Opinions on the HyperJuice 100W GaN USB Charger?

[RobIII]

I used the search but couldn't find anything on the HyperJuice 100W GaN USB Charger (which is on Kickstarter).

It looks really promising, but don't they all? It claims to be "world's first [...] Gallium Nitride charger" and "[...] 95% AC/DC Efficiency (vs 91% for Apple/Anker)". It's expected to retail at $99. I have pledged $138 for 2 a few weeks ago and it's supposed to ship soon and I am curious to know what people think. I'll be glad to provide any findings and answer any questions once I receive them but please bear in mind I am a novice amateur electro-afficionado so I do have a cheap multimeter so don't expect fancy diagrams or scope reading or whatever.

[Marco]

Looks like a decent price for a premium product, seems to use Kickstarter more for PR than to Kickstart ... but that's par for the course at this point.

Assuming you're going to use it to charge a Macbook start with a simple hand test to compare temps.

[MagicSmoker]

Heh, I see the GaN switch they are using has an integrated gate driver, neatly sidestepping one of the biggest headaches with this finicky new technology, which is the narrow gap between the threshold and max allowed voltages at the gate. That said, at $8.60 each in single quantities at Digikey, you can get a whole USB-C charger for the same price as the switch here. One question I'd ask the kickstarter team is how they plan on getting through EMC testing because they must be running at a screaming high switching frequency (I'm guessing close to 1MHz based on the power density).

Coincidentally, I wrote an article for an EV trade magazine last year comparing GaN and SiC which is (mostly) still relevant: https://chargedevs.com/features/opinion-sic-vs-gan-semiconductors-for-ev-power-converters/.

[f4eru]

interesting insights into GaN!

EMC filters get better with higher frequencies, but probably still a huge challenge here, space and BOM cost is very limited here!

Others manage to pass also : http://www.adaptertech.com.tw/series1.html

perhaps they use a forward transformer with a good shield, and good separation...

[MagicSmoker]

interesting insights into GaN!

It's probably apparent from the article that my experience with GaN has been less-than-positive, but another member here (blueskull) uses it a lot and, one assumes, productively. Then again, he might just be working with GaN to publish IEEE papers... 

EMC filters get better with higher frequencies, but probably still a huge challenge here, space and BOM cost is very limited here!

Weeelll, sort of. Parasitic inductance limits the upper bandwidth for capacitors and distributed capacitance does the same to inductors so attenuation does not keep rising with frequency in any practical low-pass filter.

[T3sl4co1l]

The relative bandwidth of a filter of given geometry is approximately constant.  So, if you need to filter 100kHz-100MHz in a traditional design, you need to filter 2MHz-2GHz in a GaN design, and the filter is proportionally smaller and (hopefully) has the same relative bandwidth.

This does assume you can miniaturize everything proportionally, i.e. instead of a 20mm long X2 film cap, you can use a 1mm long chip cap.  Anything you can't shrink at the same rate, becomes that much more significant.

Tim

[ejeffrey]

This does assume you can miniaturize everything proportionally

Which of course you can't.  Electromagnetism scales nicely but pesky material properties don't.  i.e., you don't get 10x electrical conductivity or 10x higher breakdown strength when you miniaturize.

Weeelll, sort of. Parasitic inductance limits the upper bandwidth for capacitors and distributed capacitance does the same to inductors so attenuation does not keep rising with frequency in any practical low-pass filter.

It is hard with lumped element but distributed element / transmission line low-pass filters can be made with stop bands that extend arbitrarily high.  But fair call on "practical"

[MagicSmoker]

The relative bandwidth of a filter of given geometry is approximately constant.  So, if you need to filter 100kHz-100MHz in a traditional design, you need to filter 2MHz-2GHz in a GaN design, and the filter is proportionally smaller and (hopefully) has the same relative bandwidth.

Possibly - see below - but that's not what I was saying; rather, that the usual mains EMI filter design constrained to using safety-agency approved components won't provide ever-increasing attenuation with frequency.

This does assume you can miniaturize everything proportionally, i.e. instead of a 20mm long X2 film cap, you can use a 1mm long chip cap.  Anything you can't shrink at the same rate, becomes that much more significant.

Is that 1mm chip cap UL (et al.) approved? 

[T3sl4co1l]

Exactly.

There are other ways to skin the cat -- use normal-mode filter chokes instead of safety-rated common mode; use potted chokes/assembly to eliminate creepage; or on the source side, prefer resonant topologies that generate less high frequency content in the first place; prefer balanced or shielded transformer designs; etc.

It's hard to say if any of these options will generally save or cost space, though.

Tim

[kaz911]

here are some datasheets on GaN chips

https://www.navitassemi.com/download/

I have 2 HyperJuice on order as well. Hyper products are usually OK but not quite up in the Anker quality class for small USB/PD/C chargers.

But their customer service so far has been good and replaced the USB-C dock I got from them earlier.

[RobIII]

They finally came in yesterday. I have unpacked them but haven't used them yet. I'll do some (basic) testing soon and post the results here.



For now, most reviews appear to be quite positive; the kickstarter page shows a bit of a different story:

Really poor quality. My iPad keeps making disconnection noises from time to time. I have an iPad and just an Apple Watch connected, the Watch isn't charging.

Has anyone got problems with their unit? Their US socket is made approximately 3mm too short for it to work on all wall sockets. It would only power up after an adapter is attached.

The charger was not working properly on arrival.

It was working great, but 2 of the units have failed function properly. Whenever I plug USB-C / USB-A it just connect/disconnect repeatedly non-stop until I unplug the cable from either the computer/phone/ipad or HyperJuice.

I have received only one unit and I have the same problem

I see a spark when I connect it to my mac and this one does not even want to charge its battery
I tried the hyper cable on the mac charger and it works fine. I tried the mac cable on the hyper charger and it is still not charging

Received my charger a couple of days ago, noticed these thing so far:

- When there are two devices charging, the second device always switches from charging to not charging very rapidly

- Charging more than 1 device is not possible, tried several connect / unconnect sequences

- When charging through either of the 100w ports, my MacBook Pro only charges with 30w (with both the cable I purchased as an add on, and the standard Apple USB-C cable)

- My Nintendo switch drains the battery while it is charging through the HyperJuice (only device connected)

- It feels like my iPhone and iPad charge considerably slower than with the default Apple chargers, although I did not measure it.

Similar sort of issue. When I first got it, it seemed to have been working as expected but lately my MacBook Pro often says connected, but not charging. Especially when I have two USB-C cables connected, even if one isn't connected to any device.

Have had the Hyperjuice for several months and does not allow me to charge a windows surface pro 4 and Lenovo Thinkpad simultaneously. The same issue with a combination of either on of the laptops and an ipad pro (with a USB-C to lightning cable).

Seems as though when you plug in multiple devices to this charger, it randomly stops or won't charge one of the devices.

I have received my charger but its not working.

I have received the Hyper Juice charger and something I can´t understand is that the Hyper Juice charger can´t charge the Hyper Juice batteries, I have a couple of your batteries from your last campaign and I have connected to this new charger and it doesn´t charget at all any of both.

When connecting to both my Macbook and my Pixel phone, the charging status of my phone keeps connected and disconnected every few seconds. Any idea why?

...and the list goes on for a while which is pretty... worrying.

I'll try to get to testing my two chargers as soon as possible. All I can say now is that the chargers feel quite decent and the build quality appears to be good. The silkscreen is good, they feel light but solid, the plastic has a nice touch to it and the plug adapter fits nicely and snug. The power prongs extend with a nice and satisfying snappy click. That's all I can say for now.

[TimNJ]

Here's the teardown:

http://www.chongdiantou.com/wp/archives/51078.html

[Weston]

Thanks for the link! High quality teardown. The build quality of the charger itself looks decent.

It seems that the charger has no PFC. I am forgetting / having trouble looking up the standard, but I was under the impression that Europe and the US require power factor correction above some power level, and that that power level was something below 100W.

Am I remembering incorrectly, or can this charger not be directly sold in the USA | Europe due to the lack of PFC?

[TimNJ]

Good observation. I didn't even notice. The power threshold for PFC is 75W, from IEC61000-3-2. It has a CE mark so that can't be right.

[Siwastaja]

No, there is no requirement of any certain type of PFC. There is a requirement for exceeding a certain power factor. How they achieve it, if they do, is a different question. Classical boost stage for active PFC correction isn't really the only possibility although it's quite good when the only objective is to get very near to unity PF under full load (which is a stupid goal, though!)

[uer166]

There are a lot of sketchy things, first of all there are no markings about what standards the charger is compliant with, just an ETL and a VDE listing mark. Second, I searched the ETL listing and all combinations of the part number and the manufacturer name and the listing number on Intertek's website. They either faked the listing/certification, or it hasn't been published yet, or the search function is sketchy. I did find the TUV report where the charger was tested to IEC 62368-1, which seems to be just isolation/safety related with nothing about PFC requirements. So at least it won't shock you to death 

In any case at 200W+ input power it better have PFC to be legally sold (not that anyone would care IRL).. I also searched Sanho Corporation in the ETL directory and nothing came up.

[Weston]

No, there is no requirement of any certain type of PFC. There is a requirement for exceeding a certain power factor. How they achieve it, if they do, is a different question. Classical boost stage for active PFC correction isn't really the only possibility although it's quite good when the only objective is to get very near to unity PF under full load (which is a stupid goal, though!)

Sure, there are ways of achieving an acceptable power factor without a conventional PFC boost stage, like moving the twice line frequency energy storage to the secondary side. But if you look at the teardown almost all of their energy storage is on the primary side. Given that there are no signs of power processing between the bridge rectifier and the primary side capacitors its highly dubious that they are doing anything to manage the power factor.

Are we sure the CE mark does not stand for China Export   

[uer166]

A CE (even if legit) mark is meaningless, look for a UL, VDE, or CTL listing. So far I found VDE's, but it has not been evaluated to all the correct standards, specifically not to the ones dealing with power factor.

[Weston]

Thanks for taking the time to look up the certification stuff uer166.

As a practical matter, what are the implications of this? Are they not allowed to sell the charger in the USA / Europe? Can the end consumer sue them (based on that document, Sanho Corporation is located in the USA)? I am still in academia land so I am not directly familiar with all the implications of regulatory compliance.

[uer166]

Consider how much garbage one can buy from Ebay/Amazon with a "China Export" mark as you said, and the fact that nobody has the time nor interest to enforce regulations. These guys at least tried and got some form 3rd party to look at their thing, (even if it's not all the right standards).

As it is now it's heaps worse than any OEM charger, (say Apple or any other OEM), regulation-wise, but when 90% of Amazon has literally nothing to back up safety/performance claims, it's not too bad. If it gets big enough, their lack of PFC at 200-odd watts will get noticed, and since they're based in Fremont, maybe something will happen 

If it turns out that CTL mark is indeed bogus (it's unclear as I said), then CTL will definitely be interested, that's the only sketchy aspect that might be actionable.

[TimNJ]

No, there is no requirement of any certain type of PFC. There is a requirement for exceeding a certain power factor. How they achieve it, if they do, is a different question. Classical boost stage for active PFC correction isn't really the only possibility although it's quite good when the only objective is to get very near to unity PF under full load (which is a stupid goal, though!)

True. IEC61000-3-2 defines limits for harmonic currents drawn by the equipment, and does not mandate active PFC specifically. IEC61000-3-2 applies to equipment which draws between 75-600W, excluding Class C equipment (lighting). It's possible to have acceptable power factor/current harmonics at 100W without PFC, but it is unlikely without making some sacrifices in other areas.

Interesting point about optimizing PF for the full load condition, given that typical power supply usage is probably 50-75% of its full load rating.

[RobIII]

I have ab-so-lutely no clue as to what I'm doing (still learning with my brand spanking new Hanmatek DOS1102 cheap-o-scope) but for those interested; here's some 'measurements' I took with it from the HyperJuice :



For comparison; this is the output from an Apple 12W charger:



And the output from a USB port of my HP Z4:

[f4eru]

is this under load ?

[RobIII]

is this under load ?

Ah, yes, I should've mentioned. Sorry. Yes, it's under load of charging a USB device (this one to be exact but I don't think it matters). Why, do you notice something strange or not good? Or just curious? I also should've mentioned the left ones are using DC coupling and the right ones AC coupling. All 'measured' using a 10x probe. Again, I'm new to oscilloscopes and have no clue (well, at least not much) as to what I'm doing. If there's some specific setting or measurement you want me to do / try, let me know and I'll see what I can do.

[f4eru]

180mV p-p is probably OK.
But the strange thing to me is:  2,5 kHz ?
that hint to a problem in the regulation loop, perhaps ?