TMDSEVM6678: board versions clarification

Thanks Mukul!

The customer is facing issue with SRIO connection for the DSP - of TMDSEVM6678LE units procured.
They do not observed this issue on the TMDSEVM6678LXE.
What do you think could be the issue?

Best regards,

Art

I am not sure. We no longer have in depth expertise on this family of devices.

I this is sample set of 1 EVM each on which the observations are based, it would be good to confirm that the software is exactly the same. 

There are diagnostic tests and BER tests that may be needed to ensure proper working etc. 

Hello Mukul,

The customer may need additional assistance. I might direct customer to this thread so that he can explain their scenario better.

Thanks and regards,

Art

Dear Mukul,

Please find more details on the problem which we are facing.

We notice 2 different Revision (A and B) for these boards.

Please find below photo of the serial number of the units based on the 2 variants in the Revision.

At this moment, we are doing a test on the SRIO interfaces and are connected 2 boards in a loopback configuration.

We have issues with this interfaces. We do not see any data flowing on this interface.

We observe that the boards are loaded with different FPGA image ( which configures the clock generator chip).

This is problem, because some peripheral will not work if clock generator is not configured correctly.

Could you confirm on this?

We also need DSP executable to program the FPGA on the board. We need “fpga_evmcc6678I.out”. This is not available on the website.

Dear Mukul,

Please find more details on the problem which we are facing.

We notice 2 different Revision (A and B) for these boards.

Please find below photo of the serial number of the units based on the 2 variants in the Revision.

At this moment, we are doing a test on the SRIO interfaces and are connected 2 boards in a loopback configuration.

We have issues with this interfaces. We do not see any data flowing on this interface.

We observe that the boards are loaded with different FPGA image ( which configures the clock generator chip).

This is problem, because some peripheral will not work if clock generator is not configured correctly.

Could you confirm on this?

We also need DSP executable to program the FPGA on the board. We need “fpga_evmcc6678I.out”. This is not available on the website.

Sruthy

We do not see your images of the board serial numbers.  Please try to post them again.  A PNG file is preferred.  If you cannot get the image to load, please provide text showing the serial number, model number and revision.  Please also post the FPGA revisions.

Information for this EVM is available on the Advantech website at https://www2.advantech.com/Support/TI-EVM/6678le_of.aspx.  As shown on the website, we have been shipping version A104 for the last 9 years.  We need to understand the revisions that you have on hand.

Hi all,  c6678 EVM Board with version 3A gives problem in SRIO port initialization. The SRIO port is not operational when I run SRIO bench marking test between producer and consumer.  The same sw does not give port operational error in 3B revision of the board.. I am attaching the image of the 3A boards here. We have total 3 c6678 EVM Board with version 3A, all of them does not work.

Hi 

Thanks for the additional information , however we really need full model number, serial number and revision of each board.

Can you also clarify where you purchased these from?

Regards

Mukul 

Dear Mukul, 

Please find below the part details we had recived against the order for TMDSEVM6678LE through ti.com

Model# 9792830122E , DSPM-8301E :  S.No# EPD0065402, EPD0065405, EPD0065393, LKD0091560

We are still trying to understand what hardware variants that you have.  As Mukul said, support for this old EVM design is very limited.  Please provide the following information for each EVM.  (We know this information is somewhat redundant but we need to have a clear picture.)

• Model Number
• Serial Number
• Board Revision
• FPGA Revision
• Purchase Date
• Purchase Vendor

Also, please re-check the board with serial number LKD0091560.  This does not appear to be a valid number.  Unfortunately, the picture provided of the EVM is too poor to collect any of this information.

Tom

Dear Thomas,

Please be informed that the device were purchased directly through ti.com online portal :

Purchase Date and Order reference from TI.com as below:

TI.com Order# T01303041(3 units) – 22^nd April 2021-PO Reference DIF-P-00708

TI.com Order# T01303045(1 unit) - 22^nd April 2021- PO Reference DIF-P-00708

TI.com Order# T01119985 (1 unit)- 31st January 2021 - PO reference #DIF-P-00565

Dear Thomas,

Can you please advise how can we find the

• Model Number
• Serial Number
• Board Revision
• FPGA Revision

Sruthy,

I thought that you previously mentioned differences in the FPGA version.  I may have been mistaken,  Please provide the other information for each board.  The placement of this text is shown on page 2 of the Advantech Known Issues document.  You download this from the following webpage: www2.advantech.com/.../6678le_download3.aspx.

Tom

Sruthy,

You reported SRIO connectivity issues.  How are you connecting the EVMs for SRIO testing?  What data rate and SERDES configuration are you using?

Tom

WE HAVE 3 UNITS REVA, Model# 9792830122E , DSPM-8301E : S.No# EPD0065402, EPD0065405, EPD0065393, REVA  - does not transmit the data  at all while REVB LKD0091560 is working Fine.. We had received all the devices for orders placed through ti.com T01303041 and T01303045 on same date (the reason to place as two different orders was tha the order qty was limited o 3 per order).. We would like o know the difference beween RevA and RevB.- please advise if you can clarify on any technical performance difference on both revisions.These are the devices we had bought from ti.com

Sruthy,

I am not able to find any documentation for a Revision B.  I assume that what you are describing as Revision A is actually production revision 3A.  I am also confused by the number LKD0091560.  This does not align with any of the known serial number or model number formats.  Is it physically different from the other 3?

You reported SRIO connectivity issues.  How are you connecting the EVMs for SRIO testing?  What data rate and SERDES configuration are you using?  Have you probed the SERDES to verify that it is operating at the expected rate?

Tom

Hi sruthy,

In order to troubleshoot the issue in SRIO, please do share the register configuration of SerDes register(SRIO_SERDES_CFGPLL), SerDes Receive/Transmit Channel Configuration Registers and Error detect enable (ERR_DET) registers used in code for both in Rev A and Rev B boards. Please do confirm that SRIO issue is observed when using same code on Rev A and Rev B boards.

Also do share the CCS version, SDK version.

Please follow this link for SRIO register configuration and sample code:

https://www.ti.com/lit/ug/sprugw1c/sprugw1c.pdf

Thanks & regards,

Rajarajan U

Dear Thomas,

Please be informed that physically there are no visible differences between both the models.

Both the revisions A  and B was procured from the official TI.com portal only. The revision B is perfectly working now and we are facing issue with the Revision A.

Can we please have a call or Microsoft Teams meeting to resolve/ clarify the doubts both from yours and end user technical team.

Please share us your convenient time or kindly call us : Mr. Ibrahim :0558636833 (end user)

Awaiting your response

Thanks in advance.

Dear Raj,

Please be informed that the issue is reported when we try to run the same code/application at same rate on both revison A and B.

Currently revision B is working fine, however, revision A is still having issue.

Can we please have a call with you. Please share us your contact number or email ID so that we can ave a Teams meeting request send to you for faster resolution.

Thanks in advane,

Sruthy Mathew

Hi TI team,

It is important for us to know the main difference between Rev3A and Rev3B boards. Can you let us know the differences between the 2 revisions?

Please find pictures of the 2 boards variance for your feedback.

Rev3A

Rev3B

Hello Ibrahim, Sruthy 

We were able to track down the difference on rev 3A vs 3B. 

Rev  3B has the SR2.0 silicon (latest and what customers will go to production with). Rev 3A may have older SR1.0 silicon. 

https://www.ti.com/lit/er/sprz334h/sprz334h.pdf

If you look at the errata, you will see how you can read some registers to confirm silicon revision. You will also notice that SR1.0 has several silicon errata around SRIO.

So for stable development you need to be on SR2.0 silicon based EVMs. 

Dear Mukul, Thanks a lot for the update.

Please could you advise if we can get the replacement of  RevA ( 4units with RevB) for the orders we have placed in January through TI.om.

Mr. Art had informed you would be able to help us with the details of ADVANTEC team who had fulfilled the order earlier.

Also, please confirm if we can et REV B units if we now initiate order through TI.com.

Hello Sruthy 

Sorry for the delay. We are working to internally figure out how we can initiate a swap for the EVMs. 

I do need your help to confirm one thing please.

Please refer to the silicon errata table 2 and  confirm that EVM has SR1.0 for the non working board vs SR2.0 for the working board by reading the JTAGID register value (the base address is in the datasheet) and the details SR1.0 vs SR2.0 value is in the silicon errata (Table 2) 

https://www.ti.com/lit/er/sprz334h/sprz334h.pdf

Hi Sruthy 

I was informed that at least for your case the working and non working board seem to have the same silicon rev, SR2.0 , is that correct?

If so, then do we need to go back to the question that were previously asked

You reported SRIO connectivity issues.  How are you connecting the EVMs for SRIO testing?  What data rate and SERDES configuration are you using?  Have you probed the SERDES to verify that it is operating at the expected rate?

Additionally did you figure out the difference in the FPGA firmware?

Dear Mukul,

We need the swap of 4 units Rev A  with Rev B.

The Rev B version is working fine  according to our application.

Kindly support on Urgent basis regarding:

1. Swapping the devices for the shipments we received earlier

2. we also need to place new orders for Rev B units specifically. Please help us get connected with the right team who can ship RevB.

Awaiting your response soon as the delay is affecting our project development.

 Thanks in advance, 

Sruthy Mathew

Dear Mukul,

Dear Mykul, 

Please find information on the JTAGID which we have captured. It indicated R2.0 silicon for Rev3A boards also.

Hence this is an issue.2251.JTAGID_MM_REVID.docx

Dear Mukul , t would be great if we can have a technical call aligned with the end user for better understanding of the issue and faster resolution.

please share us your email ID and convenient time  to which we can share an official invitation for  a TEAMS call. 

Awaiting your response.

Hello Sruthy

Our support is via E2E.  We typically cannot support direct calls. 

Can you please clarify what you wanted to accomplish in a technical call?

I believe you have started the process for returning the EVM. The field team assigned to you have submitted a request for 2-3 new EVMs that we are hoping will be Rev 3B. All of this is in works. 

As mentioned previously , at least from the word doc you shared it appears that the C6678 silicon on both the working and non working board is SR 2.0 , so there should not be any silicon related issues with SRIO as was the case with SR1.0. 

It is unclear why 3B works for you and 3A does not. Our next action is to get you 3B boards as part of RMR.

Please let us know what additional support is needed?

Best Regards

Mukul 

Dear Mukul,

Please can you expedite the procedure for the RMR of 4 units REV A and replace the same with REVB.

Also, yes we will need additional 2- 3 units TMDSEVM6678LE- but working version please - it would be of great help if this can be confirmed before the release of the units from your partner warehouse..

Please share us the details of the supplier who can support us on the same.

thanks in advance,

Sruthy Mathew

Hi Sruthy

Ok. I will try to track the latest status on the RMR and replacement EVMs.

We have limited quantities of these EVMs and are not planning to make new ones. We would like you to first get hold of  the 3-4 units and confirm them working prior to approving any new sales. 

Regards

Mukul

Dear,

Message noted with Thanks. Would be awaiting your update on RMR.

Also please advise us on the new replacement model for the Part number if incase you have any visibility.

Regards,
Sruthy Mathew 

Sruthy

It appears that the latest boards we have are 6678L , not LE. The difference is highlighted in the product page. L version do not have the XDS560v2 , just XDS100 support. Does that work for you?

In your email you mentioned

>> Also, yes we will need additional 2- 3 units TMDSEVM6678LE- but working version please

The EVMs you have and the EVMs we will ship , all have SR2.0  silicon, so we expect all of them to work. It is unclear why the boards you have are failing SRIO test. 

We can ship you the latest boards from inventory (TMDSEVM6678L) but they may behave the same as your other 3A boards do. However on we are willing to send these to you as part of RMR. Please confirm that c6678L boards are ok. 

Dear Mukul,

Can you please confirm if we are getting TMDSEVM6678LE Rev B for RMR or only TMDSEVM6678L for both new and RMR.

Dear Mukul, Kindly be informed that Our preference still is on getting TMDSEVM6678LE  working versions if available for both new and replacement.

Also, please advise if we get  TMDSEVM6678L replacement board as a replacement instead of TMDSEVM6678LE, is it possible to hold chances that we will still need to revert back to TMDSEVM6678LE if TMDSEVM6678L is not operating correctly.

Sruthy 

I want to re-iterate again that for now I am trying to get your RMR quantity worked out. The newer boards are "L" version. The LE version are older  *however* in both cases it appears that they have SR2.0 silicon --> which is also the case for the EVMs you currently have (even though they are 3A , per the JTAGID you provided) 

I strongly encourage that before putting an order for additional EVMs, please see if you can get the replacement EVMs working. We have limited inventory of these boards and not planning to build any more boards once the stock gets over. 

Regards

Mukul 

Dear Mukul,

Please advise if we can get one unit of test device for TMDSEVM6678L  to be shipped to Abu Dhabi -UAE.

We shall then have it tested and confirm you the replacement part number and for new orders.

Is there any chance you can support us with a free demo test unit.

Or kindly advise with the discount you can offer to us for demo purchase of TMDSEVM6678L .

Thanks in advance,

Sruthy Mathew

Sruthy 

For RMR replacement we are processing 3 EVMs request - we will send you 2x LE and 1x L.  These will be shipped to UAE as per the order entered by the field team supporting you. 

Returning in the non working EVMs can be worked in parallel separately.

There is no difference in the price. 

Regards

Mukul 

Dear Mukul, thanks for the update.

please note that the RMR device count is 4 units in total, not 3.

Dear Mukul,

Could you please expedite the RMR process for he release of TMDSEVM6678LE - 3 units and TMDSEVM6678L   1 unit at the earliest.

Awaiting your feedback with shipmen status on the same.'

Thanks and regards,

Sruthy Mathew

Dear Mukul,

Could you please advise us on the status of RMR request.

Kindly share us the shipment tracking details of the replacement units.

Your swift response would be highly appreciated.

Regards,

Sruthy Mathew

Sruthy

Please feel free to open another ticket with CSC for getting tracking # etc. What I know from our EVM planning team is that they have put a rush order for 3x LE EVMs , with latest date codes. I do not have any visibility beyond that on the logistics of shipping and tracking. 

When the assigned field team is back, they will also put another request of 1x C6678L EVM , as per your other request , that you had requested RMR for 4 EVMs and are ok with evaluating one L EVM. 

Hope this helps.

Regards

Mukul  

Dear Mukul,

It would be great if you could share us the RMR ticket details or Case# to follow up with the team.

I hope CSC- team means the representatives handling tickets raised online.

Hi Sruthy 

Fedex # 937405044357

Regards

Mukul

Dear Mukul, Please be informed that the invoice copy along with FEDEX shipment 937405044357 has different part number TS5A3159MDBVREP instead of TMDSEVM6678L   and also is addressed to wrong consignee (to China)

Sruthy

I am sorry to hear that. I am afraid I will not be able to provide further guidance on this. E2E is predominantly for technical questions and support. 

If you have been send the wrong EVMs , you will likely need to raise a ticket with CSC to track this down further. 

You can share this forum thread and the request Case No.: CS0616883, ZS Request #35533

https://ticsc.service-now.com/csm

Hope that  you and your team will be up and running soon.

Regards

Mukul