AM6442: Not getting proper waveform from AM6442 IBIS module while performing simulation

Hello Amit Gade,

Thank you for the query.

Let me check with the expert and comeback.

Regards,

Sreenivasa

Hello Amit Gade,

Please review the below document if this helps.

https://www.ti.com/lit/pdf/sdaa087

Please expect delay in response due to the US holiday.

Regards,

Sreenivasa

Hello Amit Gade,

Can you confirm the IBIS model you are using.

Our simulation expert is looking into the waveforms.

Regards,

Sreenivasa 

Hello Amit Gade,

Below are some queries from the expert. Can you please provide the required information.

Expert inputs requesting additional information:

I can run some experiments on my end but need to get some more info first.

Based on the device customer is using, I am assuming they are using SPRM810 for their IBIS model. Could you confirm exactly the SPRM # of the model customer is using?

Additionally, could customer confirm the exact pins and signal names that they are seeing issues in. Also, is customer able to share the IBIS models of the receiving devices?

 Once I get all this, I can run some quick simulations on my end to see what is going on.

Regards,

Sreenivasa

Hello Amit Gade,

app note for High Speed Parallel Interfaces showing techniques to improve signal quality for LVCMOS.

Kallikuppa Sreenivasa said:

Please review the below document if this helps.

https://www.ti.com/lit/pdf/sdaa087

Did you have a chance to review the application note?

Regards,

Sreenivasa

Hi,

please refer below requested information.

We are using SPRM730 IBIS model as we are using part number AM6442BSFFHAALV.

Receiver devices used are DP83867 for RGMII signals and S28HS512TGABHM010 for OSPI signals, IBIS models are attached

Please refer the exact pin numbers where we are facing issue, we are facing issue at AM6442 RGMII and OSPI signals driver side.

In previous post I have attached the some images of simulation circuit as well, Could you please check if you are getting similar issue at your end on same pin numbers.

IBIS.zip

Regards,

Amit

Hi Sreenivasa, 

I reviewed this application note. It suggests improving the signal by adding series termination resistors or other components. However, our first goal is to get a clear waveform with zero transmission path—meaning no PCB trace or other factors—between the processor pin and the receiving device pin. Once the simulation module shows a clear waveform with zero transmission path, In actual simulation we will adjust the series termination resistor to further improve the waveform.

Right now, at the initial stage only, we are not getting a proper waveform.

Regards

Amit

Hello Amit, 

Thank you. 

Amit Gade said:

We are using SPRM730 IBIS model as we are using part number AM6442BSFFHAALV.

 Can you please do a quick check using the below model:

https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1600269/faq-am6442-am6442-am6441-am6422-am6421-am6412-am6411---collaterals-for-reference-during-different-phases-of-custom-board-design-self-review-and-bring-up 

AM64x SR2.0 IBIS Model
SPRM810.ZIP (1889 KB) - IBIS Model

Regards,

Sreenivasa

Hi Sreenivasa,

We did a quick check using SPRM810 IBIS model, and we are getting similar waveform like SPRM730 IBIS model.

Regards,

Amit

Hello Amit, 

Thank you. 

Let me check with the expert and update you.

Regards,

Sreenivasa

Hello Amit, 

Can you please help me understand the simulation tool you are using.

Regards,

Sreenivasa

Hello Sreenivasa,

We are doing simulation on 

• Mentor Graphics Hyper Lynx 2504 update 4

Hello Amit, 

Noted and Thank you.

Regards,

Sreenivasa

Hello,

What you are seeing is the result of package parasitics in both the transmitting and receiving devices. Here is a diagram of what these parasitics (R_pkg, L_pkg, C_pkg) look like:

I went ahead and recreated your simulations, and while your diagrams have an arrow indicating probing is being done at the package pin (small black square in the image above), it seems like the probing is actually being done on the die (after L_pkg). Here is the setup and results:

Disabling package parasitic for both the transmitter and receiver will clean up the signals:

Now here is the thing, package parasitics are impossible to avoid. It is up to the engineer to design and simulate the rest of the channel to mitigate these effects and ensure simulations meet spec. Let us know if you have any further questions.

Regards,

Seif Ksouri

Hi,

Thank you for the information. Could you also please confirm whether the AM6442 supports adjustable drive strength?

Regards,

Amit

Hi Seif, noted. Thanks for the support.