Resolvers unwrapped
by Ben Martins

The introduction of high-voltage batteries in electric vehicles brings new complexities and technologies. Understandably, training is imperative when working on these vehicles and information in articles or YouTube videos must not be used as training replacement.

When it comes to electric vehicles, people often get caught up in the high-voltage system. However, the fact is that the vehicle can’t operate at all without the various inputs and outputs from the low-voltage system. One such input is from the resolver sensor. 

This sensor determines the position of the rotor, enabling the inverter to trigger the insulated-gate bipolar transistors (IGBTs) which allow the current flow to the correct winding. The resolver is a simple device that has no moving parts, a typically reliable component. However, even the most reliable parts can fail and a resolver issue can cause several knock-on effects that trigger additional fault codes unrelated to the resolver. Testing them correctly is a must.

The trouble with testing resolvers is that you need to check all three circuits at the same time to verify the signal. This is made more difficult as the circuits are floating with no common ground. This means that to visualize the signals, you have to either use three differential probes or an Automotive 4000-series PicoScope with floating inputs, the 4425 and the 4425A. The other automotive PicoScopes have a common ground, like most other available scopes. However, you will not be able to use a common-ground scope to test resolvers as explained here. 
 

If you use a common-ground scope connected across all three circuits you will cause a dead short between them and trigger fault codes. If you use a single ground to chassis, the measurement will provide insufficient detail to verify the resolver. Below is a waveform that was taken with a 2 channel common ground scope along with a reference from a PicoScope 4425 to compare.

If you follow the guided test in PicoScope 7 Automotive, you would expect to see the following pattern:

Underneath it all, a resolver is simple. Coils form a stator around the rotor, which is directly connected to the motor’s output shaft. The stator is formed from three separate windings, consisting of an exciter winding and two output windings.

An alternating current at around 10 kHz is passed into the exciter coil, which is then induced into the rotor. The sine and cosine windings are arranged 90 degrees apart so that the amplitudes of the induced currents are dependent on the relative angle of the rotor. The rotor is shaped with lobes (the number of which varies between manufacturers). These create a change in the intensity of the output winding, depending on the rotor position.

Even with this signal, it’s not easy to determine how the vehicle can establish the position of the motor. This is where having all three signals is vital, as we can introduce a math channel that removes the 10 kHz frequency created by the excitation coil to see the actual positioning. To understand the math behind these waveforms, see further guidance in the guided test. For more about maths, see the Math is Cool section on our forum.

If we plot the output of these two sine waves against each other, we can see how the motor controller determines the motor position. Another factor we have to consider is the number of poles present on the resolver, as this will dictate how many cycles there are for one rotation of the rotor. This is one of the many reasons why we should always use genuine parts, as the wrong resolver could lead to big problems!

I know it still isn’t clear how the position is determined, but it can be made clearer by plotting the sine and cosine waveforms against each other.

Gif showing the sine and cosine waveforms plotted against each other to illustrate the cycles.

This animation is for visual representation only and is not a feature in the PicoScope software.

As you can see, a circle begins as each cycle completes. The confusing part comes when there are multi-pole resolvers with more than one cycle to complete one rotation of the rotor. This all depends on the number of poles in the stator and the number of lobes on the rotor. Unless we know the part number, or remove the sensor and count the poles and lobes, there is no real way of knowing. However, from a diagnostic point of view, the information provided in this waveform is extremely useful in verifying the sensor’s output. 

You can learn more about our EV testing solutions here and if you want to discuss any high-voltage application, please email support@picotech.com. 

Comments

Add comment

Your email address will not be published. Required fields are marked *