To settle this whole thing, I think Framework just needs to make a magnetic input module+cable combo that has the main job of protecting the laptop against spikes and surges from a misaligned or otherwise compromised magnetic connector. I’m sure the proper protection can fit inside the module space. And if it has to be power only with no data at the connector then I’m fine with that as well. This platform makes customization like this an option so I say they go for it.
I would gladly pop in a magnetic “input” (charging) module if I knew i was taking my laptop to a public or high traffic area where the possibility of someone tripping on the power cable is escalated.
I think with Snapdragon, RISC, and 16 model, cards are a low priority for them, which is fine. The fact that they don’t have something as simple as a dongle hider card yet even with the high demand I think means and magnetic connector, especially one that would have mac level handshake protocols is probably very low on the list. The volt might come out, but to me that’s the worst of both worlds if they use technology similar to what they have released. Data pins are open allowing for your board to fry, but super slow data that is barely useful in a USB 4 world.
So, to get back to the idea of moving forward here, would anyone be interested in helping me design a print for the two adapters that seem like they have the most promise?
Both of these are between 32 and 33 MM long, which is just a little shorter than the cable creations extender.
For the data blocker, I think that the inset for the magnetic connector would just have to be “moved” in ~3mm. I believe the width and height dimensions are close enough where that would be a fairly easy tweak, but I haven’t had the time to learn CAD or solid works to do it.
For the data connector, I think we’d have to go with a different design as it’s as thick as the card. I think for that one, the best solution would be a two part card holder, where the usb C side basically the cutout of the extender, and then you have a “cap” on the magnetic connector side which can be screwed into the base. Essentially a hollow for the extender, with the cap allowing more rigidity and support, but you’d still probably want to add some glue once you had a good fit. Happy to dissect any of these adapters as well to confirm pins are correctly attached if I get some additional interest here.
I’ve seen a lot of discourse about how these magnetic adapters are a big risk (what with arcing, simultaneous or near-simultaneous disconnect of CC vs. power pins, etc).
I haven’t, however, seen a lot of discourse wrt. alternative cable or connector setups that mitigate the “I tripped on my charging cable and sent my laptop flying off the table” issue, which, I would argue, is also a potentially device-breaking issue, especially for households with pets and children.
Are there other solutions to this issue that don’t risk frying the motherboard?
As a side note, I somewhat wonder if there are physical mitigation factors to the arcing issue to be had. Maybe using pogo pins with different travel lengths on the power vs. CC pins, so that the CC pin disconnects first, as with normal USB-PD?
I use short “extension” cables. The issue with laptops and tripping on cords is that, while the force is usually strongly enough that it could just pull the plug out of the port, the pull force is unlikely to be aligned straight with regard to the port, so that doesn’t happen. Cords are flexible, they lay in any position & can be pulled from any angle. But only straight can disconnect the plug rather than flinging the laptop. You see where I’m going with this? A short extension cord adds a second disconnection point, and this one will be in line with the pull force. Trip over a cable with a short extension connected on the laptop side & the force pulls the female extension port / main cable joint apart.
I’ve used a couple different short extensions in the past & they worked when someone tripped on a cable. At the moment I’m currently using extension cables that I’ve made myself, only because I’d also like the plug ends to be as short as physically possible.
Here are a few from a quick search.
0.5 M 240W 20Gbps Extension Cable amazon.com/dp/B0D11G9CX4
Another 0.5 M 240W 20Gbps Extension Cable amazon.com/dp/B0D1TXY73M
0.3 M Right Angle 240W 20Gbps Extension Cable amazon.com/dp/B0CZHT4129
I’d say you’d want no longer than 0.5 Meter / 1.6 feet.
I actually meant to post this solution, since I do use it myself and it’s worked, but got distracted with work.
I’d say the only way to do this right is put an unholy amount of esd diodes, spark gaps/tubes and whatever over-voltage protection measures you got on all the data pins, fortunately you do have quite a bit more space in an expansion card compared to the regular just a usb-c plug type thing so that may actually feasible but it’ll have to be done and tested which is likely not cheap.
Wouldn’t really work. You can’t put too many (or too beefy ESD) diodes on the data lines, they’d degrade the signal too much. Spark gaps might act as tiny antennas, so probably also not a good idea, and probably not very effective either. The only safe way I can image would be isolation over fibre-optics/opto-isolators.
Good point, ridiculously high speed data lines are a complicated thing.
You could be onto something here, not that this would not be freakishly expensive, looking at optical tb cables. But hey at least you’ll only fry your 100ish $ opto isolated magnetic card instead of the 300+ mainboard if something happens XD.
I am thinking that there must be some way of doing it safely.
I guess connecting is not a problem because it starts negotiation with only small power. So the problem is around disconnection.
electricity power does not actually flow in wires, it is the field that delivers power, so some waveguide inspired connector shape might help.
have a sensor that detects any movement of the connector to automatically disconnect power. Maybe use a laser pointer and detect as soon as it strays off centre.
But all of the above needs active power to control it, so not as power efficient and green as a normal usb-c connector so probably not appealing to most potential customers.
I guess it is a cost thing.
The solution most likely to allow safe magnetic connectors is probably something that has optic fibre for all control and data lines, and just copper or some other conducting material for the power.
If one looks at QSFP+ transceivers that do 40G as a guide for how much this might cost for the optical parts, it could be about $100-$400 each end. There is also the problem of how to keep the optics clean when connected/disconnected in the relatively dirty laptop environment.
So a nice cheap usb-c is probably the best solution and avoid magnetic connectors for now.
Opto-isolate the data pins sounds doable but expensive, you’d basically need to fit both ends of a fiberobtic usb4 cable in an expansion card which looks plausible.
For the power bit you’d mostly just need to spam protection on the cc pins though having early unplug detection would be neat and could definitely help not blowing up you expensive optocoupled setup.
You can apparently get optical tb4(usb4 cables for less than 100$ so it would not be quite that bad.
Doing it right would also be quite expensive making it even less appealing. A relatively cheap power only option may still have a relatively large customer base though.
Huh? at DC the power flow is electron flow, not microwave energy. Even at the 10GHz differential pair data lines it is still going to be electron flow. Any field generated is as a result of electron flow to start with. For waveguides you need some form of launcher to set uop the field.
@Alan_Pearce
You might think that and I thought that also because my school teacher told me so. That was until someone mentioned maxwell equations and poynting vectors to me.
I like this video explaining it a bit:
I don’t think much of what I was taught in science at school has turned out to be true.
I bit off topic, but there is a title for some scientist “Particle Physics”, except that when you actually go to study the equations. e.g. “quantum chromodynamics (QCD)” you then find out that particles don’t actually exist, and it is just “fields”. I was a bit amused when a “Particle Physicist” told me “Particles don’t exist”. I then ask, In that case, why did you call yourself a “Particle Physicist”? They just laughed and said, well we thought they existed when be came up with the title, by now we know better.
Yeah, but you require electron movement to get the magnetic field that applies to Maxwells equations. Without electron movement no magnetic field.
Magnetic fields (in inductors) or voltage fields (in capacitors) are what store energy. Voltage fields are caused by extra electrons on one plate, and reduced electrons on the other.
If you can find an opto-isolator that is fast enough sure.
Putting both ends of the electronics in an optical usb4 cable would be essentially an opto isolator but the electronics capable of doing it must exist if optical usb4 cables exist. Last time I was trying to find opto-isolators for merely 5gbit usb3 I found nothing.
Do you mean for high speed data?
Or for just power delivery / protection of the CC lines needed for USB-PD?
If you want data, that won’t be trivial. And I believe the cost will be prohibitive.
You can check major parts suppliers and see the speeds available for simple “high speed” opto-isolators. Mouser link, Digikey link. You won’t even get USB 2.0 speeds with those chips.
If you want thunderbolt / USB 4, you’ll be creating the equivalent of what’s inside a fiber optic thunderbolt USB cable. And look at the price of a 40Gbps thunderbolt USB cable. I’m seeing hundreds of dollars. Even slower cables are not cheap. The fiber part of the cable is not that expensive at all, I don’t believe it’s contributing much to the price. Rather it should be the plug ends, the parts you need, where the cost is.
Edit with a bit more context: Volta is working on releasing their Volta MAX cable, which supports USB 3.2 20Gbps and 240w, which is what the Framework 16 runs at. Scroll all the way down on this Kickstarter where they officially announced it a while back: https://www.kickstarter.com/projects/voltamagneticcharger/volta-max-the-cable-for-all-fast-charging-and-data-solutions there is no new information here, but it is the first time Volta or Framework has officially spoken on this. Pictures are same as I’ve released.
This stuff also doesn’t look any different from the magnetic usb cables you can already get from china and if it doesn’t add any protection you may as well get those, I have one that does usb4 and that was less than 10 bucks (I only tested it, this stuff is way too sketchy for me to actually use it and that is coming from someone that smears liquid metal on everything XD).
