-
Who This Checklist Is For (and What It Won't Fix)
-
The Seven Steps
-
Step 1: Confirm the Torque Class Before You Confirm the Mount
-
Step 2: Match the Servo Motor Controller to the Drive Unit, Not to the Battery
-
Step 3: Route the Speed Sensor Before You Mount the Battery
-
Step 4: Check Battery Communication and Charge Port Orientation
-
Step 5: Brake System Compatibility (Where Disc Brake Conversion Kits Come In)
-
Step 6: Test Any Linear Actuator or Adjustable Assist Mechanism
-
Step 7: Under-Load Test — Not a Parking Lot Spin
-
Step 1: Confirm the Torque Class Before You Confirm the Mount
-
Notes and Common Errors
-
The Two Mistakes That Cost Me the Most
Who This Checklist Is For (and What It Won't Fix)
I handle OEM drive-unit integration and service work — mostly Shimano Steps, some Bosch and Brose on the side. I've been doing this for about six years and I've personally cost my employer roughly $11,000 in rework, freight, and one very awkward dealer call. That number isn't a badge of honor. It's why there's a laminated checklist taped to my bench.
This is the checklist I run before any Steps motor goes onto a frame. Seven steps, around 40 minutes per bike if nothing is wrong. It's written for assembly line leads, shop techs doing retrofit builds, and distributor QC people who don't want to eat the cost of a returned unit.
What it won't fix: firmware bugs that need a Shimano service center, water ingress from failed seals, or a frame that was never designed for mid-drive geometry. If you're working with no-name carbon frames, use this as a starting point, not a warranty.
The Seven Steps
Step 1: Confirm the Torque Class Before You Confirm the Mount
There's a legacy myth I still hear from newer techs — that mid-drive motors require a purpose-built frame. That was true about a decade ago, when Bosch and a couple of others held tight control over their OEM adapter specs. Today the Shimano Steps mount standard and the aftermarket adapter space have largely closed that gap. But "can fit" and "should run" are two different questions.
Shimano's Steps E5000 is rated at 40 Nm. The E6100 is 60 Nm. I've watched assembly line operators treat those as interchangeable because they share a mount pattern — they're not interchangeable, and the difference shows up in two places: the crank Q-factor on certain frames, and the assist curve when a loaded rider hits an 18% grade.
Match the drive unit to the rider profile. Commuter frames with a 40 Nm E5000 are fine up to the 25 km/h assist cutoff that EN 15194:2017 requires for EU-bound units. Cargo and longtail frames with a 60 Nm E6100 need the mount bolts torqued to the frame maker's spec — I'm talking 8–12 Nm range depending on the frame, not the "snug plus a quarter turn" you give a bottle cage. If you're swapping an E5000 for an E6100 on the same frame, check chainline, not just the mount.
Step 2: Match the Servo Motor Controller to the Drive Unit, Not to the Battery
This is the one I see people get backwards. The servo motor controller's job is to translate torque sensor signal into phase current at the motor. It doesn't care about your battery pack voltage nearly as much as it cares about the peak phase current the drive unit will draw.
The E6100 will pull somewhere in the 36 A range at peak assist under load. If you're piecing together a system — say, a salvage Steps unit running with a third-party servo motor controller — verify the controller's continuous phase current rating is at or above that spec. A controller rated at 25 A continuous will work fine on the flat and then throw a thermal fault the first time someone takes the bike up a hill with a loaded rear rack.
I learned that one the expensive way (more on the pitfall notes below).
Step 3: Route the Speed Sensor Before You Mount the Battery
Speed sensor magnet-to-pickup gap should land in the 2–8 mm range depending on the unit. If you can see daylight through the gap from a standing position, it's too wide. The symptom is intermittent E010 errors that read like a motor controller fault and are actually just a bad signal read.
The rookie mistake here, and I've made it, is routing the sensor cable alongside the brake hose. It looks tidy. It also picks up noise on hydraulic systems every time the brake lever is actuated, and the codes that come back look like drive-unit failures. Route the sensor cable separately, or at minimum put a 10 mm standoff between the two.
Step 4: Check Battery Communication and Charge Port Orientation
Mount the battery with the charge port facing downward and toward the drive side if the frame allows it. That's not a Shimano rule — it's a service tech rule. Batteries mounted with the charge port facing up collect water in the port cover seal, and you'll see the corrosion six months later on a "won't charge" ticket.
Before you torque the battery mount, verify the handshake with the head unit. Pull the battery, reseat it, and confirm the pairing light cycles. If it doesn't handshake on the bench, it won't handshake under load.
Step 5: Brake System Compatibility (Where Disc Brake Conversion Kits Come In)
A Steps-equipped commuter with a loaded rider is 15–20 kg heavier than the same frame with a mechanical drivetrain. Rim brakes are technically within spec on some 40 Nm E5000 builds, but I won't sign off on them above that.
If the frame isn't disc-ready and you're retrofitting, this is where a disc brake conversion kit enters the picture. Check three things: rear dropout spacing (135 mm QR and 142 mm thru-axle don't interchange, and frame swap is not an option), rotor clearance behind the motor mount, and caliper mount standard — post-mount, flat-mount, or IS. IS-to-post adapters are common, but they add 5–6 mm of stack height, and on tight frames that pushes the caliper into the chainstay.
Pad compound matters too. Organic pads glaze on a loaded e-bike descent. Sintered or semi-metallic — don't let a customer talk you into organic because they're "quieter."
Step 6: Test Any Linear Actuator or Adjustable Assist Mechanism
This is the step most lines skip, and it generates the strangest warranty tickets. If the frame has an adjustable mechanism — a dropper post with an electronic actuator, an adjustable suspension linkage, or on some cargo setups a servo-driven kickstand — test the full stroke before the bike leaves the bench.
What happens when a linear actuator fails on an e-bike is different from an industrial actuator failure. Industrial units typically fail to a locked position or drop to a safe stop. On a bike, that failure mode can leave a dropper post stuck up mid-ride or a kickstand that won't retract at the wrong moment. Cycle the actuator stroke three times and listen for the motor stalling before end-of-travel — that's the early failure signature.
Step 7: Under-Load Test — Not a Parking Lot Spin
Walk mode engagement first, then a 2 km loaded ride. Loaded means 20–25 kg in the panniers or a simulated rider load on the rear rack. Check three things afterward: crank bolt torque (they settle under the first ride), motor housing temperature (warm is normal, hot to the touch is not), and drivetrain alignment under assist load.
If you can only do a short test, do walk mode plus a hill start. Walk mode is the fastest way to catch a torque sensor calibration issue. A bike that rolls fine on the flat can still fail the first time the sensor sees a 60 Nm assist demand.
Notes and Common Errors
Three things I'd flag before you tape this checklist to your bench.
- Don't trust the universal wiring diagram. Shimano changed connector pinout between the E6000 and E6100 generations. Verify against the unit-specific service manual, not a 2019 forum post.
- Firmware version first, mount second. Drive units shipping with older firmware won't pair with some head units. Update on the bench, not after the bike is assembled.
- Never calibrate the torque sensor with the bike upside down. Some Steps models use gravity as a reference during calibration. Upside-down calibration produces a bike that feels mushy on the first ride and errors out on the second.
The Two Mistakes That Cost Me the Most
In my first year on the floor (that was 2019), I made the classic spec error: I treated the E5000 and E6100 as interchangeable because they shared a mount. I mounted six E6100 units on frames that were spec'd for the E5000. The crank Q-factor was off by roughly 4 mm on every one — the customer's leg felt it before our QC did. Rework came to about $1,300 in labor and shipping plus a two-week delay on the dealer's floor launch.
The second one was a penny-wise call. To save about $40 per unit on a 12-bike order, I sourced third-party servo motor controllers instead of the matched Shimano units. Three of the twelve failed with thermal shutdown inside the first month of customer use. We replaced all twelve to be safe. Net loss was somewhere around $1,900, plus the dealer relationship, which took longer to repair than the bikes did.
To be fair, the third-party controller specs were close on paper. That's exactly the trap — "close" isn't the same as "rated for peak phase current under sustained climb load."
The lesson I keep coming back to is simple. Components that look the same on a spec sheet aren't the same when the rider is climbing with a loaded trailer. That's the quality-perception problem in a nutshell — the customer doesn't feel the $40 you saved. They feel the motor cutting out 400 meters from the top of the hill.