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.

DS90UB948-Q1: MAP TOOLS testing result

Part Number: DS90UB948-Q1
Other Parts Discussed in Thread: ALP

Hi Team,

customer is using our MAP TOOLS on analysis 948.

in below picture, should i select the I2C address? for now we selected 0x64

 

in testing result, what does the SP and EQ in Coordinate System? whether EQ is needed to set through software? How could we adjust?

i attached the testing result as follow and could you help check the result?

eye-diagram-ALP.rar

BR

Brandon.

  • Brandon,

    >>>in below picture, should i select the I2C address? for now we selected 0x64

    Info tab should automatically find the 948 device and update. Scan connected devices allows user to see all the devices on the i2c bus. In case customer design has mulitple 948s on the i2c bus, then he can choose a different 948 device than the one ALP detects.

    >>>in testing result, what does the SP and EQ in Coordinate System? whether EQ is needed to set through software? How could we adjust?

    SP is Strobe Position which selects when the data is sampled. EQ is the equalizer gain. You can check the following app note and user guide as background. Even though it explains with 954 it applies equally to 948 as well.

    www.ti.com/.../snla301

    www.ti.com/.../snlu243

    >>>i attached the testing result as follow and could you help check the result?

    What's the cable length you are testing with? All green doesn't look right. Is this with EVM or customer board? If customer board, could you make sure CPU on the board is not initializing 948? We want ALP to control and run this diagnostic test without any external CPU or MCU intereference.

    Could you also make sure you have the latest ALP-PROFILE-UPDATE installed from following link

    https://www.ti.com/tool/ALP

    You can confirm your ALP is up to date by selecting ALP >> Help >> About this Program

    Date should match the latest date of the ALP-PROFILE-UPDATE (v08) version on the web.

    Thanks,

    Vishy

  • Hi,Vishy

    The cable length is 1 meter,it was tested on our TD-TECH board.

    There is an MCU connected to 948 through I2C, and Now I disconnected this MCU.

    941 is connected to a SOC that provides image data. 941 is initialized once when it is powered on, and will not be initialized later even if 941 and 948 lose connection. The test results are as follows:

    PORT0:

    PORT1:

    ALP version:

  • Hello,

    Thank you for updating to the latest ALP.

    I was looking at some Map results I had captured with our EVMs (948/949) for 1m. See below

    Above result is obtained by running Margin Analysis Script. The script version is the same as the ALP integrated GUI version except the results are output in a CSV file. You can find the script as part your ALP installation folder (PreDefScripts).

    >>>>941 is connected to a SOC that provides image data. 941 is initialized once when it is powered on, and will not be initialized later even if 941 and 948 lose connection

    I think it would get better if we disconnect this SOC i2c lines also from 941, so there is no interference in the margin analysis scan

    Thanks,

    Vishy

  • One more item: when you run the margin analysis, please make sure strobe setting begin/end selection is 0-9 and EQ level begin/end selection is 0-14. 

    Thanks,

    Vishy

  • Hi,vishy

    Thanks for your advice.

    • The following two graphs are CSV files corresponding to yesterday's test results

    • port0:

    port1:

    The connection state before disconnecting I2C between 941 and SOC is as follows:

    • The connection state after disconnecting I2C between 941 and SOC is as follows:

    if strobe setting begin/end selection is 0-9 and EQ level begin/end selection is 0-14,most of the time the results are as follows:

    • Only one test result was as follows:

  • Hi Michaux,

    a) Let's debug only running Port 0 to start. 

    b) When you disconnect SOC, could you make sure PCLK is still available. You are losing the link because there is no PCLK. 

    Please check how is PCLK generated and provided to the serializer in your setup. Ensure PCLK and link working and then run the margin analysis.

    Thanks,

    Vishy

  • Dear vishy,

    a)I find that now Port0 tests successfully every time, but Port1 tests red every time.Is the test result of PORT0 ok?How do I measure port1?

    PORT0:

    b)When I disconnected the 941 from the SOC's I2C, the PCLK was fine, but the 941's LVDS input was gone。

  • Michaux,

    Port 0 result looks much better now. I think we use DACAR535 cable. Please check with quality cable.

    Regarding Port 1, are you able to see through remote registers window port1 device and its registers? Please check the connection is good.

    - In registers tab, make sure port1 is selected on the top right

    - Check in the serializer id register (0x5B) correct address of the remote serializer (should be 0x84 in your case, 0x34 being port 0)

    - Read from the scripting tab address 0x0 of the remote serializer and get valid remote address

    - in remote registers tab, make sure correct port on the top right is selected and you are able to see the registers

    >>>>>b) When I disconnected the 941 from the SOC's I2C, the PCLK was fine, but the 941's LVDS input was gone。

    No input needed to run margin analysis. We are checking the quality of the link between the serializer and deserializer, so image input or patgen or bist not necessary.

    Thanks,

    Vishy

  • Hi,vishy

    By setting bit2bit1bit0 to 011 for register 0x5B, I can now measure Port1. The I2C of the 941 is also disconnected. The following is the test result. Please help to check whether it meets the requirements.

    port0:

    port1:

  • Hello Michaux,

    Thank you for the update. The port1 result looks very good and almost perfect. Port 0 result is ok but not ideal as channel 1 and showing more insertion loss. Is the port 0 cable same like Port 1? May be you could try with a different cable (or the Port 1 cable) and also compare/review your setup  & layout between port 0 and port 1.

    Thanks,

    Vishy