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.

DS125DF410: Stability of PRBS stream on free running VCO

Part Number: DS125DF410

Tool/software:

Hi TI Team,

We are using DS125DF410's onboard VCOs, dividers and PRBS generators to generate PRBS streams at various data rates (6Gbps, 6.25Gbps, 5Gbps, 10Gbps, 12Gbps, 12.5Gbps). Since the free-running VCOs may not be at the right frequencies, we made them right by adjusting the cap and lpf values. So setting the data output at the right rate is not an issue to us. We actually manage to use a 2nd board as a Rx to receive the signals and monitor the eye patterns of the incoming data streams.

However, we noticed that the 6Gbps and sometimes the 6.25Gbps data streams can be unstable as reviewed by the eye patterns (though they are fully locked at the Rx). All other rates are free from such an issue.

The issues happen repeatably to some channels of a Tx board. If a different Tx board is used, the channels that illustrate the issue can be different. We check the chip voltage and confirm that it's at 2.46V. It's further confirmed that the issue is from the Tx not the Rx, because the Tx channel with the issue will cause all the Rx channels to show the same issue on the eye patterns if the Rx channel is connected to that Tx channel.

Could you let us know how can we fix the issues?

Thanks.

Child

  • Hi Child,

    I'm looking into this and will provide feedback shortly.

    Best,

    Lucas

  • Hi Child,

    When you say a different Tx board is used, I am assuming the 2 boards are identical in design and have different DS125DF410 units installed? CAP DAC and LPF DAC values will vary part to part to achieve precise frequency. Therefore each unit needs to be tuned individually.

    I'm assuming the VCO frequency was measured directly to tune CAP DAC and LPF DAC values? Can you share VCO frequency waveforms? I'd like to check the frequency stability.

    If VCO frequency wasn't measured directly, can you explain how CAP DAC and LPF DAC values were tuned?

    Best,

    Lucas

  • Hi Lucas,

    You were right about the Tx boards. I was referring to the Tx boards with identical design (i.e. they were assembled physically with two different DS125DF410 chips). We understand that the CAP/LPF values need to be tuned on a channel/chip basis, as the VCO on each channel/chip may have different CAP/LPF values for a specific frequency.

    We don't have a scope with bandwidth high enough to let us see the 6Gbps tone directly. Rather, we observe the fundamental frequency of the PRBS-9 stream at 6Gbps with a spectrum analyzer. Based on the fundamental frequency, we can adjust the VCO to output the frequency we need. For streams with normal eye patterns, you will see a very clean single-tone spectrum. For the streams with eye pattern issues, you will see higher noise floor and some "overtones" around the fundamental frequency. 

    Child

  • Hi Child,

    Thank you for your clarifications.

    Unfortunately, CAP DAC and LPF DAC are the only variables which can be changed to improve the VCO frequency. If these values have already been tuned and the frequency stability is not sufficient, then this is a limitation of using PRBS generator with free-running VCO.

    If you are able to feed a sub-rate clock signal to the retimer RX, then the retimer can lock to this signal and the PRBS generator output will be synchronous to this signal. This is the recommended method to use the PRBS generator with precise frequency stability.

    Best,

    Lucas