DLPC230-Q1: OpenLDI timing

Which transmitter part number are you using or considering?  We need to find an expert on the FPDlink team that can help address this question and the others that you have.  The openLDI on the DLPC230/DLPC231 is a hard coded phy which is compatible to the standard, but we do not have detailed expertise on this team on the protocols.  Therefore, I want to move this question to the right team.

Which transmitter part number are you using or considering?  We need to find an expert on the FPDlink team that can help address this question and the others that you have.  The openLDI on the DLPC230/DLPC231 is a hard coded phy which is compatible to the standard, but we do not have detailed expertise on this team on the protocols.  Therefore, I want to move this question to the right team.

Thanks for the reply, we are considering to use SN65DSI83 as transmitter.

Our target is to generate the cable CLK to data skew, so we are calculating openLDI timing based on snla249.pdf method, this method use Rsposn, Rsposn definition seems a little different with Tskew of DLPC230-Q1, so want to know how to translate Tskew to Rsposn?

In other words, DLPC230-Q1 defined the input data position, not the receiver strobe postion (Rsposn), not strobe window mentioned as snla249.pdf definition, so can we think for DLPC230-Q1 openLDI timing margin = Tskew - Tx output data pulse variation - PCB CLK_to_data skew - ISI - input CLK cycle to cycle jitter? thanks.

Thanks for your quick reply! I have two further questions hope to discuss with you.

If,

transmitter data output pulse variation: +/-a

Clk cycle to cycle jitter: +/-b

PCB CLk to data skew: +/-c

Clk jitter due to ISI/crosstalk/reflection: +/-d

Data jitter due to ISI/crosstalk/reflection: +/-e

So the final margin=Tskew-a-b-c-d-e, is this right? 

Is it necessary to consider d and e? maybe this is too pessimistic to sum up d and e directly.

thanks!

Hi Jason,

Thanks for your help! still have several questions want to consult you.

1.Is tjitter required for cycle to cycle or peak to peak? As they are different meanings for jitter measurement. If it is for cycle to cycle, how this cycle to cycle jitter impact the timing margin for below timing diagram?

2.Is tip0~6 measured from data P/N cross point to the middle of CLK rising edge?

thanks

wenjun

Hi Jason,

can we use this equation to calculate the margin? margin = Tskew - a - b -c

• cycle to cycle jitter: +/-a
• data position variation related the current cycle CLK rising edge(right???) at receiver: +/-b
• CLK to data PCB skew: +/-c

And I am still curious why cycle to cycle jitter should be subtracted from Tskew. As from timing diagram, Tskew is only related with the current cycle CLK rising edge, not related with previous or post cycle CLK rising edge.

thanks

wenjun

Hi Jason,

Thanks for your reply and patience.

one more question: as below figure, is tip0~6 related with the present cycle clk rising edge or the last cycle clk rising edge ? if it is related with the last cycle clk rising edge, then this makes sense to subtract cycle to cycle jitter.

datasheet doesn't indicate the tip1 value, it should be equal to tskew, right?

thanks

wenjun

Wenjun,

I will need to ask this detail from the HW team to see if they know.  Please wait.

Jason

Ok, thanks!

I sent a note to the HW designer asking for more clarification.

Thanks a lot, I will wait for this information.

Should have something in a day or two

Hi Jason,

Many many thanks. 

Want to double confirm if the cause and effect (my understanding) is like below:

Cause1: As internally the 7x clock that is clocking the bits out will have +/-400ps jitter of the ideal for each bit time,

Effect1: so this make data transition point Tip1 can be up to +/-400ps relative to that clock edge(tskew).

Cause2:  as cycle to cycle jitter will impact 7x clock to sample the bits,

Effect2: so cycle to cycle jitter will increase or decrease tip1.

thanks

wenjun

Hi Jason,

sorry, still have one tip0~6 measuring point question want to double confirm with you: 

is Lx_CLK measured at middle of edge or Vih? Is Lx_data measured at P/N crosspoint or Vih/Vil? If measured at Vih/Vil, what's the value for the Vih/Vil?

thanks

wenjun

Hi Wenjun,

Please allow us a couple extra days to look into this.

Thank you for your patience,
Michael Ly

Ok, thanks!

Hi Jason,

it is clear, thanks a lot!

want to double confirm for VID definition of table 6.9, is VID definition same as IVodI relative to 0V? Or +Vod - (-Vod)=2xVod? as below figure, thanks!

ok, thanks!