DS90UB947-Q1: Schematic&PCB review

peiyq

Part Number: DS90UB947-Q1
Other Parts Discussed in Thread: DS90UB954-Q1,

Hi TI,

Our is using DS90UB954-Q1 in our project ,the serializer is DS90UB953 which is work in 4G FPD-Link PoC Network,could you check if the schmatic and PCB layout ?

Peiyuanqing

over 5 years ago

0 peiyq over 5 years ago

Expert 1400 points

DS90UB954-Q1 schematic.pdf

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Nile_MB_V07-TI.brd

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Hi TI，
Could you confirm the 49.9 resistor is closed to connector or 954 is best and RIN+single -ended 50Ω with RIN- should separated with Ground or just leave it blank space
Thanks
Peiyuanqing

0 Darryl Jeong over 5 years ago in reply to peiyq

TI__Expert 4210 points

I'm sorry, I'm confused. The subject is "DS90UB947-Q1: Schematic&PCB review" but your schematic is for a DS90UB954.
Anyway, in answer to your question:
On the FPD3 trace place the 50 ohm termination close to the connector; the AC cap should be close to the 954 input.
I noticed that C363 (going to RIN0+) you have selected 33nF, therefore C364 should be 15nF not 47nF. Also I noticed that CON2 has pin 5 as the center conductor, typically pin 1 should be the center conductor. You will need to check the CON2 symbol is correct.

0 peiyq over 5 years ago in reply to Darryl Jeong

Expert 1400 points

Hi Darry,
Sorry about confuse you ,the isusue is DS90UB954,Thanks for your support, C 364 have change to 15nf as you mentioned , CON2 is correct after we check it ,
1.i upload the new schematic and PCB ,could you check it again?
2.Regarding of the RIN+single ended 50Ω and RIN-, should we separated with Ground or just leave it blank space?
Thanks
Pei yuanqing

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

954 schematic.pdf

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Nile_MB_V07-20190312-TI.zip

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Hi Darry,
Could you check it ASAP,our customer is waiting the answer ,can i get review results before 3/14
Thanks
Pei Yuaniqng

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Hi Darry,
Could you check the IN+ trace is 36.1145 mm , 49.9 resistor is still necessary closed to connector?since this is very urgent.Could you tell us the email or phone number ?
Thanks
Peiyuanqing

0 Darryl Jeong over 5 years ago in reply to peiyq

TI__Expert 4210 points

Hi Peiyuanqing,
The recommendation is to place the 49.9 ohm resistor next to the connector and run the trace back to the IN- (not plus side as you are referring to IN+) as a 100 differential. The 49.9 ohm termination resistor you are referring to is "R6547" correct? This goes to "RIN0_M"

0 peiyq over 5 years ago in reply to Darryl Jeong

Expert 1400 points

Hi Darryl ,
thanks for your reply,
1.The 49.9 ohm termination resistor you are referring to is "R6547" correct
yes , "R6547" is 49.9 ohm termination resistor
2. we are using coaxial cables,could you check if it is can meet TI's layout rule,trace is in attachment pdf file
3. could you please double check the 49.9 ohm termination resistor closed to connector is better since the trace is 36.1145 mm long
Thanks
Pei yuanqing

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

in+&in-.pdf

0 peiyq over 5 years ago in reply to Darryl Jeong

Expert 1400 points

Hi Darry，
my Answer is：
1.yes，The 49.9 ohm termination resistor is "R6547"
2.i have uploaded PCB file in arobat file，could you check the route is meet TI‘s design guide？
3.customer is asking the in+ 36.1145 mm long ，the 49.9 ohm termination resistor is still closed to connector？
Thanks
Pei Yuanqing

0 Darryl Jeong over 5 years ago in reply to peiyq

TI__Expert 4210 points

Hi Pei,

The in+ and in- should be 100 ohm differential up to the connector so remove the ground trace between. Can they extract the trace loss (s-parameter) of the trace after the layout fix?

0 peiyq over 5 years ago in reply to Darryl Jeong

Expert 1400 points

Hi Darryl,
We had gerber out in attachment 50Ω seperated by GND since customer is pull in schedule，will test it around 4/5 .
Thanks
Pei yuanqing

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Hi Darryl，
Could you help check if single 50Ω have some side effect on CSI signal qualtiy？we meet CSI sometimes loseing one frame
Thanks
Peiyuanqing

0 peiyq over 5 years ago in reply to peiyq

Expert 1400 points

Hi Darryl,
Could you help check if single 50Ω have some side effect on CSI signal qualtiy？we meet CSI sometimes loseing one frame ,your reply is appreciated.
Thanks
Pei yuanqing

0 Casey McCrea over 5 years ago in reply to peiyq

TI__Guru* 98495 points

Pei,

The trace routing affects the high speed FPD Link path - not the CSI directly, but if the trace is not differential up to the connector, then it is easy for noise to impact the overall performance and cause bit errors which cause CSI errors. Our suggestion would be to revise the layout based on our guidelines in the 953/954 datasheets.

Best Regards,
Casey

Interface

Interface forum

DS90UB947-Q1: Schematic&PCB review