IMC102T Rotor Lock fault

[AG_Serad]

Hello,

We have a problem that we do not reproduce at home, but that our client has regularly.

We have motor error bit 7: "Rotor Lock Fault"

My customer tells me that it happens instead when a heat seal placed on the axis of the motor, produces a very small resisting torque at a given angle. But this resistant torque is really low.

In the reference manual documentation, §4.3 (Difficulty starting the engine), there is a way to adjust the engine.
We are asked:
- to make sure that the current measurement is correct: I will check

- Check that the motor parameters are correct: I, Ld, Lq etc. : it's OK

- To adjust the PI speed regulator and the speed filter: Should the Pi loop should be hard or soft? Is the integral should be greater or less? Should the filter time constant be higher or lower? in what proportion? 10%, double?...

- Adjust the minimum speed: I'm at 150rpm. what do I mean by adjust?

- Adjust the acceleration ramp: Should I accelerate more slowly? Currently, acceleration is 600rpm/s = 4s to reach VMax at 2400rpm... it is already low I think!

- Adjust the time constant of the flux estimator: what parameter is it? decrease it? increase it? in what proportion?

- Increase the motor current limit: my motor is at 2.4A, the starting current is limited to 30%, I increased it to 70% => This change could be have an impact on my problem?

- Finally, MCEWizard calculates an ISDuty at 50%. To limit the noise at startup, I lowered it to 5%. Could this be the source of the starting problem?

It's very difficult for me to see the effect of adjusting the parameters, because on my side, I don't have the problem.
It will therefore be necessary that for each minimal adjustment, I send a correction to my customer, run it for 2 to 3 days to tell me if he notices the problem or not.

I would therefore like to know which of these actions are the most predominant?

[Krupashankar]

Hi @AG_Serad,

Thanks for posting in the Infineon Community.

I would therefore like to know which of these actions are the most predominant?

---

We would like to list down things from highest priority to lowest:

- Finally, MCEWizard calculates an ISDuty at 50%. To limit the noise at startup, I lowered it to 5%. Could this be the source of the starting problem?not.

---

Yes, This could be the problem. This parameter specifies the PWM duty cycle during ANGLE_SENSING state. For better current sensing quality, in single shunt current sensing, duty cycle of angle sensing should not be too low otherwise active vector will be too short to sense the current. In leg shunt current sensing, duty cycle should not be too high otherwise there will not be enough time to sense the current during zero vector. 

Try changing ISDuty to 50% or more based on your Current shunt configuration 

- to make sure that the current measurement is correct: I will check

---

Could you please share .mc2 file along with the details of control board and the power board( Schematics if any). If sensorless algorithm is used to determine the angle the current sensing is very essential part.

---

- Adjust the minimum speed: I'm at 150rpm. what do I mean by adjust?

By varying this parameter the motor will run in open loop until motor reached this speed, you can try increasing this value at higher speed motor will be able to start reliably

- Adjust the acceleration ramp: Should I accelerate more slowly? Currently, acceleration is 600rpm/s = 4s to reach VMax at 2400rpm... it is already low I think!

You can further try reducing the ramp rate it depends on the load connected, Each ramp should be successful  to start the motor properly.

- Adjust the time constant of the flux estimator: what parameter is it? decrease it? increase it? in what proportion?

Flux estimator time constant is 4-5 times of motor winding time constant( L/R). This has to be in accordance with motor parameter.

---

- Increase the motor current limit: my motor is at 2.4A, the starting current is limited to 30%, I increased it to 70% => This change could be have an impact on my problem?

---

Yes, This will have an impact and this will  provide more starting current. You could try increasing it to start motor with high load but keep the value less than the safer operating limit of the motor.

- To adjust the PI speed regulator and the speed filter: Should the Pi loop should be hard or soft? Is the integral should be greater or less? Should the filter time constant be higher or lower? in what proportion? 10%, double?...

Kp and Ki values are chosen based on system parameters, In general Kp will be larger compared Ki. You could manually try changing the values of Kp and Ki and observe the response of the system. If you are referring to Speed feedback filter time constant- You can try increasing, this will filter out high frequency ripple.

Thanks,

Krupashankar

[AG_Serad]

Hello

Many thanks for your complete and detailed feedback.

I will answer on each point: 

- About the IS_Duty: I will try to convince my customer to accept a "clack" noise on startup.

- Mc2 file and Schematic was sent by private message.

- About minimum speed: like when problem occusr, motor not start. It make only some small movre on left, on right... I don't see why increase minimum speed could perform better the start? *

- When motor start to move, amways start is Ok. When the problem occurs, the motor make very very small move. I'm not sure to decrease the ramp will perform better the start? Also, my customer will find the startup phase too long 😞 *

- Currently, Ld inductance is 8.5mH, R is 5.8ohms, so Tau=1.46ms. I should change the flux estimator currently at 15ms to about 5.8 to 7.32ms right?

- Start current is already increased at 70%, so I must make some tests with that. For me, I don't know why not configure it to 100%?

- I try to change Ki and Kp in big proportion without big change on motor behavior...

- I can try increase Speed Feedback filter time constant, currently at 0.2ms. If I increase, this will add some delay on speed loop and can make it instable. Is 1ms seems to you a good value?

*I think I will try these solutions in the last way.

Best regards

[Krupashankar · Accepted solution]

Hi @AG_Serad,

Thanks for sharing the .mc2 file and schematics. We could find Motor Input Scaling and Current Input to ADC Offset Voltage these parameters are according to the hardware.

- About minimum speed: like when problem occur, motor not start. It make only some small movre on left, on right... I don't see why increase minimum speed could perform better the start? *

The motor will continue to run in open-loop up to RPM specified in  Ques 9- Minimum Running Speed and rotor lock fault will not happen till motor reaches that speed.

- Currently, Ld inductance is 8.5mH, R is 5.8ohms, so Tau=1.46ms. I should change the flux estimator currently at 15ms to about 5.8 to 7.32ms, right?

It is the ratio of Line to line inductance(L) not Ld it is  (L/R).  The calculation is correct, please make sure correct Line to Line Inductance is used for calculation.

- Start current is already increased at 70%, so I must make some tests with that. For me, I don't know why not configure it to 100%?

This depends on the maximum motor current and thermal stress on the switches for high current. Please make sure the current is within the safe operation limit of the power board and motor.

- I can try to increase Speed Feedback filter time constant, currently at 0.2ms. If I increase, this will add some delay on speed loop and can make it instable. Is 1ms seems to you a good value?

It depends on your application speed.

*I think I will try these solutions in the last way.

 Yes, Please try these and let us know.

Thanks,

Krupashankar

[AG_Serad]

It is the ratio of Line to line inductance(L) not Ld it is  (L/R).  The calculation is correct, please make sure correct Line to Line Inductance is used for calculation.

In fact, I used an LCR meter. I place the LCR meter between tow of three motor phase. I read the measured L value. WHen I turn the rotor, the L value change. I get the minimum value and the maximum. = > Lmin, Lmax.

Then I confiured:

Lq = Lmax/2

Ld = Lmin/2

I'm right?

If I understand what you said, and if my calculations above are right, I should set :

1) Flux estimator Time constant = 4 or 5 x Lmin/R ?

or

2) Flux estimator Time constant = 4 or 5 x Ld/R?

Thanks,

[Krupashankar]

Hi @AG_Serad,

You could choose the average line to line inductance, Lavg =( Lmax + Lmin ) / 2

Based on this Lavg = 10.5mH 

Tau = L/R = 10.5 / 5.8 = 1.81 mSec 

Ques 60 : Flux Estimator Time Constant = 5* 1.81 = 9.05 mSec

Thanks,

Krupashankar

[AG_Serad]

Hello,

My measure with LC meter @1KHz was:

Lmax = 25mH

Lmin = 17mH.

R = 11.6 ohms

It is why I configured:

Lq = Lmax/2 = 25/2 = 12.5mH

Ld = LMin/2 = 17 / 2 = 8.5mH

Rs = R / 2 = 11.6 / 2 = 5.8ohms

If I understand, we should use the line to line inductance, in other words, the inductance measured between two phases, same for Resistor:

I get the means inductance : (25mH + 17mH) / 2 = 21mH.

Then Tau = L/R = 21mH  / 11.6 = 1.81ms. => Ok, I'm agree with you.

Bellow are the measured values of the motor I have, according LCMeter frequency:

Which fequency must we use? Motor PWM frequency is 15Khz? must we keep values measured @ 1Khz or try measure @ 15KHz?

Remarque: measured Resistor @ 10KHz is strange!??

[Krupashankar]

Hi @AG_Serad,

Which fequency must we use? Motor PWM frequency is 15Khz? must we keep values measured @ 1Khz or try measure @ 15KHz?

---

You should keep the values measured at 1Khz. You should not consider PWM frequency it depends on the electrical frequency of the motor that varies due to speed. You could follow this document How to Measure Motor
Parameters 

Remarque: measured Resistor @ 10KHz is strange!??

---

This depends on your LCR meter specification and the reactance is a function of frequency, Please refer to your LCR meter datasheet.

Thanks,

Krupashankar

[AG_Serad]

Thank for the documentation "How to Measure Motor Parameters"

It is very intesresting:

-For R measurment => OK, it is what I done

-For L measurement => Ok, it is what I done, just the L value change acording motor angle then I indicate the Lmin identified as Ld, and Lmax as Lq...I hope this is correct?

- For poles paires: Ok I measure 8 positives peak, 8 negatives peak when I move motor by TWO mechanical turn. => this correspond to the motor DataSheet: 8 Poles.

- For Ke. Until today, I used the Motor Datasheet. :

Ke = 0.61 V/ rad/s

I convert it:

0.61 x 2 x Pi => V /rps

/60 => V / rpm

x1000 => V/ krpm

/Sqrt(3) => Vrms / krpm

This gave me 36.88Vrms/krpm. It is this value I entered on MCEWIzard parameter 7.

Now I applied the method given on you documentation.

The sinusoidal has a voltage Peak to peak of 76mV

The time between 2 positive peak was 156ms

The formula given on documentation:

Ke kRPMlnrms = Vpp/F  x Npolepairs x 3.40.

Like I have 8 poles, I have 4 pole pairs.

Ke = 0.076 * 0.156 * 4 * 3.4 = 0.1612416.

Your documentation give unit: kPRM ln rms.

MceWizard ask : V(ln-rms)/krpm

Please, according all theses information, which value must I fill into MCEWizard? 36.88? 0.161? other?

Thank

Edit: I made a mistake, mys scope probe was bad set with x10.

New measure give:

Vpp = 15V

T positive peak to positive peak = 184ms.

This give: 15 x 4 x 3.4 x 0.184 = 38.64... This correspond to the theorical value given by the manufacturer Ke=0.61 Vs/rad = 36.88 Vrms/krpm.

[AG_Serad]

Hello

We must trying soem modification done to say if problem was resolved or not.

- ISduty pass from 5% to 50% (original value calculated by tool)

-Flux estimator time constant set to 9ms

- Low speed current limit set from 30% to 70%

I must check current regulator bandwith. My fan has big inertia I think. THen, I will try to apply what is said into this document:

https://www.infineon.com/dgdl/Infineon-High_Inertia_Load_drive_for_iMOTION-ApplicationNotes-v01_00-E...

In any case, we must let try many week to be sure problem desapears or not. 

Currently, we are facing another worrying problem related here:

https://community.infineon.com/t5/iMOTION-controller-driver-IPM/IMC102T-Shunt-PFC-resistor-explosed/...

Thank for your support.

[Krupashankar]

Hi @AG_Serad,

Thanks for the feedback.

We are working on this thread : https://community.infineon.com/t5/iMOTION-controller-driver-IPM/IMC102T-Shunt-PFC-resistor-explosed/...

Thanks,

Krupashankar