Tuning Wiki

This is a historical revision. Read the current page or return to its history.

Page organisation in this revision
Page type
Technical article
Subject
Tuning
Parent
Simos 18
Applies to
Simos 18.1
Tool groups
None
Manufacturers
Volkswagen Group

Changes from revision 1 to revision 10

Removed text is struck through; added text is highlighted.

Applicability: Engineering identifiers below are examples from the community SC8S5031_C_OEM Simos 18.1 cross-reference unless a product or different software is named. Definitions are research leads, not universal calibration values: match the ECU hardware, ASW, A2L, indexing, units, axes, scaling, signs and actual configuration. The community extraction can contain mismatched descriptions or indices. SC8S5031 reference.

Detection and response are different

Knock-sensor signals are filtered and assessed within operating windows; the strategy can then apply rapid per-cylinder retard and learned corrections. A low or absent logged correction does not prove no knock: the detection state, signal availability, mechanical noise, sample rate and logger scaling all matter. Examine each cylinder and final timing alongside fuel, load, temperatures and lambda. COBB documents Knock Dataset Selection (filter selection can vary with injection mode), Knock Window State, individual Knock Retard Cylinder 1 through 4, and an average for supported MQB applications. These are logger display names, not calibration maps or a guarantee of availability. COBB monitor descriptions.

Measurement and detection examples

The SC8S5031 KNCK detection cross-reference lists the following calibrations:

ip_knkwb[3] and ip_knkwd[3]: start and duration of an individual cylinder's knock measurement window; example axes are speed and cylinder air mass, with crank-angle outputs. An index must be matched to the software and firing-order/cylinder mapping. ip_knks_thd_fac_1[3]: one knock-detection threshold factor, selected when the documented filter/gain index is 1. The displayed axes are speed and predicted cylinder air mass. It is one factor in threshold calculation, not a direct universal knock limit. ip_knkwb_pre_h_rng[3] and ip_knkwd_pre[3]: examples associated with the pre-window, distinct from the main window. lv_knk_win_ovlp and nl_ref[3] are runtime window and noise-reference measurements, not tables.

Window, filter, gain and threshold calibration determine what the ECU can detect; changing them to hide correction can compromise detection. An independent Mk7 noise-and-knock investigation illustrates how noise and detection thresholds affect logged correction; its vehicle test is not a universal calibration. Validate any detection recalibration with suitable cylinder-pressure and acoustic measurements rather than inferring it from a clean-looking log.

For the same SC8S5031 image, id_knks_fil_frq[3][2][0], id_knks_fac_fil[3][2][0] and id_knks_fil_qly[3][2][0] describe per-cylinder, per-filter frequency selection, weighting and quality-factor maps. Their displayed axes are engine speed and maf_h_rng_knk in mg/stroke. The frequency map stores a coded selection, not directly a frequency in hertz; the acquisition reference describes a separate filter-table selector lc_knks_fil_tab. Confirm decoding, configuration and array indexing in the target software. KNCK detection; combustion-noise acquisition.c_cycnr_knk_det defines a number of combustion cycles for which the noise calculation is frozen after a detected knock in this reference; lv_knk_win_ovlp indicates that the pre-window was shifted to avoid main-window overlap. These represent different parts of the measurement strategy, not a direct measurement of knock intensity or a permission to disregard detection. Verify enable and failure states before drawing conclusions from logged corrections. KNCK detection reference.Noise adaptation and sensor-signal contextThe SC8S5031 KNCK detection reference lists ip_nl_gain_ad as a target for centred engine-noise signal in volts versus n_32 rpm, and ip_nl_fil_gain_inc as a noise-calculation filter gain versus rpm when the noise value is increasing. The former is not a target cylinder pressure or a safe knock level. knks_thd[3] and nl_ref[3] are cylinder-indexed runtime threshold and noise-reference values in the extraction; they are not necessarily exposed by a given logger. Confirm channel scaling, indices, detection state and the selected filtering path before comparing them. Changes to the signal-processing chain can obscure real knock; validating it requires suitable independent instrumentation.Acquisition gain is not threshold gain. In the SC8S5031 knock-detection extraction, ip_knks_gain_1[3] is a per-cylinder acquisition-gain map used for the described IDX_FIL_GAIN_KNK = 1 path, indexed by engine speed and predicted cylinder air mass. knks_cmd_gain_ad[3] is a runtime adaptive gain command. These are distinct from the ip_knks_thd_fac_1[3] threshold factor and the calculated knks_thd[3]; similar names do not make the maps interchangeable. ip_itc_knk_1[0] is an rpm-indexed sensor integration-time-constant curve for that selection, and ip_itc_knk_pre[0] concerns the pre-window. Their stored numbers are not established here as milliseconds. Verify sensor/cylinder indices, selection, encoding and calibration context before interpreting a log or definition.The same extraction describes c_fac_tra_knk as a factor applied to the knock threshold in certain speed/load transients, subject to the documented threshold limit. It is distinct from the preventive ignition-retard map ip_iga_tra_knk described below. This means threshold and correction traces can change with operating state; neither a quiet log nor a particular transient factor establishes safe combustion. Independent pressure and acoustic validation is needed to assess changes to the detection chain.Severe logged correction on a modified engine is a case in point. A Golf 7R forum post reports a forged engine with a hybrid turbo, substantial logged correction and an EPC/limp event under higher load, with additional mechanical noise from changed clearances suspected. No independent pressure test, fault-code details or confirmed fix were provided; the suspicion is not proof of false knock or a basis for weakening detection. Combustion, fuel delivery and mechanical faults should be considered alongside noise sources, and severe correction investigated before further high-load testing.

Knock response and learning examples

In the KNCK control reference, ip_iga_dec_knk describes spark retard at recognised knock, and ip_iga_max_knk the maximum retard correction; the displayed examples use speed and cylinder air mass. ip_fac_iga_dec_knk is a knock-response factor against knock energy and engine speed. They are response/protection calibrations, not recommended reduced settings.

ip_iga_inc_knk appears with an XDF label suggesting decay, while its extracted A2L description describes an increase in integrated correction when knock is detected. Treat its exact behaviour as unresolved until the target's definition and implementation are checked; do not infer a universal recovery rate from its name.

iga_cyl_knk_fast[3] is a runtime cylinder-specific fast retard before adaptation, iga_ad_1_knk[3] a runtime individual adaptation value, and lv_knk_ctl_lim a control-limit flag. These are different stages, not interchangeable editable maps. Confirm index-to-cylinder mapping, sign conventions and channel availability in the logger.

The same KNCK control extraction distinguishes ip_dly_dec_ad_1_knk and ip_dly_inc_ad_1_knk for the individual AD1 learning loop from ip_dly_inc_fast_knk for the fast loop. Each example uses engine speed n_32 on its axis and describes the number of segments between changes. The extraction also uses TDC and cylinder-stroke terminology: establish the scheduler and sign convention in the target before translating a count into elapsed time or assuming a recovery rate. These delay curves, fast correction and learned AD1 correction are distinct quantities.

Transient knock prevention

The SC8S5031 transient-prevention reference describes ip_iga_tra_knk as a transient-retard map versus engine speed and predicted cylinder air mass, and ip_fac_iga_tra as a factor versus cylinder intake-air temperature tia_cyl and the adaptation factor fac_ad_knk. Preventive transient correction is not, by itself, proof of a detected knock event. fac_ad_knk is a runtime global adaptation factor, not a direct fuel-octane measurement.The SC8S5031 transient-prevention reference describes ip_iga_tra_knk as a transient-retard map versus engine speed and predicted cylinder air mass, and ip_fac_iga_tra as a factor versus cylinder temperature and the adaptation factor fac_ad_knk. Preventive transient correction is not, by itself, proof of a detected knock event. fac_ad_knk is a runtime global adaptation factor, not a direct fuel-octane measurement.

Comparing and diagnosing logs

Record fuel and actual ethanol content, rpm, gear, air mass, pressure, temperature, lambda target and actual, combustion mode, final and individual-cylinder ignition timing, and each cylinder's knock correction. Compare like-for-like operating points and note active torque intervention and detection state where available. A repeatable cylinder-specific event warrants investigation of fuel, plugs, coils, mechanical condition, sensor mounting and interference as well as calibration; an isolated event or a clean average cannot establish its cause. Do not use a single numerical retard value as a universal safe threshold.

Worked example: correction and a reported low-speed surgeA 2025 owner report concerns a 2016 CHHB 2.0 TSI with DQ250, an open intake and modified exhaust, showing corrections on several cylinders down to −4.88° on fuel described as 100 octane. With one calibration, a low-rpm symptom was described as turbo surge, and actual boost was later reported to overshoot its target during acceleration. These are owner observations, not an independently diagnosed cause or a validated surge measurement; the fuel-rating convention and comparable test conditions were not established. The report does not establish a safe correction or timing limit.In a similar complaint, synchronized boost target and actual, wastegate command or position, rpm, throttle, gear and per-cylinder correction should be captured alongside the fuel, temperature, air charge, lambda and timing channels above. Pressure units and whether readings are absolute or gauge should be checked, and the reported noise and boost control assessed separately from the knock response. Repeatable operating points should be compared before attributing either symptom to a specific calibration.For global adaptation and low-RON detection examples, see Simos 18 fuel octane and knock adaptation.

Related: ignition and knock overview · ignition timing · knock detection and control · tuning and log diagnosis · ethanol and flex fuel.

For pre-spark combustion, low-speed high-load research and diagnostic limits, see Simos 18 low-speed pre-ignition and knock.Example from a Simos 18.1 file comparisonScope of example: A user-supplied comparison of a Simos 18.1 stock file, Stage 1 and Stage 1 pops-low, with a WinOLS CSV definition export. LTEditor identified box code 5G0906259E and software 111SC800H6G00000. The files and definitions are not publicly available, so this is a single, independently unverified file-comparison observation, not a recommended calibration. The source’s stock values matched supported byte and 16-bit entries after accounting for table transposition. Selection, runtime logic, operating conditions and checksum suitability were not independently established.In this comparison, the basic ignition families ip_iga_bas_ivvt_vvl_port_h/l contain changed entries 0.375 to 4.5 crank degrees higher. Several ip_knks_thd_fac[0..3] and ip_knks_thd_fac_1[0..3] tables have changed entries increased by 0.15625. These factors are inputs to detection, not a direct knock measurement; the change does not establish its complete effect on sensitivity or actual delivered spark. Check the active filter and window, fuel, final per-cylinder timing and independent knock evidence. A clean-looking retard log alone does not establish adequate margin.

Rendered revisions

Open each version to inspect formatting, tables and photographs.

Revision 1: Simos 18 knock detection and control

Applicability: Engineering identifiers below are examples from the community SC8S5031_C_OEM Simos 18.1 cross-reference unless a product or different software is named. Definitions are research leads, not universal calibration values: match the ECU hardware, ASW, A2L, indexing, units, axes, scaling, signs and actual configuration. The community extraction can contain mismatched descriptions or indices. SC8S5031 reference.

Detection and response are different

Knock-sensor signals are filtered and assessed within operating windows; the strategy can then apply rapid per-cylinder retard and learned corrections. A low or absent logged correction does not prove no knock: the detection state, signal availability, mechanical noise, sample rate and logger scaling all matter. Examine each cylinder and final timing alongside fuel, load, temperatures and lambda. COBB documents Knock Dataset Selection (filter selection can vary with injection mode), Knock Window State, individual Knock Retard Cylinder 1 through 4, and an average for supported MQB applications. These are logger display names, not calibration maps or a guarantee of availability. COBB monitor descriptions.

Measurement and detection examples

The SC8S5031 KNCK detection cross-reference lists the following calibrations:

  • ip_knkwb[3] and ip_knkwd[3]: start and duration of an individual cylinder's knock measurement window; example axes are speed and cylinder air mass, with crank-angle outputs. An index must be matched to the software and firing-order/cylinder mapping.
  • ip_knks_thd_fac_1[3]: one knock-detection threshold factor, selected when the documented filter/gain index is 1. The displayed axes are speed and predicted cylinder air mass. It is one factor in threshold calculation, not a direct universal knock limit.
  • ip_knkwb_pre_h_rng[3] and ip_knkwd_pre[3]: examples associated with the pre-window, distinct from the main window. lv_knk_win_ovlp and nl_ref[3] are runtime window and noise-reference measurements, not tables.

Window, filter, gain and threshold calibration determine what the ECU can detect; changing them to hide correction can compromise detection. An independent Mk7 noise-and-knock investigation illustrates how noise and detection thresholds affect logged correction; its vehicle test is not a universal calibration. Validate any detection recalibration with suitable cylinder-pressure and acoustic measurements rather than inferring it from a clean-looking log.

Knock response and learning examples

In the KNCK control reference, ip_iga_dec_knk describes spark retard at recognised knock, and ip_iga_max_knk the maximum retard correction; the displayed examples use speed and cylinder air mass. ip_fac_iga_dec_knk is a knock-response factor against knock energy and engine speed. They are response/protection calibrations, not recommended reduced settings.

ip_iga_inc_knk appears with an XDF label suggesting decay, while its extracted A2L description describes an increase in integrated correction when knock is detected. Treat its exact behaviour as unresolved until the target's definition and implementation are checked; do not infer a universal recovery rate from its name.

iga_cyl_knk_fast[3] is a runtime cylinder-specific fast retard before adaptation, iga_ad_1_knk[3] a runtime individual adaptation value, and lv_knk_ctl_lim a control-limit flag. These are different stages, not interchangeable editable maps. Confirm index-to-cylinder mapping, sign conventions and channel availability in the logger.

Transient knock prevention

The SC8S5031 transient-prevention reference describes ip_iga_tra_knk as a transient-retard map versus engine speed and predicted cylinder air mass, and ip_fac_iga_tra as a factor versus cylinder temperature and the adaptation factor fac_ad_knk. Preventive transient correction is not, by itself, proof of a detected knock event. fac_ad_knk is a runtime global adaptation factor, not a direct fuel-octane measurement.

Comparing and diagnosing logs

Record fuel and actual ethanol content, rpm, gear, air mass, pressure, temperature, lambda target and actual, combustion mode, final and individual-cylinder ignition timing, and each cylinder's knock correction. Compare like-for-like operating points and note active torque intervention and detection state where available. A repeatable cylinder-specific event warrants investigation of fuel, plugs, coils, mechanical condition, sensor mounting and interference as well as calibration; an isolated event or a clean average cannot establish its cause. Do not use a single numerical retard value as a universal safe threshold.

Related: ignition and knock overview · ignition timing · knock detection and control · tuning and log diagnosis · ethanol and flex fuel.

Revision 10: Simos 18 knock detection and control

Applicability: Engineering identifiers below are examples from the community SC8S5031_C_OEM Simos 18.1 cross-reference unless a product or different software is named. Definitions are research leads, not universal calibration values: match the ECU hardware, ASW, A2L, indexing, units, axes, scaling, signs and actual configuration. The community extraction can contain mismatched descriptions or indices. SC8S5031 reference.

Detection and response are different

Knock-sensor signals are filtered and assessed within operating windows; the strategy can then apply rapid per-cylinder retard and learned corrections. A low or absent logged correction does not prove no knock: the detection state, signal availability, mechanical noise, sample rate and logger scaling all matter. Examine each cylinder and final timing alongside fuel, load, temperatures and lambda. COBB documents Knock Dataset Selection (filter selection can vary with injection mode), Knock Window State, individual Knock Retard Cylinder 1 through 4, and an average for supported MQB applications. These are logger display names, not calibration maps or a guarantee of availability. COBB monitor descriptions.

Measurement and detection examples

The SC8S5031 KNCK detection cross-reference lists the following calibrations:

  • ip_knkwb[3] and ip_knkwd[3]: start and duration of an individual cylinder's knock measurement window; example axes are speed and cylinder air mass, with crank-angle outputs. An index must be matched to the software and firing-order/cylinder mapping.
  • ip_knks_thd_fac_1[3]: one knock-detection threshold factor, selected when the documented filter/gain index is 1. The displayed axes are speed and predicted cylinder air mass. It is one factor in threshold calculation, not a direct universal knock limit.
  • ip_knkwb_pre_h_rng[3] and ip_knkwd_pre[3]: examples associated with the pre-window, distinct from the main window. lv_knk_win_ovlp and nl_ref[3] are runtime window and noise-reference measurements, not tables.

Window, filter, gain and threshold calibration determine what the ECU can detect; changing them to hide correction can compromise detection. An independent Mk7 noise-and-knock investigation illustrates how noise and detection thresholds affect logged correction; its vehicle test is not a universal calibration. Validate any detection recalibration with suitable cylinder-pressure and acoustic measurements rather than inferring it from a clean-looking log.

For the same SC8S5031 image, id_knks_fil_frq[3][2][0], id_knks_fac_fil[3][2][0] and id_knks_fil_qly[3][2][0] describe per-cylinder, per-filter frequency selection, weighting and quality-factor maps. Their displayed axes are engine speed and maf_h_rng_knk in mg/stroke. The frequency map stores a coded selection, not directly a frequency in hertz; the acquisition reference describes a separate filter-table selector lc_knks_fil_tab. Confirm decoding, configuration and array indexing in the target software. KNCK detection; combustion-noise acquisition.

c_cycnr_knk_det defines a number of combustion cycles for which the noise calculation is frozen after a detected knock in this reference; lv_knk_win_ovlp indicates that the pre-window was shifted to avoid main-window overlap. These represent different parts of the measurement strategy, not a direct measurement of knock intensity or a permission to disregard detection. Verify enable and failure states before drawing conclusions from logged corrections. KNCK detection reference.

Noise adaptation and sensor-signal context

The SC8S5031 KNCK detection reference lists ip_nl_gain_ad as a target for centred engine-noise signal in volts versus n_32 rpm, and ip_nl_fil_gain_inc as a noise-calculation filter gain versus rpm when the noise value is increasing. The former is not a target cylinder pressure or a safe knock level. knks_thd[3] and nl_ref[3] are cylinder-indexed runtime threshold and noise-reference values in the extraction; they are not necessarily exposed by a given logger. Confirm channel scaling, indices, detection state and the selected filtering path before comparing them. Changes to the signal-processing chain can obscure real knock; validating it requires suitable independent instrumentation.

Acquisition gain is not threshold gain. In the SC8S5031 knock-detection extraction, ip_knks_gain_1[3] is a per-cylinder acquisition-gain map used for the described IDX_FIL_GAIN_KNK = 1 path, indexed by engine speed and predicted cylinder air mass. knks_cmd_gain_ad[3] is a runtime adaptive gain command. These are distinct from the ip_knks_thd_fac_1[3] threshold factor and the calculated knks_thd[3]; similar names do not make the maps interchangeable. ip_itc_knk_1[0] is an rpm-indexed sensor integration-time-constant curve for that selection, and ip_itc_knk_pre[0] concerns the pre-window. Their stored numbers are not established here as milliseconds. Verify sensor/cylinder indices, selection, encoding and calibration context before interpreting a log or definition.

The same extraction describes c_fac_tra_knk as a factor applied to the knock threshold in certain speed/load transients, subject to the documented threshold limit. It is distinct from the preventive ignition-retard map ip_iga_tra_knk described below. This means threshold and correction traces can change with operating state; neither a quiet log nor a particular transient factor establishes safe combustion. Independent pressure and acoustic validation is needed to assess changes to the detection chain.

Severe logged correction on a modified engine is a case in point. A Golf 7R forum post reports a forged engine with a hybrid turbo, substantial logged correction and an EPC/limp event under higher load, with additional mechanical noise from changed clearances suspected. No independent pressure test, fault-code details or confirmed fix were provided; the suspicion is not proof of false knock or a basis for weakening detection. Combustion, fuel delivery and mechanical faults should be considered alongside noise sources, and severe correction investigated before further high-load testing.

Knock response and learning examples

In the KNCK control reference, ip_iga_dec_knk describes spark retard at recognised knock, and ip_iga_max_knk the maximum retard correction; the displayed examples use speed and cylinder air mass. ip_fac_iga_dec_knk is a knock-response factor against knock energy and engine speed. They are response/protection calibrations, not recommended reduced settings.

ip_iga_inc_knk appears with an XDF label suggesting decay, while its extracted A2L description describes an increase in integrated correction when knock is detected. Treat its exact behaviour as unresolved until the target's definition and implementation are checked; do not infer a universal recovery rate from its name.

iga_cyl_knk_fast[3] is a runtime cylinder-specific fast retard before adaptation, iga_ad_1_knk[3] a runtime individual adaptation value, and lv_knk_ctl_lim a control-limit flag. These are different stages, not interchangeable editable maps. Confirm index-to-cylinder mapping, sign conventions and channel availability in the logger.

The same KNCK control extraction distinguishes ip_dly_dec_ad_1_knk and ip_dly_inc_ad_1_knk for the individual AD1 learning loop from ip_dly_inc_fast_knk for the fast loop. Each example uses engine speed n_32 on its axis and describes the number of segments between changes. The extraction also uses TDC and cylinder-stroke terminology: establish the scheduler and sign convention in the target before translating a count into elapsed time or assuming a recovery rate. These delay curves, fast correction and learned AD1 correction are distinct quantities.

Transient knock prevention

The SC8S5031 transient-prevention reference describes ip_iga_tra_knk as a transient-retard map versus engine speed and predicted cylinder air mass, and ip_fac_iga_tra as a factor versus cylinder intake-air temperature tia_cyl and the adaptation factor fac_ad_knk. Preventive transient correction is not, by itself, proof of a detected knock event. fac_ad_knk is a runtime global adaptation factor, not a direct fuel-octane measurement.

Comparing and diagnosing logs

Record fuel and actual ethanol content, rpm, gear, air mass, pressure, temperature, lambda target and actual, combustion mode, final and individual-cylinder ignition timing, and each cylinder's knock correction. Compare like-for-like operating points and note active torque intervention and detection state where available. A repeatable cylinder-specific event warrants investigation of fuel, plugs, coils, mechanical condition, sensor mounting and interference as well as calibration; an isolated event or a clean average cannot establish its cause. Do not use a single numerical retard value as a universal safe threshold.

Worked example: correction and a reported low-speed surge

A 2025 owner report concerns a 2016 CHHB 2.0 TSI with DQ250, an open intake and modified exhaust, showing corrections on several cylinders down to −4.88° on fuel described as 100 octane. With one calibration, a low-rpm symptom was described as turbo surge, and actual boost was later reported to overshoot its target during acceleration. These are owner observations, not an independently diagnosed cause or a validated surge measurement; the fuel-rating convention and comparable test conditions were not established. The report does not establish a safe correction or timing limit.

In a similar complaint, synchronized boost target and actual, wastegate command or position, rpm, throttle, gear and per-cylinder correction should be captured alongside the fuel, temperature, air charge, lambda and timing channels above. Pressure units and whether readings are absolute or gauge should be checked, and the reported noise and boost control assessed separately from the knock response. Repeatable operating points should be compared before attributing either symptom to a specific calibration.

For global adaptation and low-RON detection examples, see Simos 18 fuel octane and knock adaptation.

Related: ignition and knock overview · ignition timing · knock detection and control · tuning and log diagnosis · ethanol and flex fuel.

For pre-spark combustion, low-speed high-load research and diagnostic limits, see Simos 18 low-speed pre-ignition and knock.

Example from a Simos 18.1 file comparison

Scope of example: A user-supplied comparison of a Simos 18.1 stock file, Stage 1 and Stage 1 pops-low, with a WinOLS CSV definition export. LTEditor identified box code 5G0906259E and software 111SC800H6G00000. The files and definitions are not publicly available, so this is a single, independently unverified file-comparison observation, not a recommended calibration. The source’s stock values matched supported byte and 16-bit entries after accounting for table transposition. Selection, runtime logic, operating conditions and checksum suitability were not independently established.

In this comparison, the basic ignition families ip_iga_bas_ivvt_vvl_port_h/l contain changed entries 0.375 to 4.5 crank degrees higher. Several ip_knks_thd_fac[0..3] and ip_knks_thd_fac_1[0..3] tables have changed entries increased by 0.15625. These factors are inputs to detection, not a direct knock measurement; the change does not establish its complete effect on sensitivity or actual delivered spark. Check the active filter and window, fuel, final per-cylinder timing and independent knock evidence. A clean-looking retard log alone does not establish adequate margin.