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.

DS90UB954-Q1EVM: connecting 954 & DS90UB953-Q1EVM over CSI-2 only

Part Number: DS90UB954-Q1EVM

I am using a 954 eval board, a 953 eval board, and a custom connector from J24 on the 954 to J1 on the 953 to connect the two over CSI-2/MIPI. I am controlling both devices over AnalogLaunchPad via USB, but the link is not passing (I am checking both "Current CSI TX Status" and "Current RX Port Status" on the 954 under Info and neither register a pass) even when the devices are both connected and powered. The link does pass when I change some jumper settings and connect the two via FAKRA cable/over FPDLink so I know at minimum both devices are working. I've checked my custom connector several times, and each data line and both clock lines are connected to their exact counterpart on the other eval board.

Essentially, I'm trying to just verify CSI-2 link from the 954 --> 953, to ensure that it works before creating a more complex setup, and I have not been able to see passing link on the CSI TX port.

More information about this setup:

-Both eval boards are being powered independently with 12V, and I can correctly measure some of the lower voltage rails with a multimeter (3V3 & 1V8 on the 953, 5V on the 954).

-The custom connector only connect the CSI2 clock and data lines, it does not pass power to the 953.

-I am using 800Mbps and 3 data lanes. I have also tried 2 and 1 data lanes, these don't work either, as I would expect.

-In AnalogLaunchPad, I first turn on the CSI2 TX 0 checkbox under Forwarding on the 954, then I turn on RX port forwarding for the port I'm using (0). The order for these instructions were found in the datasheet.

-The jumpers are set to NOT be in power-over-coax mode (since they're powered independently), but neither mode allows link to show as passing on AnalogLaunchPad.

-The jumper to put each device in CSI mode is set correctly on both eval boards.

-When I disconnect the clock lines from the 954 and send them directly into an oscilloscope via SMA cables, I see two flat 0V lines before turning on CSI-2 and port forwarding. Once I do both of those things, the P line of the clock rises to 1.2V, and the N line stays at 0V.

-I heard from a coworker that the MIPI lines might not work without an FPDLink connection to the 953, so I hooked up an additional 954 deserializer with a FAKRA cable, power and correctly set jumpers to the 953 (This setup looked like 954 ---CSI2---> 953 ----FPDLink--> a second 954). This did not change the link between the original, CSI2 connected, 954/953. 

  • Hi Emma,

    Do you use pattern generation in the 954?

    Gr, Rob

  • Hi Emma,

    I am slightly confused by your terminology since you point the arrow in the direction from 954 to 953.  In your setup the CSI data should be input at the 953 serialized and sent to the 954 where it is deserialized and sent via I2C to some SoC.  If you cannot get the devices to LOCK then the first thing I would check is whether each device is powered and what strap mode they are in.  The 954/953 product pages have users guides for their respective EVM which show how to set the clocking mode.

    Regards,

    Nick

  • Hi, if you read my post, it says that I have already checked power and strap mode on both devices and they are correct.

    What I am trying to do is set up a first link between a data source and a serializer (the serializer will then connect to a deserializer, but that connection works just fine), so that I can then insert and test a CSI2 isolation scheme between the data source and serializer.

    I am hoping to use a deserializer as a CSI data source by turning on pattern generation. If this is not an appropriate use case for the 954, then let me know.

  • Hi Emma,

    I see what you are trying to do now.  You are fine to use the 954 PATGEN as a CSI source to the 953 as long as the data/clock lanes match up correctly. Are you able to share a register dump of the first 954?  It sounds like the CSI port is not active and in still in an LP state.

    Also can you share the screenshot you mentioned from the scope I want to see the state of the CSI port.

    Regards,

    Nick