PD Sink & Source bypass solutions for maximum "efficiency"

Announcements

Live Webinar: USB-C adoption. Simple & Cost-efficient solutions | April 18th @9am or 5pm CEST. Register now !

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
DaEr_349131
Level 5
Level 5
25 likes received 50 sign-ins 5 comments on KBA

Hi Cypress/Infineon EZ-PD team,

I wonder if you have done any existing designs where you allow a PD sink "input rail" to be directly passed through to a PD source port?

Some use-cases where this would be beneficial could be:
- 45W+ dual port or charge-through applications (hubs, docks, notebooks etc) where one port is acting as a PD sink (5-20V) and one or many other ports (type-c jacks) acts as a PD source (5V or 5-20V)
- PD source applications with an internal or external AC/DC of 20.0V 65W, a Type-C PD (source) output jack where you want to offer 20V, 3.25A with minimal losses / heat.

 

Background:

Most designs utilize a buck or buck/boost in series with the source rail and the PD output, here you'll always end up with about 5-10% losses. This is usually not a viable solution when dealing with higher power PDO's.

Other designs might support a 5-15-20V sink input and a 5V source output (type-c or type-a) where you need output port to be able to deliver 5.0V even if the UFP (or charger) only offers a 5V PDO. Having a buck in series would result in too much drop on VBUS, while using a boost will come with added parts/complexity and efficiency losses.

 

I believe the right way of doing it would be to have a smart bypass path that can tie the PD sink port straight to a PD source port under the condition when the PDO's on each side matches.

To make it safe, it's probably wise to also monitor the source rail "VBUS_IN" (coming from the UFP/charger) to ensure that the requested PDO on the source port is 1:1 with what "VBUS_OUT" should be.

I ran across a similar application attached. So wonder if there has been such logic implemented in any of the CCG's?
I believe it should be a mix of state-logic + a fast ADC/comparator to safely control the bypass path.

 

Thoughts?

 

example where a buck or buck/boost would be bypassed for near 100% efficiency

Untitled.jpg

0 Likes
1 Solution
Pranava_YN
Moderator
Moderator
Moderator
100 likes received 500 replies posted 250 solutions authored

Hi,

The use case is a valid one however we do not have example implementations of this on our CCGs.

My concern is during non-5V contract on the Source side of the system, it is expected that initially 5V is provided before contract and it has to be ramped up to higher voltage post PD negotiation. In your case, initially 5V should be supplied using Buck/ Boost and it should be bypassed post contract. There might be some issue when transitioning from Buck/boost output to bypass output during this case.

Best regards,
Pranava

View solution in original post

0 Likes
1 Reply
Pranava_YN
Moderator
Moderator
Moderator
100 likes received 500 replies posted 250 solutions authored

Hi,

The use case is a valid one however we do not have example implementations of this on our CCGs.

My concern is during non-5V contract on the Source side of the system, it is expected that initially 5V is provided before contract and it has to be ramped up to higher voltage post PD negotiation. In your case, initially 5V should be supplied using Buck/ Boost and it should be bypassed post contract. There might be some issue when transitioning from Buck/boost output to bypass output during this case.

Best regards,
Pranava
0 Likes