Mine will be less than 10
Thanks for the data, I see people babble on everywhere about high voltage but we forgot that the people who are building high voltage systems are the ones most active on the forum anyways, 12s is ample and is more than enough for most people.
Can’t even imagine selling 2 chargers in months while having 498 chargers kept as inventory lol. Thanks man you helped me dodge a bullet on this one
5 Likes
The main target with using GaN is to shrink the size, not so much having massive power. SiC chargers can already do massive power, but even at 500W a SiC charger is gigantic, hard to stuff down most commuter backpacks. Imagine a 1500W charger to bring with you along group rides, how heavy would that be carrying that on your back the whole day…
I agree with you that GaN fets are much more expensive, a 5A SiC charger cost 30 USD to order in bulk, while a 5A GaN charger cost roughly 90 USD to order in bulk. The friggin PFC alone is 20 USD, but the main point of me trying to get these made is size and weight reduction, not power.
well, The surface area of the transistors are very equal. Most of the time the GaN is alot bigger (due to isolation when approaching higher voltage).
Unless the GaN trasistor has some wild benefit when rectifying the AC. I dont see how it would make the formfactor smaller. What usually takes space is the coils, transformers and electrolytic caps. Has nothing to do with the FETs. I’d love to be wrong here.
To make them more compact you’d have to look at chargers that dont cheap out. Firm examples of this is using 4 or more layers of PCB, more expensive caps and small formfactor coils, a good compact heat disspation from a heatsink (preferably fanless) + bunch of other variables. These together will shrink size considerably. At this point GaN for anything below hundreds of volt is just a marketing scheme.
Personally I think the USB-C 3.x PD looks more promising(100W and above). This still requires a boost converter before reaching a good charging voltage tho, the voltage domain of USB C is a bit limited.
6 Likes
To my understanding, the perks of using GaN is mainly due to their efficiency. GaN fets are 95%+ efficient while traditional Sillicon based fets are at best 80% efficient. This allows the usage of higher efficiency PFCs as well as that of other components, which increases the efficiency of the whole system.
What this means is you can design a system where less surface area or even passive cooling can be achieved for the same charger wattage by using GaN fets.
This is how I understood it. I might be wrong tho
1 Like
I think this is a common missconception. It all depends where you use it. The reason Silicon carbide and GaN is intresting is due to thier low Ron at higher voltages (what typically dominates losses in FETs, Pf = R x I^2) . A traditional MOSFET for say 650V. will have higher Ron than a GaN transistor. While both GaN and Silicon carbide have started to appear for lower voltage ranges, the traditional transistors still rule. Especially sub 100V domain.
Heres an exmaple from mouser: UF3SC065007K4S (SiC) has a Ron of typically 9mOhm at 650V.
While the closest Si transistor STY145N65M5 has 15mOhm. thats almost half the resistive losses! Apart from that specific Si transistor, the SiC type dominates in terms of lower Ron at a higher voltage.
If you go even higher in voltage, then it becomes very clear that GaN and SiC are very energy saving.
But for stuff around 100V its still really hard to motivate why to use anything other then a Si FET.
As you might understand, using these fancy GaN FETs at 50.4V when a corresponding Si FET has 1/6th of the Ron makes very little sense 
6 Likes
Meanwell rates their HLG series which I think use silicon fets at 94% efficiency for the whole thing
I can’t wait for 48v 5A 240w USB PD 3.1 to hit the market.
3 Likes
Calculating total efficiency from AC → DC side has to do with alot of things, not all losses occur in the FET The most difficult thing is to maintain high efficiency in switched applications. (think converting voltages) If you have a Buck converter that does a terrible job under a certain load f.e, its probably not the transistor causing all of the trouble. So these 90% numbers beeing tossed around is an accumilated number for the entire device under specific conditions where alot of different losses are added. The FET losses is a part of that but not all Usually in cheap DC supplies/chargers, the Electrolytic caps are scorching hot. Thats due to the switching in the cap, these bulk caps are usually the cheaper kind with abit higher impedance, generating alot of heat.
3 Likes
So what you’re saying is technically, you can also achieve high efficiency with Si or SiC fet chargers if you use better components, people just threw GaN in there for marketing scheme?
2 Likes
Pretty much.
Sadly thats what I’ve discovered so far. I was very excited a few years ago to start designing with these new miracle FETs only to realise they’re more tailored to E-car applications
7 Likes
Thanks for letting me in on another trade secret
I’ll be advertising the GaN side of the charger less and more on other components
3 Likes
Hope you run 10S LOL
Just misses the 50.4V threshold 
1 Like
4V/cell is plenty 
4 Likes
That and boosting from 48v to 50v seems a lot easier (smaller, more compact) than from 20v
2 Likes
Bring on the 11S boards
5 Likes
20% fluctuation is typical for these ratings
If your environment is not the Sahara desert I’d say 50.4v is fine
That is a great option. I might have to pick one up.
exactly for charging on the go, only need top off and max cargo savings.
3 Likes
haha just an xt60 or 30 on the charger and you got it covered.
Couple people said that before me actually.
Idk what else to add. Idk maybe someone could make a polariy checker and voltage readout dongle as one of the output adapters available.
1 Like
The way GaN makes things smaller is they can be switched at a much higher frequency with lower switching losses; they feature very little gate-drain charge. This allows the use of smaller magnetic components. Effectively, you can drive up the switching frequency while keeping the heat dissipated by the transistor the same. Alternatively, you can keep switching frequency the same, and move your conduction loss <-> switching loss equilibrium point to generate less heat and reduce required heatsinking; all depends on what is taking up the most space in your design: the heatsink, the magnetics, the creepage requirements, etc.
Current high voltage GaN and SiC are comparable in terms of on-state resistance, though I think SiC is marginally better, with SiC being much easier to drive and cheaper, though their maximum switching frequency is somewhat lower. GaN has excessive gate leakage and needs an extremely well regulated gate voltage.
The benefit of GaN is that you can switch extremely fast, but this poses its own set of problems. Switching really fast will excite the parasitic inductance and the output capacitance of the transistors, this can cause excessive EMI which may cause you to fail EMC. Careful layout will get you part of the way there, and possibly all the way there, but I often see snubber circuits that subsequently drive up losses and BOM cost. You also get additional losses from this phenomenon know as “impact ionization”, so the costs could end up cancelling out the benefits.
You use GaN in off-line (plugs into wall) converters because anything higher than 20W will be designed in two stages: a PFC boost stage that takes line voltage and boosts it to 400Vdc with PF > 0.99 and an isolated DC-DC step-down to the desired output voltage. GaN would be used in the PFC stage and on the DC-DC input stage.
Silicon and SiC have much more maturity and thus have more momentum behind their development. The issue with standard silicon is we are fast approaching the theoretical limits set by the bandgap and electron mobility, and so most improvements there nowadays are in the lower voltages where better packaging technology and substrate construction can have a significant impact.
There are still significant gains to be made with SiC and GaN, with GaN having much more room to grow.
12 Likes