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.

decreasing the lock time on TV5150AM1

Other Parts Discussed in Thread: TVP5150

I have a camera head containing 3 separate cameras.  So there are 3 video streams that go through and analog mux in the camera head and go back to the TV5150AM1 over a single input. I am not , muxing between the two inputs of the 5150.  The 3 video streams are all gen locked together from a common clock source. But there is as much as 5mS difference between the gen lock signals.

 

Under the current scheme, the analog mux can switch between video signals ANYWHERE within the frame.  I am finding is that the 5150 can take as long as 200mS to lock on which is causing some problems.  I would like to get this under 100mS.  Can the group shed some light on what I MAY be able to do to lower the lock time. 

 

<My original email>

  In particular, we need some help with the  TV5150AM1 which we are using in our new design.   In that camera head, up to three PAL video signals are multiplexed onto a single video cable which is connected to channel 1 of the TV5150AM1 video decoder.  All three video signals are genlocked together.  What we found is when we switch between video signals inside our camera head, the  TV5150AM1 can take as long a 200msecs to lock before we can grab the digitized frame.  This is too long for our application.  In our previous generation product, we had used an video decoder which allowed us to do this.  Unfortunately, the new  TV5150AM1 doesn't seem to allow this fast locking.

 

Our firmware developer has already searched newsgroups and found some kind of register setting "patches" to be applied to supposedly speed up the horizontal lock time, but that didn't help.  We need some assistance from TI to reduce this lock time.  The three signals have the same offset, overall amplitude and are clock from the same source and phase locked.  The only difference would be the image itself which is different for each camera.

 

Any help would be appreciated.

 

  • Lock time could be influenced by the TVP autoswitch algorithm and also the AGC algorithm.  Could you try the steps below to determine best case lock time:

    Using cameras or signal sources with nominal video levels:

    1. Disable autoswitch by forcing the TVP to the video format being used.  Check lock time after MUX switch.

    2. After bringing up the system and acquiring initial lock, force the TVP to the format being used and then set the TVP to manual gain.  Check lock time after MUX switch.

  • 1.       Set the TVP in auto mode (both switching and AGC). Wait until TVP is locked. (No video source switching during this time)

    Write 0x29 to 0x03 register. (Miscellaneous Control Register, INTREQ/GPCL OUTPUT enable, YCbCr Enable, CLOCK Output Enable )

    Write 0xC0 to 0x04 register. (Enable AutoSwitch, Unmask all for autoswitch)

    Write 0x70 to 0x06 register. (Disable Color Killer)

    2.       Wait until TVP is locked

    3.       Write 0x04 to 0x28 register. (Set Video standard to PAL)

    4.       Write 0x17 to 0x01 register. ( Freeze the AGC gain to the previously set value)

    5.       Write 0xFC to 0x03 register. (Mask all video formats from AUTO switch)

    6.       Check Video Lock

     

    Now switch the video sources to calculate the lock time.

     

    Repeated these tests with and without TVP5150 patch firmware.  Lock time is always between ~ 240-260 ms.

    (The video sources are out of phase by ~ 4-7 ms)

     

    Repeated the test by adding a mode selection command (Set to TV mode).

    In all these tests, lock time is always ~240-260ms.

     

    In our previous experiments (we didn’t use any Manual Gain ) we have seen some improvement when two video sources are in phase (close to 1ms). We have seen lock time of 170-180 ms.

    Our goal is a lock time <100mS