This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

UCC28782EVM-030: Series switch shown in "USB C Primer on Power Delivery Application and Requirements, 2016

Part Number: UCC28782EVM-030
Other Parts Discussed in Thread: UCC28782, TPS55288, TPS65987

Hello:

We will greatly appreciate if a  correct circuit-based explanation is provided for the series switch shown in the referenced 2016 TI publication named "Evolution of USB C" by Enos and Gosselin.

Pl see attached pdf file where we point out that it is preferable to operate UCC28782 with a constant out voltage, say, 20.5 ( or some such value) and let TPS55288 be the interface being controlled by WT6636 throuhg I2C more effectively,  thus offering highest overall efficiently of the connection as USB C PD. 

In what part of any USB C PD Testing/compliance will such a series switch be indicated?

And won't the body diode of the switch negate all such "switch" operations?

TPS55288 inserted as shown will provide all compliance conditions.

The only issue will be added cost of the buck boost. 

This is in exchange for clear failure modes of the series switch in almost all operations one can figure out from all possible Sink connections.

A clear circuit-based explanation will be highly appreciated.

evolution-of-USB_TI.pdfSERIES_PASS_SWITCH_USBC.pdf

r

  • Hi Robin,

    Have you looked at the UCC28782EVM? Most of the devices on this board were released after the publication listed above. The UCC28782 was optimized for USB Type C PD applications to maximize overall efficiency across the full output voltage range. On this EVM the WT6636 controls the VOUT of the UCC28782 directly and directly controls the series switch to adhere to USB type C PD requirements. There is no need to post-regulate with a DC/DC converter. 

    I hope this answers your question. 

  • Hello JC: 

    Thnx for the review. Appreciated.

    The question was indeed for the need for the series switch need: I have not found anything in the USB C specs- anywhere that the need for the series switch becomes mandatory!

    This is something I am trying to wrap my head around: would you please point out where in any USB Specs its use becomes mandatory and the only possible artifact for all USB C output topology?

    The second question was even more important: TPS55288 is not being used to do "post regulation". The crux of the matter is controlling UCC28782 over a wide output range( necessitating the use of BB pins in it) from 5V to 20 with a fixed transformer ratio with wide load variations. It surely is "doable" but of low overall efficiency compared to fixed output at 20V. Our proposition is that TPS55288 can do a more efficient job of stepping down as a buck converter to give 5 to 20v. IN THIS PROCESS, ITS PRESENCE PROIOVIDES WHATEVER THE SERIES SWITCH CAN  DO AND SOME MORE.

    Am I the only one to notice Failure Modes in the Switch Case? PL see attached Thesis from CalPoly Obispo. Like all academicians, one can see what is minimal & yet most beneficial in a circuit. A buck step down is.

    The third point was about the body diode. It makes all propositions about its "control" moot except where the sink is connected with lower pre-existing voltages, all other cases are also moot as far as control goes. 

    The buck behaves as a "load switch " as well- more effective, under full I2C control than a series switch. Wouldn't you think so?

    These are the necessary and sufficient conditions for considering a "load switch" if indeed USB C Specs mandated that a load switch as necessary clearly with a schematic!

    Controllers provide such a solution because doing our way would mean developing new code for the controller- and they do not se why they should. Would this be a proper assessment?

    For this reason alone, we would like to know about a USB C controller that we can program through GUI or like we do for all other frameworks for embedded controllers. 

    Will appreciate your analysis on this again.

    thnxSmart USB-C Wall Plug for the DC House Project.pdf

    -r

  • HI Robin,

    With respect to supporting 5 to 20V, the UCC28782 incorporates a small boost converter. So you size the aux winding for 20V and when the voltage is at 5V this boost converter is running to ensure we have a running voltage and can support the gate drive. This helps to improve the efficiency over the possible operating ranges. Before this feature was added, the solution was to add a linear clamp to prevent the aux winding from causing over-voltage to the controller when Vout was set to 20V.

    You certainly could use the UCC28782 with a fixed VOUT of 20V and post-regulate with a DC/DC.  I have not seen an apples-to-apples 4-point efficiency comparison between the two approaches. Using a DC/DC to post-regulate would likely result in a larger and more expensive solution. 

    I imagine the extra switch is required to ensure compliance with the USB specification. I do know that the USB forum puts strict limits on the output voltage and power and if the output voltage and current go beyond what is allowed, this shall be shut down and reset to the default 5V. It seems like controlling an external switch is an easy way to accomplish this. It may not be the only way. 

    That being said, I am not familiar with the exact details of the USB specification. I have forwarded this to our USB experts to comment. 

    I hope this helps. 

  • The Q6 FET in the UCC28782 is indeed required.  The reason is that USB-PD operation requires that VBUS be held below vSafe0V (0.8V) when nothing is attached or during a Hard Reset due to any fault condition.  This is achieved by disabling the Q6 FET.  Another reason is that most AC/DC applications require being "safe" under single fault conditions.  The PD controller can monitor the VBUS current and disable the Q6 FET if it is too high, this provides the required redundant protection.

    Doing an AC/DC conversion and then another DC/DC conversion is unlikely to be as efficient as just an AC/DC conversion.  However, if you did use a TPS55288 then the Q6 FET isn't necessary if the same requirements mentioned above are taken care of. 

  • Hello Deric

    Thanks for your review and comments.

    The body diode of Q6 makes every assigned operation it moot. Excepting 2 in series- as some of the others show.

    There is absolutely nothing in the 871 pages of the USB C spec that relates to any such implicit "Load switch".

    Our point is Q6 is definitely a Failure Mode ready to happen.

    I appreciate your last para: that TPS55288 will function as a load switch - and more effectively.

    Some do show this indeed.

    One thing is for sure: we need to program the chip accordingly.

    Our estimates are that fixed output from UCC28782 has many advantages.

    In moderate to high vol, 2   switches at Q6 ( for all its claimed functions) are comparable to 1 TPS55288....

    Again, very much appreciate this lengthy discussion on this important IO structure using TI parts.

    r

  • Hello Deric:

    We have absorbed all your remarks and appreciated your time and effort in reviewing our approach.

    We are finally deciding to use the TPS65987 chip instead of the UCC28782 EVM suggested chip. The main reason being we can program it as and when needed locally;  develop, test and upload to the chip a tested code. We also would provide the Alternative mode option in our modules.

    In this regard, would you please take a look at the Power Tree to the USB C receptacle attached? This approach relies upon the TPS65987 being powered up with the UCC2872 ac going active while VBUS will be clamped to 0v due to the body diode of TPS66288 buck, TPS65987 starts controlling TPS55288 right away, taking VBUS to Vsafe5 if commanded from Receptacle.

     TPS 65987 controls through I2C everything the USB C PD Specifications require particularly at VBUS. We keep the option of providing a "discharge  " path through the PPHV1 power path when needed.

    Any reviews of TPS65987 programming and as  master I2C control of the TPS55288 will be highly appreciated.Ganmar Power Tree_TI_USBPD.pdf

  • Hi Robin,

    Please take a look at apps note on how to configure the TPS65987 as the I2C master to TPS55288: https://www.ti.com/lit/an/slvaeq7/slvaeq7.pdf?ts=1680700745268&ref_url=https%253A%252F%252Fwww.google.com%252F 

    If you need further technical assistance please let us know.

    Thanks and Regards,

    Raymond Lin

  • Hello Raymond:

    Wow: special thanks for fishing out this exact APP Note out of the hidden gems' box!

    How would we have found it?

    It has been a month since I am hassling with "our scheme" of things....unbeknownst, that the App Note existed since 202....LOL

    On it. Let us get this implemented pronto.

    thnx much really.

    r