Torque setup and the torque maps

Get this right before anything else. The torque value the TCU sees drives line pressure, clutch fill, shift firmness, lockup and adaptation. Wrong torque is the number-one cause of slipping, burnt clutches and harsh shifts on an 8HP swap. Verify against the official manual and your own live data before you drive hard, this is a community guide, not a substitute for testing.

The one safety rule: overestimate torque, never underestimate it. Too high is harsh but safe. Too low means clutch slip, and slip means clutch damage. When dialling in, start high and work backwards.

Why torque is the master input

The ZF 8HP is a fully torque-modelled gearbox. It has no line-pressure sensor; it infers pressure from solenoid current and sets that pressure from the engine torque it is told. So the torque value feeds, directly:

• Line (main) pressure: clutch clamping force. More torque commands more pressure so the packs do not slip. Idle line pressure is around 17 bar at 35°C.
• Clutch fill pressure and fill time: the oncoming clutch piston is pre-filled (~100 ms) before the shift. Fill scales from torque.
• Shift firmness: the engagement ramp targets a pressure band set by torque (roughly 0.6-1 bar light, up to ~6 bar at full load).
• Converter lockup: lockup clamp and slip allowance are torque-dependent.
• Torque reduction: how much cut the TCU asks the ECU for during a shift.
• Adaptation: the box only adapts in an 80-140 Nm window, and stores values against the torque it saw. Wrong torque means wrong learned values.

Two ways to get torque in

The Torque Calculation Input selector (category Torque Calculation) decides the source. The Automatic Detection of Torque Calculation Input flag falls back to internal calculation if CAN torque is lost.

CAN torque (from the ECU). OEM ECUs (e.g. Audi 5HP profile 20, BMW 6HP Exx profile 30) and some standalones put a torque value on CAN2 (engine bus, pins 66/67). The TCU reads it. Verify it. Tuned ECUs frequently report inflated torque (raised to dodge OEM torque limiters), so trim it with CAN Read Torque Multiplier and Offset Engine Torque.

Internal calculation (MAP + RPM + TPS). Most standalone and carb swaps. The TCU estimates torque from manifold pressure, RPM and throttle. Wire MAP to pin 7, TPS to pin 6, RPM to pin 65 (or feed RPM/MAP/TPS over CAN, e.g. Haltech profile 8 with the inputs set to the CAN engine channel) and set the input selector to internal.

The torque maps (firmware V10.75)

These are the actual tables in the XDF, category Torque Calculation:

(There are ft/lb mirrors of the base and max tables; use whichever unit set your map is in.)

Diesel gotcha: Max TQ TPS Lim is still multiplied by the MAP correction even when it is the primary source, so a populated MAP correction table will scale your diesel torque without you asking. And since firmware 10.72 the base maps no longer carry a diesel-specific calibration here: the shipped table is a generic placeholder identical to the petrol map. Fill it for your engine before relying on it.

Old names on YouTube

The Zero To 60 videos (and other older content) use the pre-10.75 table names. The mapping:

Same tables, same job. Anything a video shows in “Calculated Pressure Torque” goes straight into your base torque table.

What a base torque table looks like

This is the Base Torque Table RPM/MAP decoded from the BMW E90 Petrol 8HP70 base map (firmware V10.75). Columns are RPM, rows are manifold pressure (kPa), cells are estimated Nm. It is an example so you can see the shape, your engine and box differ, pull the right base map from download.turbolamik.eu and scale it to your torque curve.

Want to see this in your map? Open your .bin in Map Studio, every one of the 316 tables, as a heatmap or an editable grid.

Setting it up

With CAN torque

1. Select the matching CAN profile and enable CAN2 (engine bus).
2. Watch the engine-torque channel live in TunerPro. At idle it should read roughly 10-20 Nm.
3. Hold 3000 and 5000 RPM unloaded, it should stay low (engine is not loaded).
4. Cross-check a known full-throttle figure against a dyno. If the ECU over-reports, pull it down with CAN Read Torque Multiplier (or Offset Engine Torque for a flat error).
5. Do not adapt until the value looks right.

Without CAN (internal calculation)

1. Wire and calibrate MAP (pin 7), TPS (pin 6) and RPM (pin 65). Power sensors from the 5V outputs.
2. Set Torque Calculation Input to internal.
3. Open Base Torque Table RPM/MAP. Note the value at 100 kPa (atmospheric). Scale the whole table by (your peak torque ÷ that value). Example: a 490 Nm engine showing 288 Nm at 100 kPa → ×1.70.
4. Turbo: fill in Base Torque Table MAP Correction so boosted torque matches reality.
5. Diesel: enable Use Table MAX TQ TPS Lim and build Max TQ TPS Lim (the MAP table is not usable without vacuum).
6. Verify idle reads ~10-20 Nm and the value tracks load sensibly. Do not adapt yet.

Dialling it in on the road

This is the method Zero To 60 used on their internal-calculation builds (most aftermarket installs, they put it at roughly 90%). Remember the rule: overestimate, work backwards.

1. Vacuum first. Program P1, automatic, hold a steady 25-30% pedal. Walk the Base Torque Table RPM/MAP up or down in 5-15% steps until shifts are smooth but not sloppy. The base table’s pressure axis tops out at 100 kPa, so this pass covers everything off boost.
2. Then boost. All boosted torque lives in Base Torque Table MAP Correction. Do boosted pulls and adjust the multiplier, on their RX7 that was +30%, then another +15%. As a sanity anchor, 200 kPa at 200% is about right for an N54 at 1 bar.
3. Read the logs, not your seat. About 0.1 s of slip at the clutch bite is acceptable. Flare (RPM rising while the clutch applies) is not, raise torque.
4. Do not zero the vacuum cells. The negative-torque area matters for the decel clamp.
5. The MAP multiplier ceils out around 5x, very high boost setups may need inflated base values instead.

Reference numbers (so you know what sane looks like): the petrol base maps carry roughly 277-288 Nm at 100 kPa mid-RPM. Normal cruise sits at 60-160 Nm around 70-80 kPa. An F80 BMW DME reports about 100 Nm at 25% pedal in cruise. Their Land Cruiser cruised around 130 Nm, and the rotary ended up at 120-180 Nm at 100 kPa. If your cruise torque reads 400 Nm, it is wrong.

Offset Engine Torque is a diagnostic, not a calibration. Suspect the base table is too low? Add +250 Nm everywhere via the offset. If the slip disappears, the base table is too low. Then fix the table and set the offset back to zero. Never ship a map with it set.

Order matters: torque before adaptations. Adaptation runs in an oil-temperature window, it starts around 50°C and stops above 80°C (Oil Temp MIN / Oil Temp Max). If you let it learn on bad torque data inside that window, it trains the box into bad behaviour you then have to reset.

Source: Zero To 60: Setting Up Internal Tq Calculation

The danger: wrong torque

Too low (under-reported): line pressure too low → clutch slip in gear and during shifts → friction material polishes off → fluid contamination → heat. Lockup cannot hold zero slip → more heat. Adaptation learns values that are too soft for the real load.

Too high (over-reported): pressure too high → harsh, banging shifts and shock-loaded friction plates → accelerated wear. The TCU also requests bigger torque cuts than needed (a power dip on every upshift).

Either way, adapting with wrong torque is worse than not adapting. The box converges on wrong fill values and you cannot fix it by correcting torque afterwards. You have to correct torque, reset adaptations, then re-learn. See troubleshooting and the manual’s adaptations and internal torque calculation pages.

Common mistakes

• Adapting before torque is verified. The single most common cause of swap-box damage.
• Trusting a tuned ECU’s CAN torque without checking it against a dyno.
• Pointing RPM/MAP/TPS at the wrong input source (analog pins vs CAN channel).
• Fixing harsh shifts by raising shift firmness instead of correcting over-reported torque.
• Not resetting adaptations after changing the torque calibration.

Sources: the firmware V10.75 XDF (table names above), plus 8speed.au torque calculation, why torque matters, and HP Academy’s 8HP guide.