Dear all,
I have a question regarding the LOL signal of the the 8T49N241 frequency translator which can be output via GPIO3. When using the device in synthesis/free-run mode without using the digital PLL, does this signal reflect the lock status of the analog PLL? The datasheet clearly states that the signal reflects the signal of the digital PLL but what confuses me is that for the Xilinx ZCU104 board, the signal is directly routed to the FPGA fabric and used as the "tx_refclk_rdy" signal for the video phy controller. When only using the TX capability, the 8T49N241 device is configured in free-run mode and, according to the datasheet, should therefore not generate a LOL signal.
My second question is about the DBITM setting in register 0x0069. According to the datasheet, this bit should have a default value of 0. In my case, it is 1 after reset. What does this setting (as well as VCOMAN and DBIT) do?
Best regards,
Jelle
Hello Jelle,
I am going to confirm a few things with our engineering team and get back to you. Would you be able to share a part of your schematic that shows this device. ( TECHSUPP-7904 )
Hi Michael,
Thank you for looking into this. Unfortunately, I cannot share the schematic directly since I am using a commercial FMC -> HDMI board but the schematics seem to be based on the ZCU104 design which can be found here: https://docs.amd.com/api/khub/documents/g76en4BfB6qIp2gHDRkQxg/content?Ft-Calling-App=ft%2Fturnkey-portal&Ft-Calling-App-Version=4.3.55#G5.602649 (ZCU104 Board User Guide (UG1267), Figure 3-19).
Can you provided your configuration information, please this will help us determine if there are any possible issues.
If you don't want to provide that publicly you can send me a private message or open a Support Ticket
Hey Michael,
I am happy to share my configuration but I don't see how it is related to my general questions:
- Can the device generate a LOL signal on GPIO3 when in synthesizer mode?
- What are the DBITM, VCOMAN and DBIT settings and what could be the reason for DBITM being '1' after reset despite its default value being '0' according to the datasheet?
Please let me know if you still need my configuration.
Best,
Yes, please do.
this is my configuration:
0A 50 00 60 67 CD 7C 01 03 00 31 00 00 00 00 0000 07 00 00 07 00 00 77 6D 06 00 00 00 00 00 FFFF FF FF 01 3F 00 30 00 08 66 66 00 01 00 00 D008 00 00 00 00 00 08 00 00 04 00 22 22 24 22 0200 02 02 00 02 00 00 0D 02 00 02 02 00 02 02 0002 00 00 02 02 00 02 00 00 00 00 00 00 00 00 0000 00 00 00 00 00 00 A0 89 0F 2B 20 00 00 00 0B04 01 FF FF 01 FF FF 00 00 00 27 00 00 00 00 0000 00 00 00
Jedi,
Can you tell me the exact part number you are using - we are wondering if you are using a (what we call) Dash code. (Example: 8T49N241-998NLGI ) Since these parts can be customized and factory programable.
Unfortunately, I cannot be sure since I am using the Alinx FH1159 board (https://www.en.alinx.com/Product/FMC-Cards/FH1159.html) which does not state any dash code but I asked their technical support.
Dear Michael,
the full part number is 8T49N241-998NLGI.
The configuration that you provided in the form of register values, are these values what the you are writing to the device or reading from the device? Or is it both, i.e. they write it and read back identical values?
I took this register sequence and manually set the register values based on the provided values.
The resultant configuration raises several errors in the Timing Commander GUI.
GPIO3 output function = 0x4, which is a reserved value. It is possible that the customer configuration was created using a older personality in which this value was not reserved. I would need the TCS file to confirm this
The VCO frequency (4744.39678125MHz) violates the valid range of 2999.997MHz to 4000.004MHz.
We compared the register contents of 8T49N241-998 against the register contents provided by you. There are 32 registers (about 25% of registers) that have differing values between the configurations. Of these 32 registers are the registers controlling the VCO frequency, the GPIO3 output function, DBITM, VCO_MAN, and DBIT. Note, DBITM = 0 in 8T49N241-998, so it is this sequence is setting it to 1.
With this many differing registers and the dash code number of 998, we can only assume that you are trying to program only select registers. This is typically okay, however, they would need to ensure the written values are valid (i.e. VCO frequency and GPIO function).
Are you programming the device yourselves?
If yes, what is it that they are trying to achieve with their configuration. We can help provide a correct write sequence.
You are right, I set individual registers (and then read back the written value to make sure that the write was successful).
Something seems odd. When I read the previously posted values into IDT Timing Commander V1.7.3, everything is OK. GPIO3 output function is defined in registers 0x34, 0x35 and 0x36 which are 00, 00, 08 in the memory posted before. This means GPIO3SEL = 1 (and the other GPIOSEL values = 0). The reason your VCO value is too high is that my board uses a 40 MHz crystal which differs from the 48 MHz default setting.
I am trying to generate a 148.5 MHz LVDS clock on clock output 2 using a 40 MHz input crystal.
From our Engineering team:
The customer is correct regarding GPIO3SEL, I made a typo. The XTAL frequency of 40MHz explains the VCO issue as well. The XTAL frequency is a meta field in GUI and as a result was picked up from the register writes provided.
It sounds like the customer has a decent understanding of the device and their configuration.
If the customer would like, I can review their configuration if they provide the .tcs file. Otherwise, I don’t believe that there is any open action items on this ticket?