SN65LVDM050-Q1: IBIS Model Simulation

Hi Josh, 

Thanks for the reply. 

Joshua Salinas said:

You mentioned "15 transceivers with last one about 20meter apart." May I ask how long is your LVDS bus length? 

The exact bus length is not yet calculated, but we can consider it will be 20 meters minimum. 

Joshua Salinas said:

The cable that I used in Hyperlnx was RG142B/U 83242. Unfortunately, I'm not aware of a good way to simulate a twisted pair cable like Cat5 in Hyperlynx

But is it  good way to consider two 50ohm cable with 100ohm termination as chosen in the circuit to try out with the differential pair cable simulation?. will it be good enough to behave as a 100ohm differential cable? 

Joshua Salinas said:

For multi-point applications, please consider using M-LVDS like the SN65MLVD203B you mentioned. That's the reason you see an improvement in signal integrity when using the SN65MLVD203B.

I am aware of this part Josh, but SN65MLVD203B is not an automotive qualified deceive  and  SN65LVDM050-Q1 is the only product with automotive qualified from TI which supports M-LVDS. This is the reason to try out with the simulation. 

Is there any other way or tool to check for the same? wanted to check till what distance SN65LVDM050-Q1  can be considered for 100Mbps and 50Mbps signaling rate and how the signal will be distorted at the end of 20meter cable length. 

Does you (TI) have s-parameter model for  twisted pair (Cat5e) cable so I can consider s-parameter model for the simulation? If so can you please share it. 

Also does you (TI) have a eye mask parameters defined for M-LVDS specifications? If so can you please share the same.  

Thanks in advance,

Kiran 

Hello Kiran,

Based on those requirements, 15 transceivers and at least 20 meters LVDS bus, this device will cause a lot of signal integrity issues. Again, LVDM devices are not meant to be used in a multipoint application.

Kiran Kumar R said:

I am aware of this part Josh, but SN65MLVD203B is not an automotive qualified deceive  and  SN65LVDM050-Q1 is the only product with automotive qualified from TI which supports M-LVDS. This is the reason to try out with the simulation. 

The SN65LVDM050-Q1 does not support M-LVDS. This device is still considered LVDS with a slightly higher current drive.

Kiran Kumar R said:

Does you (TI) have s-parameter model for  twisted pair (Cat5e) cable so I can consider s-parameter model for the simulation? If so can you please share it

I'm not aware of any S-parameter tools and eye mask parameters for M-LVDS.

Since we don't have M-LVDS Auto devices, have you looked at RS-485? TI has a handful of RS-485 Auto transceivers that sound like a better fit than using LVDM.

Regards,

Josh

Hi Josh,

Joshua Salinas said:

have you looked at RS-485? TI has a handful of RS-485 Auto transceivers that sound like a better fit than using LVDM.

RS-485 does not meet our requirements.

I tried further simulation of SN65LVDM050-Q1 using IBIS model using PCB trace as transmission line. Below are the circuit and the transmission line configuration details 

And below are the waveforms captured at the receiver and the driver respectively. 

I have added a two 50ohm transmission line and coupled them two make 100 differential pair. Despite having a 3-inch trace length, I'm observing unexpected ripples in the waveform, as seen in the attached images.

 I have followed LVDS termination scheme as also mentioned in the SLAA840. 

1. What might be the issue for such results, Is there any issue with the IBIS model?

2. Is not  the SN65LVDM050-Q1 driver impedance is 50ohm? What does RL=50ohm mentioned in the datasheet refers to? any thing needs to be modified in the transmission line impedance and the termination value? Please let me know.

--

Thanks in Advance

Kiran

Hello Kiran,

1) I also did a simple test with ideal transmission lines and saw some reflections

This IBIS model was released +20 years ago and there might be issues with the model.

2) The 50 load is simulating two transmission lines in parallel. Or two, 100 Ω, termination resistors in parallel. To improve the signal integrity the termination resistor can be lowered to 50 Ω.

Regards,

Josh