HNHacker News
TopNewBestAskShowJobs

kanetw

468 karma · joined April 10, 2016

submissionscomments
kanetw··on EU reaches deal to make USB-C a common charger for most electronic devices
The 2 resistor on CC line thing is not spec compliant.

E: you also need vRd sensing to determine the available current.

kanetw··on EU reaches deal to make USB-C a common charger for most electronic devices
Absolutely.

On top of that, I feel like USB-C is deliberately engineered to sell more ICs. Which wouldn't be that huge of an issue except as a result, all relevant ICs are sold out (or unsuitable).

CC/PD use a custom transmission standard so you can't implement it without a PHY, which basically means at least one dedicated IC.

kanetw··on EU reaches deal to make USB-C a common charger for most electronic devices
Yep. If there's no port protection IC (and even then it might not catch it), if you unplug the USB-C connector you can potentially get 20V+ on the data lines. Goodbye data lines.
kanetw··on EU reaches deal to make USB-C a common charger for most electronic devices
As a user, I liked USB-C. Until I had to implement this ridiculously overcomplicated garbage.

To be spec-conform, you need at least 1 IC. You can theoretically hack a solution together with some resistors, but it's not spec conform. If you need the high-speed lanes, add a mux/redriver to un-flip it.

If you want power delivery (>15W), you need a PD controller and port protection (or a user will fry your data lines when they unplug the connector and put 20V on the data pins). 2 ICs right there, and one of them is basically a microcontroller, so you need to deal with more programming.

If you want alternate mode, you need to implement the entire PD stack and a redriver/mux. That's 2-3 ICs right there that have to work together (so usually a single vendor). And not all of them support alt mode.

Which, ok, fine. I'm not building a cost optimized product, I just want it to work well. Except literally none of the ICs are available. Because USB-C requires ICs for everything, it's all sold out (or total garbage that's not worth designing into a product, or requires vendor support to design the firmware, or or or).

kanetw··on Paying for Productivity: Premium Workstations over Time
> Still probably should have a spec sheet, since your target demographic probably cares quite a bit more than average. I've suggested to George to add a more prominent spec sheet. It's there, but you need to scroll a bit or go to the shop, so it's not the most visible.

> There’s a general sense that it was designed by committee. Even if it was, consumers have come to expect Apple like fit and feel and polish and unity in the presentation of any electronic product over $1000. Not sure exactly what you mean here, or rather where you got the impression. Can you elaborate?

kanetw··on Simular VR: Lenses and Schematics
Not sure how other headsets do IPD, but with eye tracking it's pretty hard to get it wrong. If the eye's off center it's easily noticeable in the image.
kanetw··on Simular VR: Lenses and Schematics
https://uploadvr.com/samsung-wide-fov-patent/

VR purposes. This might be what was used in the Odyssey, come to think of it.

kanetw··on Simular VR: Lenses and Schematics
If Microsoft decides to give us access to the APIs to implement a window manager (however you'd do that on Windows...), possible. As is, would require a significant redesign of our approach.

That being said, tethered mode will be Windows compatible.

kanetw··on Simular VR: Lenses and Schematics
You can gaze anywhere without significant quality loss (slight PPD reduction past 20 degrees but it's not really noticeable and it's uncomfortable to have your eyes that far away from center)
kanetw··on Simular VR: Lenses and Schematics
Adjusting IPD is really annoying at times. Probably one of the main reasons I pushed for auto IPD in our feature set.
kanetw··on Simular VR: Lenses and Schematics
Infinite focal distance isn't really doable with the space constraints you have in a headset. We tried, it massively increases complexity for no real gain.
kanetw··on Simular VR: Lenses and Schematics
We're 2448x2448 and use optics that magnify the center more than the edges, leading to even higher resolution (36 PPD, about equivalent to a 27" 1920x1080 monitor 60cm away)[1]

[1] https://simulavr.com/blog/ppd-optics/

kanetw··on Simular VR: Lenses and Schematics
About 1-2m in our design.
kanetw··on Simular VR: Lenses and Schematics
There's a patent by Samsung on it, but no actual product.

There's always optics in front, so it's not immediately helpful. But a curved surface might simplify lens design.

kanetw··on Demoing Our Displays
I haven't gotten too deep into it myself, but the driver outputs a backlight PWM pulse a specific amount of time after a vsync. It's something I'll optimize at some point, but I have bigger problems to work on right now.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
Not with this hardware.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
The above formula only holds for diagonal FOV, which is higher than hFOV/vFOV.

The formula for e.g. PPD in the h-direction, assuming uniform pixel distribution, would be hPPD = hRes / hFOV. With a uniform distribution, 2448px at 100deg FOV would be 24.5.

We don't have uniform pixel distribution, as the human eye cannot resolve well where it can't gaze. So we magnify more wherever you can actually point your pupil at, and less towards the periphery. That way we retain eye-acuity while allowing us to "waste' less pixels where they aren't required.

We maintain >32 PPD within +- 15 degrees of eye rotation, which accounts for 86% of eye rotations.

kanetw··on Show HN: Simula One – Portable Linux VR Computer
It's indistinguishable from reality at 60 PPD (pixels-per-degree) in the foveal region. We're at 35.

Varjo has >60 PPD by having a high-res display for the foveal region and another for the periphery.

It's doable with our approach to optics (increased magnification in the foveal region) with... about 4k x 4k?

kanetw··on Show HN: Simula One – Portable Linux VR Computer
Yes, the compute pack has several DP outs and can be used without the headset.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
I haven't turned my camera on in months, but I get your point.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
Not for this version, but definitely for subsequent ones-- assuming we have the volumes to make AMD consider us.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
Not yet using the prototypes (there's basically literally two samples of the displays we're using and we need them for dev) but once we have more of them produced, yeah, definitely. George used a Vive Pro for early software dev a lot.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
I'll post a through the lens image in a few weeks but this is equivalent to a 1080p monitor two feet away anywhere where you can direct your gaze. (Slight dropoff towards the edges but still over 30 PPD and 20um RMS spot size)
kanetw··on Show HN: Simula One – Portable Linux VR Computer
Correct. Visual-inertial SLAM using the AR passthrough cameras, with possible additional input from mmWave RADAR.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
The focal point in our headset is on the order of ~2m. Infinite conjugate adds significant difficulty and basically nobody does this (we tried in our first iteration). But 2m is comfortably far, imo
kanetw··on Show HN: Simula One – Portable Linux VR Computer
AR mode is tricky, and it's one of the things I need full integration for. This will take a few months to get the sensors, FPGA, etc. sorted.

So far with prototype units it was fine, but haven't tested for longer periods of time.

kanetw··on Show HN: Simula One – Portable Linux VR Computer
We have several external DP ports via USB-C. Web conferences... hm. Difficult to have some form of webcam with the headset on, but maybe there's a good way we'll discover down the line.
kanetw··on Show HN: Simula One – Portable Linux VR Computer
It's really incredibly hard to get a loan nowadays. And finding a VC that won't stick their nose in too much is also really hard.

George knows more about it, but even an established company I'm familiar with could not get a loan without personal guarantees from the CEO last year.

kanetw··on Show HN: Simula One – Portable Linux VR Computer
Thank you! We want to avoid being bought at all costs :)
kanetw··on Show HN: Simula One – Portable Linux VR Computer
Swappable isn't planned for the v1, yeah. V2, maybe.

Charge time should be about an hour (1C charging).

← PreviousPage 3 of 9Next →