• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 30 subscribers
  • Views 477 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

PGA411-Q1: Resolver Moving By Itself

Derek Cook
Intellectual 880 points
Part Number: PGA411-Q1

Hello, 

I have a problem where when I stop around 270, the resolver seems to move by itself. Looking at the troubleshooting guide, this seems common, and I should set Auto phase Config in DEV_OVUV4 to 0, which I have read and set to 0. This does not help me in my case. Actually, none of the settings for the phase config (Auto Open, Auto Short, or Manual) seem to work for me. 

Looking at the OSIN and OCOS pins, my signal seems to go from 100mV to 1.5Vpp. Looking in the datasheet I see the minimum should be 600mVpp, so I increased the COS and SIN gain from 1 to 2.25 (there are very little options when it comes to tuning this wide range). My exciter voltage was also distorted, so I lower my exciter preamp until the voltage was clean. The setting I finished with was preamp of 1.25 with 7VRMS. Table 1 in the datasheet seems not to recommend this setting? The voltage I read on the Excitation is now 5.2VRMS. My Output sin and output cos with these settings runs from 700mV to 2.5Vpp. This fixes the resolver moving by itself. That seems to be caused by the OSIN and OCOS being less than 600mVpp? 

My problem now is I am occasionally faulting on the signal integrity OMIZ faults because I am getting some of my 8kHz PWM motor control frequency on my resolver lines since my signal to noise ratio is worse now with the higher SIN and COS gain. The tuning range on these faults are already pretty loose, and the added 0.3V in each direction does not help. 

My question is, is there a better way to solve this problem using the settings in the TI chip that will not make my signal to noise ratio worse? It seems to me the only other option is change my resistors on the board to change the GAINSIN and GAINCOS?

Thanks for the help!

  • 0
    TI__Mastermind 23810 points
    Derek,

    The better solution is to adjust the AFE resistors to set your desired gain. This way, you can keep using the COS/SIN gain setting as 1.

    Table 1 in the datasheet has an issue which we will be updating very soon, so there is nothing wrong with what you chose.

    I don't understand your reported range of OCOS/OSIN amplitudes. The peak amplitude should occur at 0, 90, 180, and 270 degrees. At those angles, either OCOS or OSIN will be at the max amplitude and the other will be at 0Vpp. How are getting a minimum of 700mVpp? Unless you are referring to possible variation in the resolver sensor coupling?

    Also, what is your OSHORTH/OSHORTL settings? These settings work with the autophase algorithm when AUTOPHASE_CFG = 0. Usually, the default of 0 works well.
  • 0 in reply to Clancy Soehren
    Intellectual 880 points
    Clancy,

    In section 7.3.3 I see that the recommended differential voltage is between 600mVpp and 1.5Vpp as measured at OSIN and OCOS. Does this mean the minimum amplitude should be 600mVpp or should this always be 0Vpp? I am reading OSIN and OCOS to GND. It seems my minimum was increased to 700mV when I increased the gain. When I try to set OSIN and OCOS to 1.5Vpp, I see the resolver moving by itself. When I increase the AFE gain to 2.25V this goes away. My OSIN and OCOS is then 2.5V. Why would this cause the resolver moving by itself to go away?

    My OSHORTH and OSHORTL settings are set to 0. AUTOPHASE_CFG is also set to 0.
  • 0 in reply to Clancy Soehren
    Intellectual 880 points
    Hey Clancy, I am looking at another post: e2e.ti.com/.../2271070

    This is extremely similar to my issue. In this post you said OSIN and OCOS should be centered around 2.5V. This is not the case for me. My signal seems to be centered around 700mV. Would this be causing my problem? If so why? I am using the Auto Phase with Short Detection like you suggested in the other post. Just wanting to know exactly why this does not work.
  • 0 in reply to Derek Cook
    TI__Mastermind 23810 points

    Derek,

    That could absolutely be causing the issue. The autophase algorithm uses the zero crossings of OCOS/OSIN to implement the phase adjustment. In this context, zero crossing refers to the COMAFE voltage (typically 2.5V). If your entire signal is centered around 700mV, then the problem makes much more sense.

    If your signal is centered around 700mV, then there must be a DC offset between the differential input signals somewhere else in the signal path. Do you know where it is coming from? Are you using your own PCB or the EVM?

  • 0 in reply to Clancy Soehren
    Intellectual 880 points

    Clancy,

    I am using my own PCB. COM AFE is equal to 2.5V - I thought this was ground. There was a 700mV offset in my probe. Measuring OSIN and OCOS to ground now centers the signal at 2.5V. When the resolver is moving by itself, I have a 1.5Vpp signal.  

  • 0 in reply to Clancy Soehren
    Intellectual 880 points
    Clancy,

    I got it working. The problem was my OSHORT threshold was too low. The signal was not getting below the 125mV peak default at the zero crossing, so the AUTOPHASE_CFG was not working. I increased the threshold, and the resolver is no longer moving by itself.