eBike & eScooter Diagnostic Guide

A spare throttle is always a good-to-have in my opinion. The nothing-fancy one on my ebike is ~$30 from Rad Power. If you and a friend or family member each have the same throttle, a good one can be used as a diagnostic tool but be careful not to break the good one trying to diagnose or fix the bad one. Take it slow. Slow is smooth; smooth is as fast as it gets.

Some common causes for throttle issues: 

• Pinched, crimped, or cut wires.

• A disconnected plug (like if it gets snagged). Note that some plugs will be internal and may be difficult to see or access.

• The sensor inside the throttle may be broken. It's actually pretty delicate and not designed to take the full weight or the rider or crash impact.

• If there's an error code, any problem with the throttle will usually throw the same code.

• A false-positive from one or both brake sensors.

Common error codes for throttle issues:

• Segway: 14

• Hiboy: E2

• Lectric: E008

• Rad Power: E22, E16

If you want to have a basic understanding of how a throttle works, most of them use a hall effect sensor with 3 wires. It senses a magnet. More specifically, it senses how close a magnet is to the sensor. When you move the throttle, you're moving a magnet closer to or further from a sensor.

3-wires gives us 6 possible ways to hook it up (3x2x1) but there are also 2 default states for the sensor to exist and 2 plug ends (male/female) per type of plug = 24 possible configurations per throttle per plug type. This is why I always consult the manufacturer for replacement parts as the first step. There are so many options, it's easy to get the wrong one. Many sellers don't list full wiring diagrams or even good/accurate pictures.

3 wires are (+) (-) and a (sensor wire) and this is a low voltage line, usually between 3-7V. The sensor wire is a 2nd (+) line but it passes through a percentage of supplied (low) voltage either close to full or close to zero depending on the default state. A throttle which is normally open (passes most/all voltage through by default) is never going to be directly compatible with a throttle that is normally closed (passes through zero or close to zero by default) because they work in opposite directions. A normally open throttle drops [sensor] voltage as you twist it, a normally closed throttle's voltage rises as you twist it. If you connect a normally open throttle to a bike which is intended for normally closed, the wheel will likely engage fully when you power it on and slow as you twist the throttle, it will probably not stop completely without brake cutoff. Normally closed throttles tend not to work at all with normally open systems. Vendors don't often list this spec but we can test for it.

If the bike or scooter powers on, we can find the live wires in the throttle disconnect to map out which is which but we can also map the wiring diagram ourselves with a a low voltage power source, multimeter, and needle probes. A breadboard with patch cords and alligator clips is also advised. Connect the (+) end of the power source to the (+) throttle wire, connect the signal wire to the multimeter (+) probe, connect both the multimeter (-) and throttle (-) to the power source.

Common causes for motor cutout:

• Brake Sensors False Positive - Most ebikes/escooters have a safety cutoff which cuts power to the motor when brakes are engaged.

• Throttle and/or Pedal Assist Sensor (PAS) issues

• Cut Motor Wire, especially where the cord comes out the end of an axle or if there are any spots where a tire might rub.

• Does the motor wheel roll forward but not backward? Bad planetary gears which may be replaceable.

• Does the motor sound like it's coming on but the wheel doesn't spin? Bad planetary gears or clutch. Might be serviceable.

Are there brake lights and if so, are they stuck engaged?
Can you disconnect the brake sensors to test w/o them?
Does the cutout correspond with a whole bike shutdown or does the bike stay on?

Does the motor cord come out of the end of the axle and if so, is the bend in the cord concealing any cuts to the housing or internal bundle of wires?
Did we expose any motor wires from rubbing on a disc brake?
Are there any cuts from the axle rocking in the frame?
The cord doesn't need to come out of the end of the axle to be cut on its way out of the wheel but the cords which do come out of the end of the axle are most likely to get cut.
Regular flexing will eventually result in mechanical fatigue from repeated bends.

Error codes, at best, give us a general area to look for problems. Many of them, like E10 and E30 often represent non-specific communication errors which can be bad continuity in wiring, mismatched programming language between display/controller, or other issues. Each manufacturer can set its own codes, it's always best to look them up for your specific make/model. I try to put up accurate information but it may have changed between when I wrote this and when you're reading it. It's never a bad idea to take a picture of an error code chart that sometimes comes in a manual. Text or email yourself with some keywords that would make it easy to find later.

Displays for eBikes and eScooters are far more important than they seem. It's not like the dashboard instruments in a car which are only indicators. The display is where all settings are actually stored and without it, an ebike is just a bike. Some displays have more features and settings than others - more isn't necessarily better. But beyond that, the display has an internal language to talk with other components. The components need to speak the same language. This language (or protocol) is not published but it's why Lectric and Aventon displays are not compatible even when they share the same physical plug, they're speaking different internal languages.

Can you roll the motor backwards but it won't engage? If so, possible reasons include false positive brake sensor, short in wiring harness, bad throttle, bad controller, bad motor. You might notice that's all the components. There are basically no redundant parts in ebike electronics, it's an all or nothing deal and each component can fail in multiple ways that can affect the whole system differently.

If the wheel will spin forwards but not backwards, the motor planetary gears likely have broken teeth which have jammed the motor. You can still spin forward because the clutch mechanism in the planetary gear is able to freewheel which allows the wheel to coast but the motor won't be able to engage, it is jammed. Sometimes these nylon gears can fail partially, resulting in a loud grinding sound. If allows to progress, the motor will either jam or start spinning freely. The nylon gears inside are a sacrificial part, intended to save the motor. Their sacrifice is rarely necessary. Changing to metal gears is not really an upgrade but it might be to-spec. ex: Super73 uses a steel-nylon-steel sandwich on a larger bearing that is not compatible with any known vendors. Nylon planetary gear replacements are reasonably available on amazon. Expect to need a 4" gear puller to remove the clutch and take great care not to strip bolts in the motor casing, cheap ones may strip no matter what and require extraction.

If direct drive motor (motor w/ electronic braking), the wheel will likely seize completely but possibly move in stepped increments given enough force. Cause is often a short in phase wiring which can be in motor, controller, or both. If phase wires are disconnected does the motor unseize? Test for continuity between phases of the motor, there should be none. Test for continuity between the power connectors on the controller and phase wires from the controller.

The wiring harness, sometimes referred to as a spring cable by Segway, is basically a bundle of extension cords that connect a single plug on the controller to a bunch of plugs at the handlebar (brake sensors, throttle, lights, & display). They have a tendency to get snagged or kinked if not cable-managed well but also I have found that exceptionally tight zip ties, even from the factory, can crimp and cause problems in this very delicate bundle of wires, especially after where the bundle splits into individual ends. The alternative to having a wiring harness is long wires for each connection in a dedicated run all the way to the controller. I think I prefer the harness bundle vs individual runs. 

In the shop I went through the trouble of making a unit that was able to connect to all the handlebar-end connections at once which speeds up the testing process. A multimeter with needle probes and some patience is all that is technically required. Pen & paper to chart your progress will remove a lot of guesswork and I strongly recommend getting a basic breadboard kit if you don't have one. The jumpers fit better into the female Julet plugs and the breadboard makes testing easier. Testing continuity across every wire in the harness can rule out issues related to any handlebar controls; brakes, throttle, display, lights. 

It's important to note that some internal wires are shared across plugs and that's ok. Within each plug on the handlebar-end, each wire needs to have good continuity to only 1 pin at the bundled-end. If that's true, it's probably fine. But if you lack continuity on a single wire, the corresponding component will not work. You might be able to go without a brake sensor, light, or even the throttle; but the display is a critical component. 

It's good form to check the functionality of a component when you find the wiring harness to be bad. The component can also be bad. For example, if the throttle isn't working and you find that there's a continuity issue in the harness, you should still test the throttle. Best case you confirm it's good using equipment that is already out. The worst case would be ordering a harness only to discover the throttle is also bad and now you have to wait longer.

Did you recently bypass a speed governor? A little bit faster can be a LOT more power draw. You should put the governor back.

Barring a speed limit bypass, your battery is probably dying. If under warranty, submit a warranty claim. They have a limited number of charge cycles before capacity starts to degrade. Most warranties consider up to 20% capacity loss over the warranty period to be normal wear but this degradation will continue with use. My original ebike battery which used to easily carry me ~35miles now only holds ~10miles worth of charge after a several hundred, maybe a thousand, charge cycles. I would love to get more out of it but the bike is almost 4 years old at this point with nearly 20,000 miles. I put down some charge cycles. The math adds up.

Sometimes, heavy loads can cause a battery to cut out. Stiff headwind, a 2nd rider, a lot of cargo, or a steep hill will all cause heavy power draw relative to cruising around in the Phx metro area, which is pretty flat. You may be able to dial back the load but ultimately this battery is on its way out.

The battery is often the single most expensive component in an ebike or escooter. You can think of it like paying for a battery and a motor and the rest of the bike/scooter comes included. Batteries also have shelf lives. Which is part of why they are not stocked locally in a lot of shops. Too many options and if the stock doesn't move quickly it goes bad on the shelf.

Load testing a battery: first test the voltage coming from the battery. Then, connect a modest load to the battery, a lightbulb is a common choice but try to use an incandescent bulb if available for higher power draw, note that brightness scales with voltage in an incandescent bulb so it will only barely glow when connected to a 36-48v battery. With the load connected, use your multimeter to check the voltage drop across the load. With only a modest load we should see negligible voltage drop across it. If the drop is significant, the battery is bad. If the drop is not significant, check for things that could be artificially increasing motor load. I have found rocks and other debris lodged between wheels and frames before. Stuck brakes will cause the motor to work harder.

Unless the manufacturer has an app that lets you alter the speed limit, you're asking for a motorcycle. Sorry, but I won't help you turn your bike or scooter  into a motorcycle. Here's why:

In my experience, if an ebike is really capable of traveling faster, the manufacturer will include a procedure to remove the limit.

Legal reasons: Once it's a motorcycle, it needs to be registered and insured and you need motorcycle endorsement in order to ride on public streets. We're still in the wild west right now but Scottsdale and Gilbert PD have already started to crack down on roving gangs of kids riding e-motorcycles masquerading as ebikes. In some cases, the parents have been fined for the reckless riding of their children. Which is a shame. It should be closer to all cases.

Safety reasons: With few exceptions, the people who ask for this service are not skilled riders who show respect for the safety of others. It's one thing to do something dumb and hurt yourself in the process, it's something else to hurt an innocent bystander. Even 20mph is very fast on a bike yet some people want to go faster. I challenge you to maintain 20mph via pedal-only for even a single mile. Crashes over 20mph have much greater chance of serious injury as well. I won't be party to that.

Service reasons:  Going even faster will not significantly reduce your travel time but it will greatly increase wear and tear on the bike and dramatically reduce range. My ebike draws ~70W to cruise along at ~13mph in the lowest assist mode, that's about as slow as it can cruise. To cruise at ~17mph on flat ground draws about 160W. Cruising at 20mph draws ~350W. That is a real objective measurement of the power involved to cruise at various speeds on my bike, which is usually towing a trailer. Less than double the speed requiring 5x the power. Now consider all the bumps you experience while riding. You're hitting them faster which means you're hitting each bump harder. And ebikes tend to be heavy. And if you got an inexpensive one, it probably didn't have the best quality control or cargo capacity in the first place. All of these factors and more are why ebikes have speed limits built-in.

This one can be difficult to test depending on the bike but if your display has a settings menu with a bunch of parameters like P01, P02, P06, etc. For example, on a Lectric bike, P10 is the setting that controls whether the PAS is active. 0 is PAS only, 1 is Throttle only, 2 is PAS+throttle. If someone was fiddling around and changed a 0 or 2 to a 1 in that setting, your pedal assist sensor would suddenly stop working. Consult the display manual for your specific parameter settings, not all displays have all options.

If the P-settings didn't do the trick and if there's an external PAS on the bottom bracket spindle (where the cranks are attached, can be either left or right side), we can test the PAS similar to how a throttle can be tested but instead of a range of results corresponding to the physical twist range of the throttle, a pedal sensor is going to send on/off pulses as it turns. These on/off pulses will correspond with the pass of a magnet across a hall sensor.

If your bike has a torque sensing bottom bracket, I do not currently have a method of testing it independently short of connecting another bike's torque sensing bottom bracket with an extension julet cord. They tend to use the 6-pin (purple or black) connection. A note about torque sensing bottom brackets is that one side cup is keyed and must be installed in the frame before inserting the replacement bottom bracket or you will clip the cable on the frame when you try to torque it and the cable will fall to the ground ever so gently. Ask me how I know.