SMO / eSMO HVPMSM_Sensorless

A GIR

Hello,

I'm working with a f28377 microcontroler to control a PMSM motor.

We use many feedback like resolver, serial, QEP... Exactly what is done with the TMDSIDDK_v2.0

Now we want try a solution without any feedback.

To this, TI provide an exemple HVPM_Sensorless which use the SMO library. On this forum, I found the eSMO library compiled for the f28335 with FPU32.

I tested both and compare with motor angle provided by the already implemented feedback.

I found many issue which is bloquant for us:

- In reverse sense, the estimated angle SMO/eSMO has shift of 0.5.

- On motor deceleration, (so brutal then motor current pass trough 0 and motor can become like a generator) the angle switch to a very bad value.

- When motor running, I try to switch from the real feedback to the estimated feedback, this run but it is very not stable. As soon as we apply a torque on the motor axis the control Loop come crazy.

SMO / eSMO give same result.

Please is my observation are the convenient of this method or Can I change something to perform this algorithm?

over 6 years ago

0 Ramesh Ramamoorthy over 6 years ago

TI__Mastermind 27240 points

In reverse rotation, the shift of 0.5 was missing in original SMO and that was one of the fixes we did on eSMO. It has been years since eSMO release and have not received a comment like this. Pls verify again. I have nothing else to suggest.

0 A GIR over 6 years ago in reply to Ramesh Ramamoorthy

Genius 3615 points

Hello

Yes you are right, the shift of 0.5 in reverse sense is corrected on eSMO.

But what about the stability?

0 Ramesh Ramamoorthy over 6 years ago in reply to A GIR

TI__Mastermind 27240 points

No issues with stability, it should work well.

0 A GIR over 6 years ago in reply to Ramesh Ramamoorthy

Genius 3615 points

Which way do you advise me to try to improve the thing?

Which parameter can be modified to improve performance?

How to deal with the problem of fast deceleration or a very weak current or engine stopped?

Thank

0 Ramesh Ramamoorthy over 6 years ago in reply to A GIR

TI__Mastermind 27240 points

With eSMO, we have not seen any stability issues with or without load. With SMO, of course, there were some challenges with loading, which is all resolved in eSMO.

The same is true for reverse direction as well.

There are some issues at low speed (<10% rated speed) when the angle estimation becomes sensitive to parametric errors in motor model. eSMO is recommended for speeds above 10% rated speed.

To your query on parameters to trim to fix stability / performance issues, I don't know if it is the choice of PI cooefficients in the current loop or speed loop. May be you need to look into that.

0 Ramesh Ramamoorthy over 6 years ago in reply to Ramesh Ramamoorthy

TI__Mastermind 27240 points

If your issue is resolved or if there is no more queries, let me know if we can close the thread. Thanks.

0 A GIR over 6 years ago in reply to Ramesh Ramamoorthy

Genius 3615 points

Hello

I tried eSmo provided into the above link (eSmo.lib compiled with FPU float32) and use it on F28377.

My Hardware is equivalant to the board given for TMDSIDDK_v2.0 (with ControlCard F28377 and MainBoard IDDK R2.2

http://e2e.ti.com/support/microcontrollers/c2000/f/171/p/434527/1578710?tisearch=e2e-sitesearch&keymatch=esmo%20float#1578710

1) Is the exemple given into HVMotorCtrl+PfcKit_v2.1\HVPM_Sensorless_2833x could working same with my hardware?

2) What does mean the value of "smo1_const.Vb" set to BASE_VOLTAGE:

#define BASE_VOLTAGE 236.14 // Base peak phase voltage (volt), Vdc/sqrt(3)

My application is generated from Monophase 230V, then the Continus DC-Link is 325V.

Which Value should I set into Vb? (325/sqrt(3) ?)

3) Which is the scale of volt1.DcBusVolt?

1.0 = 236V fot eh Exemple? Then Which should be the value in my application according answer to the point 2)?

4) smo1_const.Ib = BASE_CURRENT;

Same as Vb, which is the aim of this value? In my application, I get a value of +/-1.0 for +/-10A in one motor phase. I Can measure up to 16-17A with my hardware.

Which scale must I use into

clarke1.As
clarke1.Bs
?

Thank

0 Ramesh Ramamoorthy over 6 years ago in reply to A GIR

TI__Mastermind 27240 points

Some one will get back to you when available.

0 Ramesh Ramamoorthy over 6 years ago in reply to A GIR

TI__Mastermind 27240 points

1. The concept is same, but code may need porting over as the processor clock speed and ADC are different.

2. Base voltage is the dc bus voltage when the ADC reads maximum value representing Vdc. It depends on the hardware. For the HVMotor kit, it is about 408V, and in the software, it is treated as if 408 / sqrt(3).

3. It depends on your hardware platform, and specifically the voltage feedback scaling.

4. same as 3, but depends on current feedback sensing.

0 A GIR over 6 years ago in reply to Ramesh Ramamoorthy

Genius 3615 points

Hello

Ok, This work correclty now.

I try via other post to obtain the HFI library.

Thank for your support

https://e2e.ti.com/support/microcontrollers/c2000/f/171/t/781426

C2000™︎ microcontrollers

C2000 microcontrollers forum

SMO / eSMO HVPMSM_Sensorless