Shift quality, pressure and adaptation

How the 8HP actually makes a shift, and how to make it feel right. Covers the shift-pulse model, the pressure-correction tables, temperature compensation, gear-start behaviour, and what TurboLamik's adaptation really learns (it is not OEM clutch A-E learning). Get torque setup right first, every table here is scaled by torque.

How an 8HP shift happens

The 8HP has no line-pressure sensor. The TCU commands solenoid current, which sets clutch pressure, and it does this as a timed pulse: pre-fill the oncoming clutch, ramp pressure to take up the load, release the offgoing clutch, then settle. A good shift overlaps the two clutches for the right length of time at the right pressure. Too little pressure or too short a pulse = slip and flare (revs climb, then catch). Too much = a bang and shock load.

Everything below shapes that pulse. The size of the pulse is anchored to the torque value, which is why wrong torque ruins shift quality no matter how you set these tables.

The shift-pressure tables (firmware V10.75)

These live under Gear Pres Correction and Lockup Settings in the XDF (the firmware groups the shift mechanics with the lockup tables):

The TimeUp / TimeDown value on the active program (see program tuning) selects how firm and fast the pulse is, 0 softest, 2 fastest. Use the program value for the broad feel, then the tables above to fine-tune.

Temperature compensation

Cold fluid is thick, so a pulse that is perfect at 80°C will be harsh at 20°C and soft when the box is hot. The firmware corrects for this automatically with dedicated tables, do not fix a cold-shift-harshness by detuning the warm pulse:

Set these against logged oil temperature. Oil temp Warmup and the warmup multipliers handle the first few minutes from cold. If shifts are only bad when cold or only bad when hot, this is the section to edit, not the base pulse.

Gear-start (launch from rest)

Pulling away uses its own clutch-apply tables, separate from gear-to-gear shifts:

• Parameter 1st start / Parameter 2nd start / Parameter Reverse start (15 × 3): the apply shape from a standstill in 1st, 2nd or reverse.
• Gear Start Press Correct (5 × 3): pressure trim for the standstill apply.
• Minimum Clutch Press. Drive / Neutral: the floor pressure held in gear and in neutral.
• Clutch closing time: how quickly the clutch takes up when you select a gear.

A shuddery or slow take-up from rest is a gear-start problem, not a shift-map problem. Raise the start pressure / shorten close time for a crisper launch; soften for a gentler one (Start Gear 2 in the program also gives a softer pull by starting in 2nd).

How adaptation works

Adaptation learns and refines clutch fill and handoff timing on a per-clutch basis rather than a per-gear basis. Because the ZF 8HP engages different clutch combinations for each gear, specific shifts are used to learn each individual clutch:

Relevant calibration settings:

• Enable Adaptation Upshift / Enable Adaptation Downshift: Master toggles per direction.
• Downshift ADAPT. Max Value (7 × Gear): Caps how far the box may adapt each gear to prevent bad readings from running away.
• Adaptation of the torque reduction time: Learns the torque-cut timing window during shifts.

Adaptation activation codes (Live Diagnostics)

In live telemetry, the Adaptation Activation Parameter channel reports why adaptation is active or blocked:

How to drive the adaptation cycle

1. Warm transmission fluid to 55°C – 80°C.
2. Select Program 1 (Road / Drive).
3. Drive on a flat, open road with light, steady throttle (25–30% pedal, 55–140 Nm torque, 1600–3000 RPM).
4. Perform 15–20 gentle upshifts from 3rd through 7th gear, followed by gentle 6 → 5 downshifts.
5. Verify the letter "A" displays on the LCD / dash during shifts.

How to reset adaptations (Clone TCU toggle rule)

[!IMPORTANT] Flashing a map with zeroed adaptation tables does not automatically overwrite learned adaptation memory in the TCU without setting the clone flag.

To execute a complete adaptation wipe:

1. Read the current configuration from the TCU.
2. Zero out all cells in the adaptation maps.
3. In calibration settings, check the flag Save Parameters > Send Adapt. (Clone TCU).
4. Upload the configuration to the TCU (this forces an overwrite of the internal adaptation registers).
5. Immediately uncheck Send Adapt. (Clone TCU) and upload to the TCU once more. Leaving this flag enabled will overwrite adaptations on future map adjustments.

Always: correct torque calibration → reset adaptations → re-learn. Adapting on an uncalibrated torque reading causes the TCU to learn incorrect fill times that cannot be fixed by altering torque afterwards.

A tuning order that works

1. Torque verified (see torque setup). Nothing else is meaningful until this is right.
2. Program TimeUp/TimeDown for the broad firmness you want per mode.
3. Pulse Correction UP/DOWN for global "too soft / too harsh".
4. Time Correction + Slow Ram for the catch (flare vs bang).
5. Temperature tables for cold/hot consistency.
6. Gear-start tables for take-up from rest.
7. Enable adaptation last, with sane max-value bounds, and let it learn.

Shift cut with an aftermarket ECU

Full-throttle flat shifts need the engine to drop torque exactly when the box swaps clutches. The TCU broadcasts its torque-reduction request on CAN; the ECU cuts on it. This is the Haltech recipe Zero To 60 set up and ran, the principle transfers to any ECU that can read an analog CAN channel.

TCU side (V10.75 names):

• Send ID = 705, which makes the TCU pretend to be a Haltech I/O Expander Box B (704 = Box A, use that if B is taken). Turn on the Send ID+0 and Send ID+1 flags.
• Analog Can Out maximum value = 4095.
• Can2 Speed = 4 (1 Mbps, to match the Haltech bus).
• Analog Can Out channel assignments used: 13 oil temp, 14 gear, 18 torque reduction, 16 program. The torque-reduction channel is inverted: 0 V = 100% cut requested.
• The gear channel uses a voltage table: R = 0 V, N = 0.5 V, 1st = 1 V, then 0.5 V per gear up to 8th = 4.5 V. Now the dash and the logger know the gear.

Haltech side:

• Bring the torque-reduction channel in as a paddle-shift input (the flat-shift trigger), with thresholds of 0.1 V on / 0 V off.
• Flat-shift action: ignition cut, ramping from 90% to 0% over 180 ms, with a 350 ms maximum cut time as the safety ceiling.

Set up this way, the car held 520-550 whp flat to the floor through every shift.

One warning from the same car: with shift-cut working and the clutch re-engagement set to no slip, clutch-kick re-engagement broke two axles in 30 minutes on street tyres. The box can apply drive faster than your driveline would like. Leave some slip in the re-engagement unless your axles are up to it.

Source: Zero To 60: Setup TURBOLAMIK TCU & Aftermarket ECU for Shift Cut

Sources: firmware V10.75 XDF (table names above), the manual's adaptations page, and 8speed.au shift-setup notes. Tables and values vary per gearbox, start from the base map for your exact box at download.turbolamik.eu.