TPS65988DK: Failure to negotiate 20V Sink PDO

Hi David, 

Can you provide the PJT file so I can take a look at your configuration from my end? Also what is the end application for this use case?

Thanks and Regards,

Raymond Lin

5A Version.pjt

It is a USB peripheral with 2 USB-C Ports.  When only connected to a PC from Port A, the port is used to receive power from the PC and does data transfer.  The second port is used for power only.  More power to the peripheral opens additional features.  Any power received on port B in excess of what is needed is offered to the PC on port A.

Hi David, 

Thank you for your patience, I tested your PJT configuration on a TPS65988DK EVM without any modification and saw no issues with achieving 20V contract. I've attached PD logs below and which can be opened using the software linked here: https://www.totalphase.com/products/data-center/ 

Can you provide the PD logs from your end so I can review? Also have you tested with different chargers to see if the behavior is the same? For context in the logs I provided (20V/5A vs 20V/2.25A) these were tested with two different PD chargers with different current capabilities at 20V. 

Thanks and Regards,

Raymond Lin

Port A at 20V/5A: https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/5A_5F00_Version_5F00_Port_5F00_A_5F00_20V_5F00_5A_5F00_Contract.tdc

Port A at 20V/2.25A:https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/5A_5F00_Version_5F00_Port_5F00_A_5F00_20V_5F00_2.25A_5F00_Contract.tdc

Port B at 20V/5A: https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/5A_5F00_Version_5F00_Port_5F00_B_5F00_20V_5F00_5A_5F00_Contract.tdc

Port B at 20V/2.25A: https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/5A_5F00_Version_5F00_Port_5F00_B_5F00_20V_5F00_2.25A_5F00_Contract.tdc

Thanks so much for your help. Attched is the negotiation that gets to 15V.  This was created using this app.   softwaretools.infineon.com/.../com.ifx.tb.tool.ezpdprotocolanalyzer

Opening the .csv with it will interpret the data fields.  Here is the Advertised sourc capabilities from the charger:

aNegotiate only 15V.csv

Here is what is requested

Best regards,

David

Hi David,

Are both ports exhibiting the same behavior when you plug in with this charger? What kind of charger are you using for test and have you tested with different 20V capable PD chargers to see if it still only negotiates to 15V? 

Another recommendation to try, I notice for Sink PDO2 you have it set as a variable sink as shown below:

Just for testing purposes, can you re-configured PDO2 to be a fixed type for 20V and see if the behavior is different? See screenshot below for reference:

Thanks and Regards,

Raymond Lin

Hi Raymond,

We found a powersupply that will negotiate 20V.  IN addition to those shown, a supply that shipped with a Lenovo Laptop was also tested.  Only the Amazon Basics supply negotiated 20V all others negotiated 15V.  

.

I will try your suggestion of parameter changes and post soon.

Both ports A and B worked the same with both chargers

I made the suggested change to Port. Both chargers that would only negotiate to 15V supplied 5V.  The charger that worked before, also worked with this change.

Hi David,

Going back to the original variable sink PDO, can you try these settings independently and test them again with the 20V chargers:

1. Set the operation current from 0.9A to 3A

2. Disable the "Ask for Max" bitfield 

3. Combine setting 1 and 2

Thanks and Regards,

Raymond Lin

Hi Raymond,

Neither of these had any effect on the outcome.  We found another charger for a different laptop that only outputs 5V when connected.  Here is the label:

Thanks for your work on this.

David

Hi David, 

Does your system have an EC that is communicating/configuring TPS65988DK during operation? I'm not entirely sure why specific chargers you tested with are not able to achieve 20V contracts when the EVM I tested your PJT config on was able to work. Do you also have a PD log for this specific Lenovo charger you can share with me? 

david hilderman said:

I made the suggested change to Port. Both chargers that would only negotiate to 15V supplied 5V.  The charger that worked before, also worked with this change.

Looking back on your previous feedback when you tested sink PDO2 as a fixed type, just to clarify you were able to see TPS65988 able to negotiate a 20V successfully then? Or was it still only able to negotiate 15V? 

If you are able to, I also recommend flashing your PJT configuration on a TPS65988DK EVM and testing with the same chargers you've been using. 

Thanks and Regards,

Raymond Lin

Hi Raymond,

Yes we have an EC in our system.  We need to query the amount of power available to determine what features can be enabled on our product. Then after that, if there is power left, offer that power to the PC connected.  I am getting our FW developer to remove all code that touches the TPS65988DK and let me test to see if somehow this is causing the problems.  As far as he can tell, it shouldn't but we won't know until the code is removed.  

After that if issues are still there I will provide the PD log.

Thanks,

David

Raymond Lin said:

Looking back on your previous feedback when you tested sink PDO2 as a fixed type, just to clarify you were able to see TPS65988 able to negotiate a 20V successfully then? Or was it still only able to negotiate 15V? 

It was able to negotiate 20V

The FW on the EC was changed so that it only read from the TPS65988DK.  No improvement in operation.  

A new piece to the puzzle.  If i use a Thunderbolt Cable to connect to the charger that normally will negotiate 20V, it then only negotiates 15V.

Attached is the PD log for the Lenovo supply that only results in 5V

Lenoovo 5V result.csv

Is there some timing thing that could be causing reads of the TPS65988DK to mess things up?

Best regards,

David

Hi David, 

During your previous testing when you were only able to achieve 15V contract, what kind of cable were you using? Are they normal USB-C cables or special ones that can do 5A/TBT/etc.? 

Thanks and Regards,

Raymond Lin

It is a special TBT active cable. 

I loaded the .pjt file on the EVK.  All supplies and cables succesfully negotiate 20V.  With the Jumper configuration set as below it should be identical to our board- I will go through things to make sure of this.

Here are the differences between the EVK and our board with the Jumpers a in the previous post

Other than that we use a different power source for the comparitor for the external switch on Port A.

I will work with the FW developer to have him remove ALL code that touches the TPS65988, even reads.

I see that my screen shot did not come through in the previous post.  Resent:

We removed all EC code and tried things.  No change.  Still only one supply negotiated to 20V.  On that supply, if I connected/reconnected multiple times, I was able to get it to negotiate 15V once.  

Hi David, 

Do you have the PD logs for the one supply that can negotiate 20V? What kind of cable did you use to achieve this 20V or is it directly attached to the supply adapter? 

Can you also send me your schematic design for your hardware? You can send me a direct message if you do not wish to upload directly to this e2e thread. I assume there is some hardware difference between your hardware design and the EVM if you're able to get the 20V result consistently with EVM on your end. 

Thanks and Regards,

Raymond Lin

Hi Raymond,

Attached is the PD log for teh supply that negotiates successfully.  This was done using a low cost USB-C to USB-C Cable.  It I use an active Thunderbolt cable with the same supply, it negotiates 15V.  I have also attached the PD log.Negotiate 15V w TB cable.csv

20V success negotiate.csv

Hi David, 

Let me review the schematic and PD logs to see if I can find anything outstanding that will cause this irregular behavior. 

Just to clarify from your testing, you were able to see TPS65988DK EVM to be able to achieve 20V using the same adapters and cable combination? 

Thanks and Regards,

Raymond Lin

Hi Raymond,

I was able to get proper operation- that is negotiate 20V with all adapters and cable combinations when using the EVM with the .prj I sent you.  

Thanks,

David

Hi David, 

Can you try checking to make sure the addressing is set correctly and also clicking on "Share Settings Across All Devices" is enabled and test on your hardware? See screenshot below:

Although it looks like the GUI addressing is matching the hardware (in your schematic ADCIN2 is tied to ground I believe) just want to double check. If the Share Settings check box works on your hardware then you might need to check the board assembly and make sure the R34 is properly populated. 

Thanks and Regards,

Raymond Lin

Thanks Raymond. 

ADCIN2 is definitely tied to GND.  Today I tried erasing the flash on the EC and tried again with the Share Settings Across All Devices checked.  No difference.   Such a mystery.  

When comparing operation of the EVK with our product, I have found one difference using the PD Protocol Analyser.  When connecting the USB C cable to the EVK, CC2 is 1.75V and less regardless of how you plug in the USB C cable.  On our product, in one direction  CC2 is 1.7V, in the other orientation, it is 0.9V.  In the oreintatoin where is is 0.9V, it sometimes spikes up to  1.1V for a ms or so.

I am using this application with the PD Protocol Analyser 

https://softwaretools.infineon.com/tools/com.ifx.tb.tool.ezpdprotocolanalyzer?_ga=2.23900794.119261010.1718040368-1844051579.1715792448

Here is the saved file.  In this file I have captured the following:

Our product USB-C cable in one orientation

Our product USB-C cable in the other orientation

EVK USB-C cable in cable in one orientation

EVK USB-C cable in cable in the other orientation

If you use the application, you can see the waveforms and voltages of the CC lines.

I do not see the file to be opened by the application in my previous powet.  I will try again with it zipped here.

Both Orientation DUT EVK.zip

It is most important that we are able to negotiate maximum power from port B in our design.  I experimented with parameters on this page.  I have tried

Offer Priority= Choose Highest Voltage

This had no effect.   I experimented with many different things on this page- Some changes that I made resulted in only 5V being negotiated, but I was never able to get 20V.

We do not use the TPD6S300A protection IC in our design.  Could this be causing problems?