Quickshifter Load Cell Setup on a Mectronik ECU: MeCal 3, Step by Step

A load cell is not a switch, and setting one up is not the same job. A push-pull microswitch has one thing to get right: where it clicks. A load cell hands the ECU a continuous force signal, which means the ECU has to be told how to read that voltage as newtons before any of it means anything. Get the reading right and the trigger force is obvious. Get it wrong and you will spend a session chasing shifts that fire twice, or not at all, with settings that look perfectly reasonable on the laptop.

This is the process we use on our own bike: a Mectronik MKE7 (WSS600_A), firmware 2.27.44, running a load cell on analog input AN5 on a Triumph Street Triple 765. The numbers below are ours. Some of them you should copy, and one of them you absolutely should not — we will be clear about which.

What your ECU needs

Before anything else, check your ECU can do three things. If it can, a load cell will work. If it cannot, no amount of calibration will save it.

A spare analog input reading 0-5 V. This is where the force signal goes. On our MKE7 that is AN5, pin B4.

A gearshift strategy that accepts a load-cell input. Specifically, one where you enter a gain in N/V and an offset in volts, and set the trigger as a force. If your gearshift strategy only offers a voltage threshold, you can still make it work, but you lose the thing a load cell is for — you are back to a threshold that means nothing physical.

The ability to disable the pull-up resistor on that input. Non-negotiable, and the subject of the next section.

Disable the pull-up first

This is the one that catches almost everyone, so deal with it before you touch a single calibration number.

Analog inputs on a race ECU often ship with a pull-up resistor enabled, because the most common thing wired to them is a simple single-switch shift sensor, which needs one. A load cell does not. Leave the pull-up on and it will drag the sensor off zero — you will read a standing force with the lever completely free, and every threshold you set afterwards will be built on top of that error.

In MeCal 3 this lives under Calibration → ECU IO, not in the gearshift strategy file. That separation is exactly why it gets missed: you can set up the entire LoadCell sensor page correctly and still have this wrong. Set the pull-up on your chosen analog input to DISABLE.

MeCal 3: Calibration → Sensors → LoadCell

Five parameters. Two are fixed, one is yours alone, and one has a sign you have to check rather than assume.

InChannel — the analog input you actually wired to. Ours is 1015, which is AN5 on pin B4. Yours will differ; use your own pinout.

Conversion.Gain — 500 N/V, signed. Whether it is +500 or −500 depends on which way the voltage moves when the lever is loaded. Do not guess this. Watch the raw analog voltage while you press the lever: if it falls, enter −500 N/V; if it rises, enter +500 N/V. A sign error does not look like a sign error in the logs — it looks like a sensor that never triggers.

Conversion.Offset — your own measured no-load voltage, read with the lever completely free. Ours is 1.677 V. Do not copy that number. It is a function of how your particular cell sits in your particular linkage, and every install lands somewhere different. Copying someone else's offset is the same mistake as leaving the pull-up on: a constant error baked into everything downstream.

KFilter — 10. KFilterDForce — 0. These are good starting values and there is rarely a reason to move them until everything else is proven.

Trigger force, and why hysteresis matters

Once the ECU reads real newtons, the trigger becomes a number you can reason about instead of a number you tune by feel.

Logged at race pace on our bike, peak lever force is roughly 350 N on the up-shift and 190 N on the down-shift. The down-shift is lower because the motion is different, and that asymmetry is normal — do not try to make the two numbers match. From those peaks:

Up-shift: TrigForce 200 N, ReArmForce 100 N.
Down-shift: TrigForce −120 N, ReArmForce −60 N.

Keep ReArmForce meaningfully below TrigForce. That gap is your hysteresis: the force has to fall back past ReArmForce before the strategy will fire again. Set them equal and a single shift can trigger twice as the force wobbles across the threshold on its way down — which shows up as a phantom double shift you will struggle to reproduce in the pits.

Your own forces will differ with lever ratio, rod position and rider. The reliable way to set them is to log a handful of shifts at the pace you actually ride, look at your own peaks, and set the trigger at roughly 60% of peak. High enough that road-going lever pressure never trips it, low enough that a genuine race shift always does.

Do it in this order

Offset, then gain, then trigger force. The order is not arbitrary — trigger force is meaningless until the first two are right, so setting it early just means setting it twice.

1. Offset. Lever completely free. Read the raw analog voltage and enter it as the sensor offset.

2. Gain. Enter 500 N/V with the sign you verified above by watching the voltage move.

3. Trigger force. Set it from your own logged peaks, then save to flash.

4. Verify. Power-cycle the ECU and confirm the force reads within a few newtons of zero with the lever at rest. This step exists because a setting that survives a power cycle is a setting that is actually stored — and because if the pull-up is still enabled, this is where it shows up.

Mount it near mid-span

One mechanical detail that gets overlooked. Adjust the rod length so the resting voltage sits near the middle of the sensor's output span. A resting point pushed toward one end of the range costs you travel in that direction before the signal clips, and clipping is invisible in the settings — the force simply stops climbing, and a hard shift reads the same as a moderate one.

Ours rests at 1.677 V, which leaves noticeably more room in tension than in compression. That is a compromise we chose knowingly for this linkage; check where yours lands rather than assuming it is centred.

Check for drift before you trust it

The last thing worth doing is nothing at all: set your zero in the morning and check it again at the end of the day. A shift sensor that drifts gives you a threshold that quietly moves out from under you, and you will blame the strategy long before you blame the sensor.

We logged our own cell across a full 20-lap test day. With the lever at rest the reading stayed at −0.02 N. That is the standard worth holding a load cell to, whichever one you run.

The sensor we run

Everything above was developed on our own Quickshifter Load Cell — ±1000 N, bidirectional, 500 N/V, 5-16 V supply, M6 both ends, IP65, terminated in a 3-pin JST that mates with a Solo Engineering harness. It is ASBK Next Gen legal and fits any bike running a compatible Mectronik World Supersport ECU.

If your bike or ECU is not one we have listed, the three requirements at the top of this guide are what actually decide it. Send us your ECU model and we will tell you straight.