Rotor Angle measurement from QEP through TMS320F2812
Altair EmployeeSubmitted by rajinmlinus on Thu, 02/06/2014 - 11:57
Dear Sir/ Madam
I am using TMS320F2812 dsp Processor. To get the rotor angle i have used QEP encoder and the QEP encoder will produce the pulse of 2000 per cycle. In visim i am using F281x quadrature encoder. please any one explain how we can get the rotor angle from the block and i need to give this rotor angle for the DQ to ABC conversion of the PMSM stator current. I have gone through the example available. I am in unable to understand.
Also, can any one tell me in worked example for one cycle how much pulses they have considered.
Answers
-
Submitted by Anders89 on Sun, 02/09/2014 - 11:21.
To use the angle in DQ to ABC (Park) transform as they are provided by VisSim/ECD, you must normalize it to unity scale, so multiply it by 1/(encoder count). This normalization to unity is also required for the QEP Speed calculator. See attached diagram below. (you must be signed in to see it).
0 -
Submitted by rajinmlinus on Tue, 02/11/2014 - 10:59.
Thank you sir for your reply. My doubt is the processor frequency is 4 times faster than the outside frequency. If the encoder pulses are 2000 counts for one cycle we have to multiply the output by (1/8000)=0.000125 right?. This is happening in the file pmsm31B2.vsm. If i am wrong please correct me.
Rajin M Linus
0 -
Submitted by Anders89 on Tue, 02/11/2014 - 11:22.
Not sure what you mean by the processor frequency is 4 times faster than the outside frequency. After normalizing the encoder counts you must multiply by the number of pole pairs and do modulo 1 so that you get a complete angle 'revolution' for each pole pair. So if there are 2000 counts per mechanical revolution, and there are 4 pole-pairs of magnets, you multiply the encoder count by 4/2000 = .002, then you take modulo 1 to get 4 mini-rotations per mechanical rotation.
0 -
Submitted by rajinmlinus on Sat, 08/02/2014 - 18:27.
Dear Anders Is the attached file is for the encoder pulse of 2000 counts? If so can you tell me how the theta offset is -1 in the attached file. Thank you
Rajin M Linus
0 -
Submitted by pete on Tue, 08/05/2014 - 20:13.
Rajin, in the file 'PMSM31B2.vsm' that was attached previously, in the quadrature encoder normalization logic, there is a fixed-point sum of -1 where the input has been negated. This is fed into a merge block that tests to see if the angle is negative and adds one if it is. The reason for subtracting -1 instead of simply adding 1 is due to the fx1.16 fixed-point format being used. Since there is only a single sign bit and the rest are fractional bits, you can represent -1 but not 1 due to two's complement arithmetic The conditional addition of 1 is only needed to account for use of a negative angle offset, in the PMSM31B2.vsm diagram the angle offset is not used, so you can remove that logic if you want.
0 -
Submitted by stenod on Fri, 01/29/2016 - 16:21.
In vissim source code file, I am placing eQEP1 (embeded>F280x>)block. What should be its(eQEP block) parameter setting(count mode=? position reset=? etc) . and how will i see output of eQEP in debug file. Shall I connect any block at QPOSCNT1 terminal of eQEP block in source code. Please suggest
0