8HP controller comparison — gateway vs standalone vs ECU-native

There are three ways to make a modern OEM automatic or dual-clutch transmission work outside its original car. They are genuinely different architectures with different failure modes, and the right one depends on your build rather than on which is "best".

We ship two of the three, which is why this page can be written honestly: we have no stake in talking you out of the approach that suits you.

The three approaches

Gateway Standalone controller ECU-native
The factory control unit Kept Removed Kept or removed, depending
What runs the gearbox The factory unit The controller Your engine ECU
What the device does Recreates the CAN environment the factory unit expects Drives solenoids and reads sensors directly The ECU does both jobs
Boxes in the car Engine ECU + gateway + factory TCU Engine ECU + controller Engine ECU
Shift quality out of the box Factory What you calibrate Depends on the implementation
Control over shift points None — they are in the transmission's tune Complete Complete
What you must calibrate The engine picture, chiefly torque Everything: pressures, fill, adaptation Whatever the ECU exposes

What each is genuinely good at

A gateway is the least work for the best baseline. The factory calibration is the product of enormous manufacturer effort, and keeping it means keeping shift quality you are unlikely to match by hand. You supply an accurate engine picture and the transmission behaves as designed.

The cost is that you inherit its decisions. Shift points, drive modes and lockup are in the transmission's own tune, and no gateway parameter changes them.

A standalone controller gives you everything, and asks for everything. Every pressure, fill time and adaptation is yours. That is the only route to genuinely custom behaviour — and it is also the route where a poor calibration is your poor calibration. It is the answer when you need control the factory unit will not give you, or when the factory unit is not available.

ECU-native is fewer boxes and one place to tune. If your engine ECU already controls your transmission acceptably, that is a real advantage: one device, one software, one wiring loom, one vendor. Nothing here is an argument against it.

Choosing

Use a gateway when you have the factory control unit, you want factory shift quality, and you do not need to change how the transmission behaves.

Use a standalone controller when you do not have the factory unit, you need control it will not give you, or you are running a transmission with no gateway path at all — DQ250 and DQ500, for instance.

Use ECU-native when your ECU already supports your exact transmission variant to a standard you are happy with. If you are already running an ECU whose transmission control covers your gearbox, that is a reasonable choice and you should take it. One less box, one less loom, one less thing to learn.

The question worth asking of any ECU-native option is how far its support extends across generations and variants — transmission families change electrically between generations, and support for one is not support for all.

Where the real difficulty sits

This is the part the architecture choice does not change, and it is worth knowing before you pick:

Engine torque is the hard problem in all three. Every approach needs a trustworthy torque value, because that is what sets clutch pressure. An under-reporting signal slips clutches; an over-reporting one shifts harshly. No architecture removes this, and none of them validate it for you.

Approaches differ in where the torque value comes from — the ECU's own model, a correction factor applied to it, a table you populate, or an independent calculation from fuel and air. What matters is whether the number is right, and that is a calibration question rather than an architectural one. See engine torque estimation and the ten-minute check.

The signals are the other half. Whatever you choose needs RPM, load, throttle, pedal, brake and temperatures, and a brake signal is missing from almost every aftermarket ECU stream. That work is the same in all three cases.

When IOcan is the wrong choice

Stated plainly, because finding out during commissioning is expensive.

You want to change shift points, drive modes or lockup. A gateway cannot. Those live in the transmission's tune — why. Either re-tune the transmission, or use a standalone controller.

You do not have the factory control unit. A gateway has nothing to talk to.

Your gearbox has no gateway path. DQ250 and DQ500 are standalone only.

Your ECU cannot act on a torque-reduction request. The gateway sends a percentage; something has to turn it into ignition retard and pedal position. If your ECU has no facility for that, the transmission still shifts, but upshifts are harsher and downshifts do not rev-match. The standalone drivers can execute the cut themselves; the gateways cannot — why.

Your gearbox is on the withdrawal list. ZF 4HP, Honda NSX, Mercedes 722.6, Aisin AA80E and A960E are being removed. Do not start a new build on them.

You need it road-legal in a regulated market. These are off-road-use devices.

You are on DKG or PDK and cannot be an early adopter. 8HP is the mature platform; the others are documented more thinly and are actively changing. See platform maturity.