DRV8312-C2-KIT: Motor works but gets hot

From this discussion: e2e.ti.com/.../717130

Can you post a picture of the waveform on the other BLDC controller? What you've shown in the previous thread for IB-FB and IC-FB is very noisy. It's possible there is some harmonic distortion that is adding to your overall power loss in the motor windings. What PWM frequency are you using? You may benefit from increasing it.

I'm surprised you're able to drive the motor well with such a noisy current waveform

Sean

Sean,

Here are the motor voltage wave forms .We have purchased this controller and don't have much information about it.

PWM frequency we are using is 45 Khz.

.'

Is 1 pole pair correct? Perhaps you can try increasing your PWM frequency to 60kHz to see if you have any improvement on the current waveform. Also, it appears you are using overmodulation and field weakening, correct?

Sean

Yes its 1 pole pair.I am currently using lab 10a but I haven't modified the modulation index or checked field weakening.

Secondly this motor is delta connected. Heres the info of the resistance.

Does Instaspin determine the bldc to be wye connected by default for calculation of Rs?

Please try limiting the modulation index 0.57 to limit the overmodulation to only ignore 1 phase instead of 2. If you are unable to reach the target speed at that point, you may consider enabling field weakening at that point.

Sean

Sean,

I tried lab 9 which is without over modulation and increased the PM freq to 60 kHz ,but its still the same.

Can you send me the waveform of the desired output.

 

The waveform is hopefully sinusoidal but without much noise. Have you made any alterations to the EVM? Does your trapezoidal controller drive the motor to the same speed as InstaSPIN?

Sean

No,we haven't made any alterations to EVM.The trapezoidal controller can drive upto 35K rpm while Instaspin can drive upto 38 Krpm.
If we drive the same motor at 30K rpm on both the controller ,there a temperature difference of about 8 degree Celcius. Its 8 degree higher using instaspin.

Perhaps the bemf profile of this motor is trapezoidal instead of sinusoidal? Can you show the bemf waveform? If so, the identified motor parameters may not be exact for InstaSPIN, which relies on sinusoidal bemf waveform for best performance

Sean

This waveform is using lab9

Top:IC-FB

Bottom: IB-FB

 

Here's the Back emf of the motor running at 15Krpm and 10Krpm respectively.

"the back EMF of a BLDC motor is not truly trapezoidal in shape. In reality, its shape is more sinusoidal. This is why BLDC motors can use trapezoidal or sinusoidal commutation."

I am still not sure if its sinusoidal or trapezoidal.

did you use proj_lab02c to ID the motor parameters? what are the USER_MOTOR settings you used?

I think the Ls value is likely too high effecting the estimator and the automatic setting of the current controllers. The current is not well controlled, and you can see the Id value is not being regulated to 0.0

have you changed any of the PI settings to the current or speed controllers?

did you use proj_lab02c to ID the motor parameters?
We used lab 2b to ID motor parameters.
what are the USER_MOTOR settings you used?
(Motorware v18)
#elif (USER_MOTOR == Mymotor)
#define USER_MOTOR_TYPE MOTOR_Type_Pm
#define USER_MOTOR_NUM_POLE_PAIRS (1)
#define USER_MOTOR_Rr (NULL)
#define USER_MOTOR_Rs (0.4278004765510559)
#define USER_MOTOR_Ls_d (0.0000011488098152)
#define USER_MOTOR_Ls_q (0.0000011488098152)
#define USER_MOTOR_RATED_FLUX (0.0129088694229722)
#define USER_MOTOR_MAGNETIZING_CURRENT (NULL)
#define USER_MOTOR_RES_EST_CURRENT (0.6)
#define USER_MOTOR_IND_EST_CURRENT (-0.6)
#define USER_MOTOR_MAX_CURRENT (3.0)
#define USER_MOTOR_FLUX_EST_FREQ_Hz (60.0)
#define USER_MOTOR_FREQ_LOW (58.33) // Hz - suggested to set to 10% of rated motor frequency
#define USER_MOTOR_FREQ_HIGH (583.3) // Hz - suggested to set to 100% of rated motor frequency
#define USER_MOTOR_FREQ_MAX (699.96) // Hz - suggested to set to 120% of rated motor frequency
#define USER_MOTOR_VOLT_MIN (1.8) // Volt - suggested to set to 15% of rated motor voltage
#define USER_MOTOR_VOLT_MAX (12.0)

I think the Ls value is likely too high effecting the estimator and the automatic setting of the current controllers. The current is not well controlled, and you can see the Id value is not being regulated to 0.0

As per Motor specifications:

Motor resistance is 0.78 ohms (Delta Connected)
And Inductance is 52uH. But GUI shows its 2.1758 uH.

have you changed any of the PI settings to the current or speed controllers?

Yes,I have referred the GUI Quick Start Guide: InstaSPIN UNIVERSAL for modifying the values.

 This is using lab 2c.Motor still gets hot and reach temp up to 51 C.

Yes,that was my mistake.I know we have to make the conversion as per the GUI Quick Start Guide.

There's definitely some improvement,but the Isc is still around 419 Amps.If we add inductance what values should we use.
Currently as per data sheet the motor Inductance is 40-52uH. If we check it using LCR meter it shows around 39-40uH,but on GUI its very low.Do you think adding an inductance of 25uH in series will reduce the Isc?

We can also try improving the quality of the motor. Do you think its just the inductance or even resistance should be improved and by what? I just want to know so that I have an idea while selecting the motor for future

it depends on what you want the motor to do.
it appears you have selected a low inductance, high speed motor typically used for "hobby" type applications

these are usually very cheap as they don't include many magnets and very few turns of copper wire.

With lab 2c, everythings fine .But in lab 10, I see Id value is non zero.it keeps fluctuating .I will try modifying the current and speed controllers tomorrow.

With lab 2c , I can achieve max speed of 30K rpm.Motors max speed is 37Krpm. I am able to reach speed upto 36Krpm in lab 10a. Do you think voltage scaling will help us achieve the max speed ?

Secondly what is  Id_A ? and how to reduce it 

that is what lab10 is doing, it is increasing the voltage that is put into the inverter.

but it makes it harder to sample the currents required for FOC.

Will hardware scaling help for this?

I don't understand what you are suggesting, so I will have to say no.

the problem with current sampling is related to the design of your motor and using a low side shunt technique

We have decided to contact the motor manufacturer to redesign the motor for a few samples.


Assuming,

L=75 uH

R= 0.05 ohms

therefore,

R/L= 666.66 Hz =40,000 rpm (Motor is rated for 37000 and has 1 pole pair so kept it slightly more)

Isc=Flux/L = assuming flux as per the current motor.(It might increase as we increase the windings)

0.0129 / 75uH
=172 Amps
actual value will be more that 172A due to increase in flux. (This is still large)

Am I correct in this calculation ?

Should We further increase the inductance to reduce Isc? How much impact will it have due to increase in resistance ?
and How max we can get with resistance values.

We need to give the parameters to the manufacturer, so want to get it verified once.
The values we decide wont be exact,but atleast it will better that the current one.

Please suggest.