EEVblog 1749 - Werewolf VFLEX USB-C Power Supply Adapters

[EEVblog]

Review of the Werewolf VFLEX USB-C power supply cable system.
https://werewolf.us/collections/werewolf-products/products/vflex-3-pack-starter-kit

[SiliconWizard]

What's their max rated power? With such a tiny PCB I wonder about max current. Would be interesting to run these at max power and take a look with a thermal camera.

There's also one thing that makes this potentially problematic - everyone having worked on USB-PD devices will probably know: USB-PD negotiates whatever the source can deliver. It's guaranteed to output the requested voltage as a maximum, but can output a lower voltage if the source can't output higher.  This could make some devices you power them with malfunction, and there is no way to know for sure what voltage it is until you measure it with a voltmeter, using the power source you'll be using, not a mobile phone. You can set it say, for 19 V in the app but once connected to some USB-PD power source, you may get only 12 V. For instance. That's the beauty of USB-PD and why its "universal nature" is de facto questionable *from the user's point of view* (also see discussion in the thread about the EU regulation).

That makes this kind of adapters potentially misleading unless you're very careful and have the equipment to check the output voltage, which the average Joe probably doesn't. Just because you set it to a given voltage in the app doesn't mean that's what you'll get with a given USB-PD power source. All you know (unless there is a bug in their app/firmware) is that it won't output *higher* than what you set. I have designed a USB-PD power supply, and I know Dave has too, so believe me/us, this isn't as simple and as rosy as it sounds.

Also, but I may have missed it, there doesn't seem to be any protection on these adapters so they fully rely on the protection in the power source itself. While that may seem reasonable in this case, I personally wouldn't design a power adapter of any kind, even this, without some added protection against short circuits or even against undervoltage (which would mitigate the issue above but would definitely add complexity to the board). But if I have missed protections they offer, please let me know.

Just my 2 cents.

[thm_w]

What's their max rated power? With such a tiny PCB I wonder about max current. Would be interesting to run these at max power and take a look with a thermal camera.

5A max. Very short path through the PCB so it should be OK.


As for the malfunction, I think he showed the LED will light up red if it can't negotiate the full voltage.

[EEVblog]

As for the malfunction, I think he showed the LED will light up red if it can't negotiate the full voltage.

Correct.

[lgbeno]

Willing to discuss the VFLEX design more, as already mentioned, VBUS is a straight passthrough from the USB connector to the output contact.  This is also done in 2oz copper.  I've run tests running at 5A continuous and there is just a modest temp rise.  Most of the heat generated comes from the USB-C cable that is connected to VFLEX.  In the end, with a good quality cable, there is about a 30 deg F temp rise when running continuously at 5A.



Regarding protection, the supplies provide this.  A series switch inline or other protection there would be a lot of added physical space, cost and much more heat generated in the normal use case.  In the end it just did not seem practically necessary.  Just like any power supply, that you plug into something, there are risks involved.

The statement about the reverting to 12V when targeting 19V, I don't think that this is an action taken by the supply, it could have been a feature of that particular trigger chip.  For VFLEX, if it cannot achieve the target voltage, it will always revert back to 5V and show the LED.

We also independently monitor the voltage with on chip ADC so if you see a green LED it means that negotiation was successful AND that the ADC measures a voltage that is within tolerance.

There are 3 failure states:
Solid Red LED = Negotiation success but ADC voltage out of range
On / Off Red Blink in 1s interval = Negotiation not successful
Fast Blink RED = eMarker Cable error

Especially for voltages over 20V, the supplies also enforce a cable check so you need to consider compatibility of the cable that you are using as well.  VFLEX tries its best to distinguish those types of error.

Really enjoyed the video and happy to answer any questions that people may have!

[Bzzz]

Thank you, very much appreciated to highlight this in a separate video. I was aware of the granularity of modern USB chargers and people discussed this as a low-tier lab supply option back then when the spec was announced, but I haven't seen a device that actually makes use of it. Once this type of device goes so mainstream that I can order it for work at the usual suspects, I very much will put these in our R&D offices.

[lgbeno]

Thank you, very much appreciated to highlight this in a separate video. I was aware of the granularity of modern USB chargers and people discussed this as a low-tier lab supply option back then when the spec was announced, but I haven't seen a device that actually makes use of it. Once this type of device goes so mainstream that I can order it for work at the usual suspects, I very much will put these in our R&D offices.

I see that you are in Germany, curious who you think is a good outlet to retail through?  Amazon?  are there others?  We are starting to list on Digikey but international shipping rates are a little ... teuer

[negativ3]

With it being programmable via an app, I'm interested in offline capabilities to do so and fault tolerance.

[Bzzz]

At my company, adding new vendors is an arduous task, so unless we really really REALLY need something from a new company, we stick with the large electronics distributors - Farnell, Mouser, Reichelt, Digikey, RS (usually in that order). Amazon Germany is possible if they send a proper invoice, which isn't always the case. Amazon abroad is a no-go. Just to make that clear how silly we operate: Whenever some express fee can speed up R&D work, we opt for that. 2k extra for some PCB prototypes to arrive one week earlier? Done, paid, let's go. Earlier this month, we dismissed a 40-bucks, 2-week delivery special-made product from a Chinese seller in favour of a local company that we have dealt with before - at a 400€ price tag and 6 weeks lead time. Simply because we couldn't get proper paperwork from the guys in China. It's absurd.

So - if you can make e.g. Digikey sell and ideally ship your product, we're in!

[lgbeno]

With it being programmable via an app, I'm interested in offline capabilities to do so and fault tolerance.

For Offline, the app nominally works in offline, we still have some improvements to make on the inventory offline sync but you should be able to configure voltage offline.

In any case, there is also https://lib.vflex.app where we have a open source CLI and some browser code that can be downloaded and run in the browser offline.  You can also use this to write your own stuff as well.

[lgbeno]

So - if you can make e.g. Digikey sell and ideally ship your product, we're in!

Sounds like a great customer!  (but also yes, silly corporate policies...)

We have one listing here via Digikey Marketplace: https://www.digikey.com/en/products/detail/werewolf/VFLEX-DIY/28793751?s=N4IgTCBcDaIGoDEAyBRAGiAugXyA

Will be adding the entire catalog soon.  Can always check back on the company page for more details.
https://www.digikey.com/en/supplier-centers/werewolf

Since it is marketplace, these items ship from our warehouse but from a purchasing perspective, it is transacted by Digikey.

[SiliconWizard]

Willing to discuss the VFLEX design more, as already mentioned, VBUS is a straight passthrough from the USB connector to the output contact.  This is also done in 2oz copper.  I've run tests running at 5A continuous and there is just a modest temp rise.  Most of the heat generated comes from the USB-C cable that is connected to VFLEX.  In the end, with a good quality cable, there is about a 30 deg F temp rise when running continuously at 5A.
(...)

Hi, thanks for detailing your design here. Monitoring the actual output voltage is a nice addition.

Regarding the negotiated output voltage, it all depends on the PD controller you use. If it's fully programmable (which appears to be what you use?), then you can query a specific voltage and if the source does not acknowledge it, you can act as you decide to. Some simplified integrated PD controllers do automatically fall back not to 5V, but to the nearest, less than or equal, voltage the source can provide. If I understand, in your case, you fall back to 5V. While that's reasonable, that can still be a problem with some devices. Some devices that, say, must be powered with 9V DC, may not behave all that well if they are powered with only 5V for longer than just a few tens of ms. That's a case that you don't necessarily take into account when designing a DC-powered product. So a given product may malfunction if powered continuously at 5V and possibly even get damaged. Now, good thing you have an indicator (red LED) for that.

Now of course, to mitigate that, you would need to add a power switch instead of a direct connection, as I don't think (?) you can make a USB PD source shut down VBUS entirely. And so we're back to your rationale of making something simple and compact.

Ditto for adding some protection, that would sure add complexity, cost and size. I'm just wondering if, from a CE or UL POV, it would be fully compliant. I'm just raising the question, I don't have a final answer to that. But as an active device, even if you consider there is only a passive connection on the power path, I wonder if you can be fully compliant with safety requirements with no specific protection. So, opening the question, this device being unconveniently somewhere between a fully passive connection device and an active device.