HTC Vive Wireless Adapter Review: No Tether, No AMD

Headline seems a little misleading to a degree...

y. Performance was strong on two Intel-powered systems we tried, but after a software update, we also experienced problems connecting to the Vive Pro from these computers. In other words, if you use a Ryzen-powered computer or connect to a Vive Pro rather than a regular Vive, you may want to wait for updates.

This suggests its just more than AMD issues, which the headline does nothing to reflect.


Intel also played a critical role in developing the Vive Wireless Adapter. The company adapted its WiGig technology to enable wireless communication from the host PC to the untethered headset. Intel introduced WiGig at IDF 2014 and touted it as a wireless peripheral dock of sorts, but the technology failed to gain traction in that market. With the emergence of VR headsets, Intel saw an opportunity to revive the development of its high-bandwidth wireless data transmission solution.

This begs the question, based on the shenanigans that Intel has pulled lately as to "Did Intel play dirty in the firmware to ensure issues with AMD processors?" I can't say that they did, but it can make you wonder. They may not have, or may have... its circumstantial at least until someone examines the firmware closely.
 

husker

Distinguished
Oct 2, 2009
1,202
220
19,670
I have a Vive Pro and AMD system and the wireless adapter works great for me:
Processor: AMD RYZEN 7 2700X
MB: MSI X470 GAMING M7 AC
Graphics: GIGABYTE Radeon RX VEGA 64XTX
 

BaRoMeTrIc

Honorable
Jan 30, 2017
164
16
10,715


I'm sure it has more to do with Ryzens latency than an Intel conspiracy, or possibly a windows 10 issue with this adapter and ryzens systems
 


Yes, those are other possibilities. As I said, its circumstantial right now... and correlation does not always equal causation... too much not investigated/addressed to know what the true cause is at this time.
 


Can't speak about HTC, but It's certainly in Valve's best interest to make this work on as much hardware as possible, with as few hitches as possible. I also imagine Valve has enough money that they wouldn't care about the kickbacks. I *also* imagine Valve would come down pretty hard on HTC if they ever found out HTC did something like that.


#*&$* on the AMD compatibility...but the battery pack thing is hardly a problem...just put the battery pack on a belt and put the belt on over whatever you want to wear when you play. You could even paint the belt and pack yellow and pretend you're batman.
 

Jeff Fx

Reputable
Jan 2, 2015
328
0
4,780
>and there’s no question that once you cut the cord, you won’t want to go back to a wired VR experience.

I went back to wired from the TP Cast because of fuzzy video, and tracking glitches. I hoped this would be better, but it's being reported that there's no compression chip on the Vive wireless card, and that compression is being handled by the CPU, harming VR performance.

>The problem is, the wireless system doesn’t always work the way it should.

Which is why I'm sticking with wired until a wireless systems works as well as a wired installation.
 


Or it could be the fact that short of the PCIe x16 slot on their board all the other slots are PCIe 2.0. The article did state:

HTC’s instructions, which you can find here, suggest installing the WiGig card first. The expansion card features a PCIe x1 interface, which should make it compatible with any PCIe slot in your motherboard as long as it supports PCIe Gen3.

Per the manual from Asus the only slot that's PCIe 3.0 on that board is the top x16 slot. The rest are PCIe 2.0.

WiGig transmits at up to 7Gb/s while PCIe 2.0 x1 is limited to 5Gb/s. If the Vive is needing that bandwidth then that PCIe 2.0 limitation of the AMD chipsets will be an issue.

I guess it would depend on what slot he used for the X99 system as the only PCIe 2.0 slot is the x1 slot but all the PCIe x16 slots are 3.0. I think we need to see this tested with a newer Intel system and maybe a Threadripper system with PCIe 3.0.

Or it could just be a driver issue.
 
Nov 1, 2018
1
0
10
As a note, the claimed struggle with AMD isn't true for everyone. I'm running it perfectly fine on two different ryzen systems.
 


Well, It sounds like an idea... a good one at that... The PCIe 2.0 vs 3.0 limitation may very well be the culprit in this case. Doesn't Zen+ and the 4xx chipsets bump up the count on PCIe 3.0 slots? (i.e. 2700 or 2700x)
 

John Nemesh

Honorable
Mar 13, 2013
53
0
10,630
Yeah, maybe you guys need to look into the issues a bit more. I have a Ryzen 2700x system, Asus CH7 wifi motherboard, Asus ROG Strix Vega 64, and Vive Pro. The wireless kit works flawlessly for me. Some are speculating that it may be a problem with Nvidia on Ryzen systems causing the problem...but that would be something I would expect YOU to research before publishing an erroneous article with a misleading headline....
 


Nope. 16 lanes from the CPU are PCIe 3.0, the chipset are all PCIe 2.0:

uikcg3hVXONqfMDf.jpg




Not sure how nVidia on a Ryzen system would cause the wireless to have issues.

Either way it needs to be investigated more so by the developer than a tech publication.

The headline is not erroneous as people with Ryzen setups are having issues and it is not working as it should for everyone.
 

shabbo

Distinguished
Nov 29, 2011
31
4
18,545
PS5 with Navi will solve all these problems. By then AMD should have a better solution for wireless display. Intel is crap, I've had nothing but problems with intel's over-hyped beta-ware.
 

That begs the question, why not? something tells me that a dedicated PCIe card is not going to be a requirement for most of the wireless headsets that will be coming in the near future. More likely, you'll just plug a self-contained transmitter into your video card, just like how wired headsets attach.


A few bits of plastic and a short cable for $60. I guess that should be a good indication of how much markup there is on all this stuff.

Also, this article seemed to miss some of the device's best features...

Pros
- Receiver doubles as ears to complete Robo-Shrek costume.
- PCIe card gives your tempered glass gaming PC that 1990's dial-up modem look.
 
I highly suspect the problem with the Ryzen test system is the X370 chipset. I've ran into more than one PCIe card, x4 and x1, that refuse to work correctly, depending on the slot they are installed on X370. They all work fine in the x16 and x8 PCIe gen 3 slots coming straight off the CPU, but as soon as you install them in a chipset managed slot, no amount of tinkering can make the devices work correctly. Not a gen 3 issue as the cards don't need gen 3 to work correctly.

X470 sorts out the incompatibility issue.

I would recommend Tom's update their motherboard and try again, since investigating the superseded X370 chipset's PCIe compatibility problems seems a bit like a dead end.

It does beg the question in my mind. If there was nothing wrong with the X370 chipset, and it's still handles all incoming Ryzen CPUs perfectly fine, why did AMD go to the trouble of pushing out an X470 chipset that has an identical feature set (save for StoreMI which is not a feature of the chipset)? My personal suspicion is to quietly fix some edge case errata.
 


AMD is going to come out with something better than WiGig?

Just an FYI, AMD is part of the WiGig alliance. They helped to develop it. Intel had a bigger part, mainly because they do more with networking, but AMD still helped to define the system.

As well in networking the best NICs are typically Intel, who also helped to develop 10Gbe.



If thats true then that kind of kills the upgrade path for people who use the older chipset if they are wanting VR.

They might also be able to update the drivers for the chipset to fix it. Guess we will have to wait and see.

But I agree TH should retest with a X470 board and even a Threadripper system.
 
May not completely kill the upgrade path for X370 owners, but it certainly would limit hardware card support, provided there are alternative cards available (doubtful) that can be found to work without the issue, especially if HTC's software support is currently a mess.

On a personal level, I'm curious if the add-in card is just another temporary solution to push a product to market, a bit like the 1st gen TPCast setup, to keep the money train rolling while we wait for the VirtualLink USB-C Port equipped graphics cards to hit more general availability? Seems you should be able to buy something that would just plug into a VirtualLink USB-C port to broadcast your wireless VR, once the ports are available on more sanely priced gaming cards, and you can guarantee the ports on video cards are going to be hooked to good PCIe lanes as long as the graphics card is working correctly.

The new RTX cards have the VR ports now, and AMD is part of the consortium and will likely include the ports on future cards that could be considered for VR, so I really don't see why wireless VR going forward wouldn't make use of the newer VR ports instead of the gimmicky seeming WiGig card rerouting the PC's video output internally through the PCIe subsystem, which to me sounds like a recipe for just this sort of disaster. (Seems like one of the worst ways you could go about trying to engineer this particular mouse trap!)

Straight from VIVE's support site:

Having problems with the PCIe WiGig card?
If you're experiencing one of these issues after installing the PCIe WiGig® card:

● Your computer doesn't boot up.
● Your computer can't detect the PCIe WiGig® card.
● Your computer freezes.
● WiGig® can't connect.

Try the steps below.

1. Check if the PCIe WiGig® card is inserted firmly in the PCIe slot.
2. Check if the cable from the wireless link box is connected firmly to the card.
3. Reboot your computer.

If the problem persists, try installing the card in another PCIe slot.
Not sure why the author of the article has been waiting so long for tether free VR, since TPCast's solution at least appears to have been working in the marketplace, despite it's issues. If the second gen TPCast ever gets to market, I'm not actually sure what sort of solution VIVE would be trying to provide other than a higher priced alternative.
 
Nov 2, 2018
1
0
10
Thank you for this article! It saved me from spending another night trying to get mine to work. I have a Z370 MSI board, an I7-8700, 3200 MHZ Ram, and a GTX1080Ti. My Vive Pro is unusable for me with the wireless card. The judder and frame drops are severe. I've reinstalled Windows, moved cards, changed BIOS, and relocated the antenna. Nothing helps. I realize others don't have these issues, but some of us do, and my PC's specs should not be causing the issue. It's well above the requirements.
 
Nov 2, 2018
1
0
10
Plug it into the second x16 slot, switch your BIOS to force PCIe GEN3.

Get out of SteamVR Beta if you are in it, uninstall Oculus software if you have it, make sure you have async repro on, and set your supersampling to 120% or lower.

Best you can hope for, but I'm waiting myself for an update.(Vive Pro, 9700k, Z390)
 
Jun 29, 2018
88
3
135
All Ryzen Motherboards use PCIe 2.0 and not PCIe 3.0 for non GPU slots... this problem cant be solved whatsoever because the AMD chipset does not give any PCIe 3.0 lanes.

 
I think you're skipping over the part where Intel systems are having issues too, despite the PCIe gen 3 support, and there are usually 2 slots on Ryzen systems that are run from the CPU's gen 3 lanes. The author mentions trying multiple PCIe slots for the WiGig card, but doesn't mention if he tried it in either of the Ryzen system's gen 3 slots. To be thorough, test the card in all slots, despite having to move the GPU around. The loss in bandwidth to the GPU may not even be noticeable in this situation, but that too can be evaluated and thrown into the review.

The ultimate issue is HTC's poor decision to route video, all the way from it's end-point, back through the system, when the system is likely under it's highest workload. Can it be done? Sure, but not when the engineers can't wrap their head around coding it correctly.

Why did the engineers go this route, so they could handle things in software, save a chip on their setup, keep most of the setup internal to the PC, or perhaps to avoid a loop-back cable? Not worth it if the final product reliability is this flimsy. So much for the days of buying an expensive piece of equipment and expecting it to be able to perform as advertised.

I highly suspect HTC should invest a few dollars in more computer equipment in their dev lab that gives a better representation of systems that might be expected to run their product.
 
Nov 29, 2018
2
0
10
Have a Z97-Pro WiFi and having similar problems with computer freezing up while installing the WiGig driver. I have tried the second PCIe X16 slot with it set to Gen3 in the Bios and still no luck. I have also tried the first PCIe 1 slot above my Graphics Card but that can only be set to Gen1/Gen2 in the BIOS. Same results.

The card has the blue light in the back ... and I also tried another card with the same results. So unlikely to be the Wireless card. Guess my only hope right now is that ASUS updates the Z97 series BIOS or Intel release a new WiGig driver/install.