LV04EVK01 Evaluation Kit Does NOT lock to 74MHz LVDS Clock + Data DS92LV0421 DS92LV0422

Hi Jaimy,
Do not understand the R6 you mentioned. If possible, please send us the circuit diagram on how you connect DS92LV0421 and DS92LV0422, and let us know what is the cable loss (or cable type and length) that connects between the SER and DES.
Please also provide scope picture at DOUT+- (before the cable) and RIN+- (after the cable) measured with a differential probe.
regards,TK Chin

Hi Jaimy,

I looked at the EVK schematic for the Tx board. If the scope shot across R6 on the Tx board (100-ohm for RXCLKIN+/-) is clean, then the CML output should be fine. As T.K. mentioned, please send us captures of DOUT+/- and RIN+/-. Are you using a USB-to-miniUSB cable to connect Tx to Rx.

Another request. Could you let us know what your jumper pin settings are for Tx and Rx board and whether any LEDs on the Rx board are turning on?

Thanks,

Michael

Hi Jaimy,

I have recently been contacted by our local sales rep James Kay with some more details of this issue. For the benefit of anyone else that may be incurring the same issues, I'm including the information in this post as we work to figure out what is going on.

==================================================

In some of the boards before the most recent build, we found that there have been some incorrect silkscreen labels on the serializer and deserializer board. If the silkscreen labels have not been corrected yet, then please refer to the following directions.

DS92LV0421 Serializer Board:

The position of the jumpers should be set up as shown below.

  

Config[0:1] labels should be swapped as shown below, and this affects the ability of the SerDes to lock.

 

DS92LV0422 Deserializer Board:

The position of the jumpers should be set up as shown below.

The incorrect labels are detailed below:

 

I remember after setting up the board with the jumpers in the places shown above, things began to work and the “Pass” and “Lock” Orange LED on the DS92LV0422 board turned on when operating at Clock = 50 MHz. It should work with 75 MHz as well.

==========================================

I've also received the waveforms you are seeing (from James).

DS92LV0421 Input Differential LVDS Clock:

DS92LV0421 Tx Output (DOUT+/-):

I expect to see at minimum a distinct "01" in the eye monitor to indicate a low "Stop Bit" and a high "Start Bit," followed by 28 scrambled bits, which I am not seeing here.

Some thoughts:

1. Can you send a photo of how and where you are probing the DOUT+/- pins on the Tx board and what the termination is on the scope/probe you are using? Is it a differential probe with 100-ohm impedance?

2. Have you tried running at a lower clock frequency, such as 40 or 50 MHz? Any difference?

Thanks,

Michael

Hi Jaimy,

I received the setup configuration and the waveforms from James, which I am posting below:

Setup:

Tx Board:

Rx Board:

DS92LV0422 RXIN at 30 MHz:

DS92LV0422 RXIN at 50 MHz:

DS92LV0422 RXIN at 70 MHz:

DS92LV0422 RXIN at 75 MHz:

I am seeing signs of under-equalization on the incoming RXIN signal, since there are faint lines for long-transitions that reach the +/-400 mV amplitude, but then at higher speeds, that amplitude starts to close.

I have a few thoughts about this:

1. Since you are using a USB-to-miniUSB cable, it is possible that the cable quality may be limiting the operating frequency, especially when you measure at the Rx board side. Can you probe the +/- nods of P1 on the Tx board to ensure that the CML DOUT+/- signal is clean at 75 MHz clock?

2. There may be a stub effect due to the presence of R8 and R9 on the Tx board that is creating an alternate path to SMA connection options. I picked up a LV04EVK01 board in our lab and saw that these pins appear to be populated already. Having a stub can reduce the effective operating frequency by 5-10 MHz.

3. Try using the de-emphasis signal-conditioning feature in order to improve the signal quality that you measure at RXIN. This can be done by applying a shunt jumper on JP17 of the Tx board and then varying the potentiometer resistance on VR3.

4. USB cables are typically designed to 90-ohm differential impedance. The DS92LV0421/22 is optimized for 100-ohm differential impedance. For better signal integrity between the two boards, you should find better performance if you solder on the SMA connection options (J7, J8) on the Tx board and connect from DS92LV0421 to DS92LV0422 with matched 50-ohm cables. Doing this would also required that you remove R26 and R27 on the Tx board, remove R24 and R25 on the Rx board, and then populate R3 and R4.

Thanks,

Michael