Manifold air pressure:Rescale project \| 5120]]<ref>[http://nefariousmotorsports.com/forum/index.php?topic=3027.0 5120 hack]</ref> hack, you run the risk of maxing out ps_w, which is a bad thing (see below).		Scale MLHFM (constant throughout scale vs stock!) to get your idle and partial short term fuel trims (STFT) near zero and your WOT [[AFR]] to roughly match your requested [[AFR]]. Another good MLHFM sanity check is to log ps_w vs actual boost at WOT - ps_w should generally be slightly below actual boost. If it is over, your MAF is probably overscaled, and if you are running the stock MAP and not a [[Manifold air pressure:Rescale project \| 5120]]<ref>[http://nefariousmotorsports.com/forum/index.php?topic=3027.0 5120 hack]</ref> hack, you run the risk of maxing out ps_w, which is a bad thing (see below).

	If your MLOFS is non-zero (e.g. most Bosch MAF files), ''you cannot do a simple arithmetic scaling of MLHFM''; you must offset it down (subtract 200) by MLOFS, then scale it, then offset it back (add 200).<ref>[http://nefariousmotorsports.com/forum/index.php?topic=4839 MAF scaling when MLOFS!=0]</ref> Alternately, you can set MLOFS to zero and shift MLHFM down accordingly, after which you can scale MLHFM arithmetically as usual. Note that MLOFS takes effect after the CPU takes a moving window average of the output of MLHFM, so technically this operation does not yield a completely identical result as stock.		If your MLOFS is non-zero (e.g. most Bosch MAF files), ''you cannot do a simple arithmetic scaling of MLHFM''; you must offset it down (subtract 200) by MLOFS, then scale it, then offset it back (add 200).<ref>[http://nefariousmotorsports.com/forum/index.php?topic=4839 MAF scaling when MLOFS!=0]</ref> Alternately, you can set MLOFS to zero and shift MLHFM down accordingly, after which you can scale MLHFM arithmetically as usual. Note that MLOFS takes effect after the CPU takes a moving window average of the output of MLHFM, so technically this operation does not yield a completely identical result as stock.
Line 271:		Line 271:

	*LAMFA - requested lambda. Not very flexible for high load enrichment, uses requested torque instead of actual torque. Also, the stock LAMFA torque axis is (incorrectly) restricted to 0.5%-1.0% on 2.7t ECUs. You will have to change it to full range (multiply each value by 100 to result in 50%-100%) to use LAMFA, otherwise, the 1% column will be used at ALL requested torque points!<ref>[http://nefariousmotorsports.com/forum/index.php/topic,141.msg9068.html#msg9068 Mysterious LAMFA axis values]</ref>		*LAMFA - requested lambda. Not very flexible for high load enrichment, uses requested torque instead of actual torque. Also, the stock LAMFA torque axis is (incorrectly) restricted to 0.5%-1.0% on 2.7t ECUs. You will have to change it to full range (multiply each value by 100 to result in 50%-100%) to use LAMFA, otherwise, the 1% column will be used at ALL requested torque points!<ref>[http://nefariousmotorsports.com/forum/index.php/topic,141.msg9068.html#msg9068 Mysterious LAMFA axis values]</ref>
	*ZKLAMFAW~~/ZKWLAFWL~~ (0x1C3EE) - time constant ~~for~~ ~~driver~~ ~~requested~~ ~~lambda~~ ~~filtering~~. ~~This~~ ~~filter~~ ~~smooths~~ LAMFA. Unfortunately, the smoothing has the side effect of delaying the effects of LAMFA.<ref>[http://nefariousmotorsports.com/forum/index.php?topic=2637.0 Effect of ZKLAMFAW on LAMFA]</ref> You may want to bring in LAMFA fairly aggressively and/or ~~increase~~ ZKLAMFAW significantly to compensate.<ref>[http://nefariousmotorsports.com/forum/index.php?topic=2637.msg38067#msg38067 Effect of ZKLAMFAW]</ref>		*ZKLAMFAW (0x1C3EE, stock 0.5 s) - time constant of the low-pass filter on the enrichment target. The ECU first takes the richest of LAMFA, the knock enrichment (<code>lamfawkr_w</code>, below) and the warm-up enrichment (KFLAFWL), then filters the result, so ZKLAMFAW delays knock enrichment as well as LAMFA. The M-box XDF mislabels this constant as ZKWLAFWL and shows it as a percentage (stock 20%); the real ZKWLAFWL (0x18F6A) is the time constant for blending in the warm-up enrichment. Unfortunately, the smoothing has the side effect of delaying the effects of LAMFA.<ref>[http://nefariousmotorsports.com/forum/index.php?topic=2637.0 Effect of ZKLAMFAW on LAMFA]</ref> You may want to bring in LAMFA fairly aggressively and/or make the ZKLAMFAW filter significantly faster to compensate. A larger raw value (a larger percentage in the XDF) is a shorter time constant.<ref>[http://nefariousmotorsports.com/forum/index.php?topic=2637.msg38067#msg38067 Effect of ZKLAMFAW]</ref>
	*TLAFA - time delay before driver requested lambda active. Already set to 0 on many files. On others, you may have to reduce it to get LAMFA to take effect in a timely manner.		*TLAFA - time delay before driver requested lambda active. Already set to 0 on many files. On others, you may have to reduce it to get LAMFA to take effect in a timely manner.

			The result is <code>lamfaws_w</code>. The ECU combines it with the knock and warm-up enrichment, filters it (ZKLAMFAW), and snaps it to 1.0 when it is within DLAMFAW of 1.0 (stock 0). That result is <code>lamfa_w</code>.
	The result is "lamfa_w"

	===LAMFAWKR===		===LAMFAWKR===
Line 333:		Line 333:
	===Result===		===Result===

			The lowest (richest) value becomes your requested AFR. LAMFA and knock enrichment are combined first: <code>lamfa_w</code> = filtered min(LAMFA, <code>lamfawkr_w</code>, warm-up enrichment). The requested lambda is then min(<code>lamfa_w</code>, <code>lambts_w</code>).
	The lowest value of the three (lamfa_w, lamfawkr, lambts) become your requested AFR.

	Thus, you can use the combination of LAMFA, BTS, and KFLAMKR to tailor your AFR. In particular, pre-emptive rich fueling at low actual loads (but high requested load) may inhibit knock sooner than the KR approach. You can use BTS to do further enrichment when EGTs get dangerously high, unless you have BTS activate at very low actual loads (and/or calculated EGTs). The latter may be undesirable for fuel economy reasons; it is more fuel efficient to have BTS activate much later, let LAMFA handle pre-emptive, high load request (but low actual load) enrichment, and let KR handle real time knock inhibition.		Thus, you can use the combination of LAMFA, BTS, and KFLAMKR to tailor your AFR. In particular, pre-emptive rich fueling at low actual loads (but high requested load) may inhibit knock sooner than the KR approach. You can use BTS to do further enrichment when EGTs get dangerously high, unless you have BTS activate at very low actual loads (and/or calculated EGTs). The latter may be undesirable for fuel economy reasons; it is more fuel efficient to have BTS activate much later, let LAMFA handle pre-emptive, high load request (but low actual load) enrichment, and let KR handle real time knock inhibition.

Nyet: Add NWS explanation, formalize M-box L-box etc. usage

2026-10-02T00:52:27Z

Add NWS explanation, formalize M-box L-box etc. usage

Show changes

Nyet: fix KFZW2 wnse, add a few comments to KFURL and KFPRG

2026-10-02T00:45:00Z

fix KFZW2 wnse, add a few comments to KFURL and KFPRG

← Older revision		Revision as of 00:45, 2 October 2026
Line 176:		Line 176:
	\| 1 \|\| '''KFMSNWDK''' \|\| Changed throttle body or intake manifold \|\| Throttle body flow map (throttle angle × nmot → normalized mass flow, then scaled by the KLAF pressure-ratio curve). The single most important map for alpha-n accuracy.		\| 1 \|\| '''KFMSNWDK''' \|\| Changed throttle body or intake manifold \|\| Throttle body flow map (throttle angle × nmot → normalized mass flow, then scaled by the KLAF pressure-ratio curve). The single most important map for alpha-n accuracy.
	\|-		\|-
	\| 2 \|\| '''KFURL''' \|\| Changed cams, head work, port work \|\| Pressure-to-load slope (%/hPa, ~~cam~~ angle × nmot) — how much load each hPa of manifold pressure produces		\| 2 \|\| '''KFURL''' \|\| Changed cams, head work, port work \|\| Pressure-to-load slope (%/hPa, valve overlap angle × nmot; overlap comes from cam position) — how much load each hPa of manifold pressure produces
	\|-		\|-
	\| 2 \|\| '''KFPBRK''' \|\| Changed cams, head porting, compression ratio \|\| Combustion chamber pressure correction factor (nmot × load → factor)		\| 2 \|\| '''KFPBRK''' \|\| Changed cams, head porting, compression ratio \|\| Combustion chamber pressure correction factor (nmot × load → factor)
	\|-		\|-
	\| 2 \|\| '''KFPRG''' \|\| Changed cam timing, exhaust, turbo \|\| Residual gas partial pressure (~~cam~~ angle × nmot) — the manifold pressure at which cylinder filling = 0		\| 2 \|\| '''KFPRG''' \|\| Changed cam timing, exhaust, turbo \|\| Residual gas partial pressure (valve overlap angle × nmot) — the manifold pressure at which cylinder filling = 0
	\|-		\|-
	\| 3 \|\| '''Adaptation reset''' \|\| After ANY map changes \|\| Clear msndko_w (additive offset) and fkpvdk_w (multiplicative factor), then drive with MAF connected to re-learn		\| 3 \|\| '''Adaptation reset''' \|\| After ANY map changes \|\| Clear msndko_w (additive offset) and fkpvdk_w (multiplicative factor), then drive with MAF connected to re-learn
Line 747:		Line 747:

	ME7.1 has a two point variable cam timing system; there is a base timing map for each cam timing state based on KFNW:		ME7.1 has a two point variable cam timing system; there is a base timing map for each cam timing state based on KFNW:
	* KFZW - VVT via KFNW inactive (fnwue=0, wnwi/~~wnwise~~=0)		* KFZW - VVT via KFNW inactive (fnwue=0, wnwi/wnwse=0)
	* KFZW2 - VVT via KFNW active (fnwue=1, wnwi/~~wnwise~~=22)		* KFZW2 - VVT via KFNW active (fnwue=1, wnwi/wnwse=20)

			The cam weighting factor is <code>fnwue</code> = min(<code>wnwi</code>, WNWIMXT) / DWNWEMXT. On the M box both are 20°KW, and the target angle <code>wnwse</code> is WNWIMXT when the cam is switched on.

	While the cam is moving, the two maps are blended: base timing = <code>fnwue</code> × KFZW2 + (1 - <code>fnwue</code>) × KFZW.		While the cam is moving, the two maps are blended: base timing = <code>fnwue</code> × KFZW2 + (1 - <code>fnwue</code>) × KFZW.

Nyet at 00:37, 2 October 2026

2026-10-02T00:37:52Z

← Older revision		Revision as of 00:37, 2 October 2026
Line 49:		Line 49:
	# [http://nyet.org/cars/ECUxPlot/ ECUxPlot] to graph logs. Also has a HP/TQ calculator and various other tools.		# [http://nyet.org/cars/ECUxPlot/ ECUxPlot] to graph logs. Also has a HP/TQ calculator and various other tools.
	# [https://github.com/nyetwurk/trim-heatmap Trim Heatmap tool] - Plot STFT heatmaps from your ME7Logs to assist tuning KFKHFM!		# [https://github.com/nyetwurk/trim-heatmap Trim Heatmap tool] - Plot STFT heatmaps from your ME7Logs to assist tuning KFKHFM!

	=== RS4 K-box logging ===		=== RS4 K-box logging ===
			There are reports that the K box will only accept an ME7Logger connection while the engine is stopped. Its SLOW-0x00 connection sequence checks that engine speed (<code>nmot_w</code>) is zero at three points, so SLOW-0x00 only connects with the engine stopped. Other boxes such as the M box have the same checks and are normally logged with the engine running using SLOW-0x11, so try SLOW-0x11 on the K box first.<ref>[http://nefariousmotorsports.com/forum/index.php?;topic=14119.0 Connect ME7Logger with engine running using SLOW-0x00]</ref> If SLOW-0x11 does not work, these patches remove the checks so that SLOW-0x00 connects with the engine running:
	There is a bug in the K box file <ref>[http://nefariousmotorsports.com/forum/index.php?;topic=14119.0 Allow logging start before motor is running]</ref> which forces you to connect only once the car is running.

	A workaround is to patch these locations:

	0x35FD2 - 3D 06 -> CC 00		0x35FD2 - 3D 06 -> CC 00
	0x77DE4 - 3D 08 -> CC 00		0x77DE4 - 3D 08 -> CC 00
	0x78CB6 - 3D 03 -> CC 00		0x78CB6 - 3D 03 -> CC 00

			Each replaces a "skip if engine speed is not zero" branch with a NOP. The RS4 Q file uses the same offsets.

	=Terminology=		=Terminology=
Line 1,007:		Line 1,008:

	Define these as 16 bit (do not designate LoHi or LSB), output type will be Hex digits. '''Still under development - may not work properly'''		Define these as 16 bit (do not designate LoHi or LSB), output type will be Hex digits. '''Still under development - may not work properly'''

			Each location is the RAM address in an instruction that reads a pre-cat O2 sensor's raw voltage: 1A91 is bank 1 (<code>uusvk_w</code>) and 1891 is bank 2 (<code>uusvk2_w</code>). These are the only three reads of each sensor, and the ECU derives the sensor voltage used by lambda control (<code>usvk_w</code>, <code>usvk2_w</code>) from them. Pointing one bank's reads at the other bank's sensor makes both banks' lambda control run from the remaining sensor. The patch does not change the heater outputs or the post-cat sensors. The offsets are the same on the M and L boxes.

	===Delete B1===		===Delete B1===

Nyet: under/overboost clarification again

2026-10-01T23:47:45Z

under/overboost clarification again

← Older revision		Revision as of 23:47, 1 October 2026
Line 696:		Line 696:
	===Overboost===		===Overboost===
	<span id="Negative deviation (overboost)"></span>		<span id="Negative deviation (overboost)"></span>
			'''P1555 (17963 "Boost Pressure Control: Upper Limit Exceeded")'''

	If you don't get all of this just right, and your actual boost goes too far above requested boost, you may experience overboost [[throttle cut]], which is ME attempting to get boost back under control by temporarily closing the throttle plate. This happens when the predicted overboost exceeds DPUPS (300 hPa stock), and needs no DTC. If actual boost stays above requested by more than KFDLULS (250-400 hPa at normal boost levels) for TULV1 (0.5 s), you get P1555 (17963 "Boost Pressure Control: Upper Limit Exceeded" in VCDS, <code>E_ldo</code>). The car then goes into limp mode: <code>rlmax</code> is capped by LDORXN (stock 115% falling to 90% above 4000 RPM), and overboost (KFLDRXO) is disabled.		If you don't get all of this just right, and your actual boost goes too far above requested boost, you may experience overboost [[throttle cut]], which is ME attempting to get boost back under control by temporarily closing the throttle plate. This happens when the predicted overboost exceeds DPUPS (300 hPa stock), and needs no DTC. If actual boost stays above requested by more than KFDLULS (250-400 hPa at normal boost levels) for TULV1 (0.5 s), you get P1555 (17963 "Boost Pressure Control: Upper Limit Exceeded" in VCDS, <code>E_ldo</code>). The car then goes into limp mode: <code>rlmax</code> is capped by LDORXN (stock 115% falling to 90% above 4000 RPM), and overboost (KFLDRXO) is disabled.

Line 702:		Line 704:
	===Underboost===		===Underboost===
	<span id="Positive deviation (underboost)"></span>		<span id="Positive deviation (underboost)"></span>
			'''P1556 (17964 "Charge Pressure Control: Negative Deviation")'''

	Alternately, if your requested boost is far too high for a given load/rpm point, you may experience [http://nefariousmotorsports.com/forum/index.php/topic,871.0title,.html underboost] limp mode. This occurs if actual boost is too far under requested boost for too long. The result will be the P1556 Negative Deviation code (see [[#Boost control DTCs\|boost control DTCs]] below), and from then on out, WGDC restricted to 10%. Specifically, the ECU sets <code>E_ldra</code> when, near full load (RPM above NDLDRAPU and load above SDLDRL), with the PID I term maxed out and overboost inactive, <code>lde</code> stays above EDLDRP (100 hPa) for TDLDRA (1.5 s). It then switches to open-loop boost control, where duty comes from KFTVLDRE (10.2% everywhere in the stock map).		Alternately, if your requested boost is far too high for a given load/rpm point, you may experience [http://nefariousmotorsports.com/forum/index.php/topic,871.0title,.html underboost] limp mode. This occurs if actual boost is too far under requested boost for too long. The result will be the P1556 Negative Deviation code (see [[#Boost control DTCs\|boost control DTCs]] below), and from then on out, WGDC restricted to 10%. Specifically, the ECU sets <code>E_ldra</code> when, near full load (RPM above NDLDRAPU and load above SDLDRL), with the PID I term maxed out and overboost inactive, <code>lde</code> stays above EDLDRP (100 hPa) for TDLDRA (1.5 s). It then switches to open-loop boost control, where duty comes from KFTVLDRE (10.2% everywhere in the stock map).
	* If your MAF scaling is too aggressive, your load may be reading high, which might enable underboost diagnosis too early during a pull. Fix this by scaling back MLHFM		* If your MAF scaling is too aggressive, your load may be reading high, which might enable underboost diagnosis too early during a pull. Fix this by scaling back MLHFM
Line 711:		Line 715:
	* If none of the above helps, consider tweaking NDLDRAPU and SDLDRL (there is no SDLDRA in the M-box), or EDLDRP and TDLDRA.		* If none of the above helps, consider tweaking NDLDRAPU and SDLDRL (there is no SDLDRA in the M-box), or EDLDRP and TDLDRA.

	Most common (non-tuning) related causes include a broken n75, boost leaks, blown turbos, or wastegates that are stuck open. Boost leaks will often be accompanied by fuel trim issues, and are more likely to set P1557 (17965 "Charge Pressure Control: Positive Deviation" in VCDS) than P1556 (17964 "Charge Pressure Control: Negative Deviation").		Most common (non-tuning) related causes include a broken n75, boost leaks, blown turbos, or wastegates that are stuck open. Boost leaks will often be accompanied by fuel trim issues, and are more likely to set '''P1557 (17965 "Charge Pressure Control: Positive Deviation" in VCDS)''' than P1556 (17964 "Charge Pressure Control: Negative Deviation"). A leak between the turbos and the throttle is a form of underboost: air the MAF has measured escapes before the throttle, so the manifold pressure the ECU models from MAF airflow (<code>ps_w</code>) is too high for the boost pressure the sensor actually sees (<code>pvdkds_w</code>). '''P1557''' catches that mismatch, often before the boost controller runs out of headroom and sets P1556.

	'''If you are running K04s, you probably want to use the RS4 maps for all of these areas as a starting point, rather than the stock K03 maps.'''		'''If you are running K04s, you probably want to use the RS4 maps for all of these areas as a starting point, rather than the stock K03 maps.'''
Line 718:		Line 722:
	The boost DTC names are confusing for two reasons.		The boost DTC names are confusing for two reasons.

	First, ~~Bosch~~ ~~and~~ ~~VAG~~ ~~use~~ ~~opposite~~ ~~signs~~ ~~for~~ "deviation". Bosch's control deviation <code>lde</code> is requested minus actual boost, so in ~~Bosch~~ ~~terms~~ positive ~~deviation~~ is ~~underboost~~ ~~and negative~~ deviation is ~~overboost~~. VAG's ~~fault text~~ ~~uses the opposite sign:~~ P1556 is "Charge Pressure Control: Negative ~~Deviation" and P1557 is "Charge Pressure Control: Positive~~ Deviation".<ref>[http://wiki.ross-tech.com/index.php/17964/P1556/005462 Ross-Tech: 17964/P1556]</ref><ref>[http://wiki.ross-tech.com/index.php/17965/P1557/005463 Ross-Tech: 17965/P1557]</ref> So ~~the~~ ~~VAG~~ ~~text~~ ~~makes~~ ~~P1557~~ ~~sound~~ ~~like~~ ~~overboost~~.		First, underboost is called a positive deviation in Bosch documentation but a negative deviation in VAG fault text. Bosch's control deviation <code>lde</code> is requested minus actual boost, so underboost makes it positive; this is the "positive deviation" in Bosch and NefMoto discussions. VAG's underboost code, P1556, is "Charge Pressure Control: Negative Deviation".<ref>[http://wiki.ross-tech.com/index.php/17964/P1556/005462 Ross-Tech: 17964/P1556]</ref> P1557 is "Charge Pressure Control: Positive Deviation",<ref>[http://wiki.ross-tech.com/index.php/17965/P1557/005463 Ross-Tech: 17965/P1557]</ref> which next to P1556 reads as overboost, but as described below it is not a boost deviation check at all.

	Second, the Bosch underboost diagnosis (<code>E_ldra</code>) has two separate checks, and they report different codes. Each ME7 fault path has four DTCs, one each for a max, min, signal and plausibility error. <code>E_ldra</code> reports its min error as P1556 and its max error as P1557.		Second, the Bosch underboost diagnosis (<code>E_ldra</code>) has two separate checks, and they report different codes. Each ME7 fault path has four DTCs, one each for a max, min, signal and plausibility error. <code>E_ldra</code> reports its min error as P1556 and its max error as P1557.
Line 728:		Line 732:
	P1556 and P1557 both trigger the same open-loop boost fallback (KFTVLDRE). In practice, a boost leak between the turbos and the throttle is the most likely cause of P1557, while turbos, wastegates or an N75 that simply cannot reach requested boost (with no leak) should set P1556.		P1556 and P1557 both trigger the same open-loop boost fallback (KFTVLDRE). In practice, a boost leak between the turbos and the throttle is the most likely cause of P1557, while turbos, wastegates or an N75 that simply cannot reach requested boost (with no leak) should set P1556.

			'''In some non-S4 files the same checks report different codes, chosen by the market codeword CWCDCUM.''' In 1.8T Golf/Jetta files (e.g. 06A906032HJ), the ratio check that sets P1557 on the S4 reports P0234 (16618 "Boost Pressure Regulation: Limit Exceeded (Overboost Condition)" in VCDS), and underboost reports P0235 (16619 "Boost Pressure: Regulation Range not Reached").<ref>[http://nefariousmotorsports.com/forum/index.php?topic=7524.0 P0234 incorrectly reported by VCDS?]</ref> In newer files (e.g. 06A906032NL, 8N0906018CJ), overboost and the ratio check both report P0234, and underboost reports P0299 (16683 "Boost Pressure Regulation: Control Range Not Reached"). This is the ECU's own code assignment, not a VCDS error, so a P0234 from these files does not necessarily mean overboost.
	'''Also note that in some non-S4 files (e.g. Golf Jetta LP file), it may also be (incorrectly) reported as P0234 overboost.<ref>[http://nefariousmotorsports.com/forum/index.php?topic=7524.0 P0234 incorrectly reported by VCDS?]</ref> or even negative/positive swapped entirely.<ref>[http://nefariousmotorsports.com/forum/index.php?topic=7524.msg69436#msg69436 negative and positive deviation swapped]</ref>'''

	==PID vs ME7 boost limitations==		==PID vs ME7 boost limitations==

← Older revision		Revision as of 01:14, 2 October 2026
Line 11:		Line 11:
	==Flashing utilities==		==Flashing utilities==
	You need some way to read and write files to the ECU. Some options are:		You need some way to read and write files to the ECU. Some options are:
	* The [http://nefariousmotorsports.com/forum/index.php?topic=12861.0 Nefmoto Free ECU Flashing Software] and an eBay USB VAG KKL FTDI FT232 based cable (or a CH340 based cable). Does not rely on [[boot mode]], but can also flash bricked ECUs ~~because~~ boot mode ~~was recently added~~. You can also use (recent) VCDS (Ross-tech) cables with Nefmoto, but '''you may have to disable "intelligent mode" in the VCDS setup dialog''' (intelligent mode causes the VCDS driver software to constantly poll the cable in the background, which may interfere with non VCDS apps). For more information, see the [http://nefariousmotorsports.com/forum/index.php?topic=6537.0 Nefmoto Flashing Guide.] If you want to use the RossTech cable as a standard OBDII cable on a COM port (not just USB), install [http://www.ross-tech.com/vag-com/usb/virtual-com-port.html RossTech's VCP driver]. Make sure you uninstall all of their other drivers first. Also, when you upgrade VCDS, do not let it install drivers over your VCP drivers.		* The [http://nefariousmotorsports.com/forum/index.php?topic=12861.0 Nefmoto Free ECU Flashing Software] and an eBay USB VAG KKL FTDI FT232 based cable (or a CH340 based cable). Does not rely on [[boot mode]], but can also flash bricked ECUs in boot mode. You can also use (recent) VCDS (Ross-tech) cables with Nefmoto, but '''you may have to disable "intelligent mode" in the VCDS setup dialog''' (intelligent mode causes the VCDS driver software to constantly poll the cable in the background, which may interfere with non VCDS apps). For more information, see the [http://nefariousmotorsports.com/forum/index.php?topic=6537.0 Nefmoto Flashing Guide.] If you want to use the RossTech cable as a standard OBDII cable on a COM port (not just USB), install [http://www.ross-tech.com/vag-com/usb/virtual-com-port.html RossTech's VCP driver]. Make sure you uninstall all of their other drivers first. Also, when you upgrade VCDS, do not let it install drivers over your VCP drivers. The Nefmoto flasher is [https://github.com/NefMoto/NefMotoOpenSource open source]. Its boot mode support (flash, and the 95040 EEPROM on ME7.1/ME7.5) is derived from [https://github.com/EcuProg7/C167BootTool C167BootTool]. Fast init works on the M-box and connects much faster; if it fails on another ECU or cable, use slow init (the default).

	* [http://www.nefariousmotorsports.com/wiki/index.php/Galletto_1260_Flashing_Cable Galletto 1260 cable] and [http://nyet.org/cars/files/apps/Proper%20english%20Galletto.rar software] - Useful in an emergency to recover a bricked ECU. You may have to flash your ECU using [[boot mode]] for it to work if your ECU is bricked. However, it may also work w/o [[bootmode]] (unlike Galletto 1250).		* [http://www.nefariousmotorsports.com/wiki/index.php/Galletto_1260_Flashing_Cable Galletto 1260 cable] and [http://nyet.org/cars/files/apps/Proper%20english%20Galletto.rar software] - Useful in an emergency to recover a bricked ECU. You may have to flash your ECU using [[boot mode]] for it to work if your ECU is bricked. However, it may also work w/o [[bootmode]] (unlike Galletto 1250).
Line 35:		Line 35:
	#* [http://www.nefariousmotorsports.com/forum/index.php/topic,447.0title,.html ME7Check] is a free executable that will tell you whether the checksums in a file are valid or not. It does not correct them, but if you aren't using WinOLS, you should always use it to check files before flashing them to your ECU.		#* [http://www.nefariousmotorsports.com/forum/index.php/topic,447.0title,.html ME7Check] is a free executable that will tell you whether the checksums in a file are valid or not. It does not correct them, but if you aren't using WinOLS, you should always use it to check files before flashing them to your ECU.
	#* [http://www.ecufix.com/shop/ ECUFix] - $150 - Stand alone checksums correction utility		#* [http://www.ecufix.com/shop/ ECUFix] - $150 - Stand alone checksums correction utility
	#* [https://github.com/nyetwurk/ME7Sum ME7Sum] - Open source ~~ME7.x~~ checker/corrector. ~~Currently~~ ~~under~~ ~~development!~~		#* [https://github.com/nyetwurk/ME7Sum ME7Sum] - Open source checksum checker/corrector for VAG ME7.1/ME7.5 and Porsche ME7.2. Check your stock file with it first; some tuner-modified files deliberately leave checksums wrong.
			# (Optional) A disassembler, to see how the firmware uses a map: [https://github.com/NationalSecurityAgency/ghidra Ghidra] with the [https://github.com/keyhana/c166-ghidra-module C166 processor module].

	== Logging utilities ==		== Logging utilities ==
Line 43:		Line 44:
	#* If you are logging M-box, there are a bunch of variables that ME7Logger cannot autodetect that you may want to add to your .ecu file - https://github.com/nyetwurk/ME7L/tree/master/ecus		#* If you are logging M-box, there are a bunch of variables that ME7Logger cannot autodetect that you may want to add to your .ecu file - https://github.com/nyetwurk/ME7L/tree/master/ecus
	#* A list of commonly logged variables is here (lines with a leading semicolon are comments, and thus variables that are not being logged) - https://github.com/nyetwurk/ME7L/blob/master/logs/typical.lst		#* A list of commonly logged variables is here (lines with a leading semicolon are comments, and thus variables that are not being logged) - https://github.com/nyetwurk/ME7L/blob/master/logs/typical.lst
	# Nefmoto [http://nefariousmotorsports.com/forum/index.php/topic,1155.0.html logger] - more or less equivalent to ME7Logger, but writes out an .xml log file that can't ~~really be used anywhere (yet)~~.		# Nefmoto [http://nefariousmotorsports.com/forum/index.php/topic,1155.0.html logger] - more or less equivalent to ME7Logger, but writes out an .xml log file that ECUxPlot can't read.
	# Ross-Tech's [[VCDS]] (aka VAG-COM) - very limited. Useful for reading/clearing DTCs, monitoring fuel trims and misfires, and that's about it.		# Ross-Tech's [[VCDS]] (aka VAG-COM) - very limited. Useful for reading/clearing DTCs, monitoring fuel trims and misfires, and that's about it.
	# APR's [[ECUx]] - No longer actively supported by APR. Avoid if at all possible		# APR's [[ECUx]] - No longer actively supported by APR. Avoid if at all possible