Review Index:
Feedback

ASUS PQ321Q 31.5-in 4K 60 Hz Tiled Monitor Review

Author: Ryan Shrout
Subject: Displays
Manufacturer: ASUS

Setup Process on AMD and NVIDIA, Dual HDMI Testing

Unfortunately for ASUS, setup and use of the PQ321Q isn't quite as simple as plug and play quite yet.  In our initial testing video you saw some of our problems and conflicts getting the monitor up and running on both AMD and NVIDIA test beds - it was kind of annoying and embarrassing enough for NVIDIA to get them to push out a new driver to address SOME of the problems in under a week.  But let's talk about the better experience first with AMD.

 

Setup with AMD Radeon HD 7000-series

Connecting the ASUS PQ321Q to our AMD Radeon HD 7970 GHz Edition card via the included DisplayPort cable was by far the easier of our two experiences.  With the monitor turned on and MST mode enabled in the DisplayPort STREAM options, we booted into Windows right away but the OS and driver was seeing either head of the display as a unique monitor.  As a result, we had the start menu of Windows 8 on the left hand side and your standard desktop wallpaper on the right hand side.  But, right away, the screen was usable as AMD has identified each panel and set it up as a spaned display configuration, each at 1920x2160.

View Full Size

Great, but how do you set it up as a single display with a single resolution?  Eyefinity.  I entered into the AMD Catalyst Control Center and ran the standard Eyefinity configuration utility and created a two-panel option.  Boom - we were up and running.

View Full Size

As with any other Eyefinty result, the operating system sees a single panel, but we just happen to be running it at 3840x2160 with a refresh rate of 60 Hz. 

 

Setup with NVIDIA GeForce GTX 700-series

Things weren't quite as simple with NVIDIA as it turned out.  There are some bugs and issues that were addressed with some driver fixes they shared with me this week but there are other problems that won't be addressed until a new firmware is pushed out by ASUS and STMicro later in the month.  The first problem was an occasional ability to fail a cold boot - the motherboard would sit and stare at me with a "62" on the POST code and do nothing else until I hit the reset button.  Afterwards, it would start over and make its way into Windows nearly 100% of the time.  Annoying, but workable.

View Full Size

Another potential problem was the lack of ability to see the POST screen and BIOS during boot - most of the time our PQ321Q only showed something on the panel when we had entered Windows 8 and gotten to the login screen.  That can be pretty annoying and would require a user to have another monitor handy if any troubleshooting ever needed to happen with their rig.

In my experience it was also very hard to get anything on the screen if you don't at first boot into Windows with your NVIDIA graphics card with the monitor in SST mode rather than MST mode.  Once you are in Windows, you can switch the panel to MST mode and the system will recover, and you can reboot back into Windows each time after that without switching modes.

View Full Size

After talking with NVIDIA and ASUS about these problems, both companies seemed very confident that most of if not all would be addressed with a forth coming firmware updated provided by STMicro and pushed out by ASUS to its customers.  The current estimate for that release is the end of July so I'll be sure to test out the changes once we have the FW installed on our test unit.  (A recent driver release by NVIDIA helps with some other compatibility issues.)

After finally getting MST mode working over the DisplayPort cable with our NVIDIA GeForce GTX Titan graphics card, the rest of the setup is actually easier than you have with AMD.  NVIDIA has never supported 2x1 Surround configurations on GeForce graphics cards and that doesn't really change with these driver releases.  Instead, NVIDIA is looking for these panels as a matter of special case and configuring the setup automatically for you.  There is no need to go into the driver and setup a grouping or tell it which side is right or left, etc.

In truth, that is the promise of the upcoming VESA 1.3 standard that includes new requirements to make tiled displays a better supported category.  Once THAT firmware is released by ASUS (they are claiming later in the year) and with further tiled monitors, the display and graphics card will communicate without user intervention about how many heads are included, their orientation and refresh rates and set it up automatically.  NVIDIA is basically attempting to emulate this with updated drivers but it will be an industry wide feature once VESA 1.3 becomes the norm.

 

Dual HDMI Testing

I had several readers asking about the support for Dual HDMI mode with the ASUS PQ321Q and I have had mixed results.  Because none of the graphics cards on the market (or at least that I have here) include dual HDMI 1.4 outputs, I was forced to use one native HDMI connection and then adapt either a DVI or DP connection to HDMI for the second input.  With AMD, no collection of adapters would allow me to run 3840x2160 at 60 Hz and after a talk with AMD it seems that we need an ACTIVE DisplayPort to HDMI 1.4 adapter (which I have ordered but are still not yet available) to get it to work correctly.  You are able to run at 4K resolution with a 30 Hz refresh rate but that kind of defeats a primary selling point of this $3500 monitor.  So currently dual HDMI mode appears to be out for AMD cards.

The same was true for NVIDIA last week but over the last few days they were able to provide a brand new driver (not the one posted publicly this week) that was able to enable 3840x2160 @ 60 Hz using the native HDMI port and a passive DVI to HDMI adapter for the second input.

View Full Size

With this driver and configuration the system acted identically to when it was connected via DisplayPort with MST enabled.  Yes, that does include the POST and cold boot issues, but once in Windows it was seamless.

View Full Size

 

Obviously there is a lot left to be desired still for the ease of setup and configuration of the ASUS PQ321Q monitor.  The AMD setup experience was by far the easier of the two but NVIDIA made a lot of strides with a couple of driver revisions in the last seven days.  Most of the other problems with NVIDIA will likely persist until a firmware update or two filter their way out.

July 19, 2013 | 11:28 AM - Posted by Shambles (not verified)

Whenever I see an external power brick I automatically assume they were cutting as many corners as possible when designing the product. Not that I was going to drop all that money on a monitor anyways, but if I was going to I sure wouldn't be doing it on one that I can't use with a normal power plug.

July 19, 2013 | 01:23 PM - Posted by willmore (not verified)

I'll take an external power brick for any device that I want to live a long time. By separating the power supply and the device, you separate most of the MTBF of the devices. I've had power supplies fail in monitors and they're very hard to fix. Need a generic 19V power brick? That's COTS.

July 19, 2013 | 07:47 PM - Posted by Anonymous (not verified)

Agreed. After years of servicing equipment, I'll gladly take generic external power bricks over internal custom power setups. I even like that they're doing it with a lot of the budget PCs now.

July 19, 2013 | 07:48 PM - Posted by MoFoQ (not verified)

while it is true that most failures of LCD monitors are caused by failing internal power supplies, it is not true that they are difficult to fix.
Some can be a challenge to open, yes but once open, they tend to be easy to repair.
Typically they fall under two categories: blown capacitor(s) or arcing/cold solder on the input power lead. The last two monitors I've repaired, the input power lead had a cold solder which lead to it cracking...then arcing (tiny amounts) plus with the stress of the power cord being plugged in and out, the gap got big enough to do some serious arcing; enough to partial burn that part of the trace.
Fix was easy: clean, resolder or bypass.
For blown capacitors, the hardest part is ordering the right high-quality capacitors.

With that said, using an external power brick has a few additional advantages such as heat.

July 19, 2013 | 09:15 PM - Posted by Inflex (not verified)

$20 replacement power pack, versus (at least) 30 minutes of repair work and replacing $3 worth of caps. Economically in terms of service jobs the $20 replacement pack wins out.

Of course, if you're doing it yourself as a home job or for some fun, no worries, replacing caps is a walk in the park. That said, sometimes it's more complex than that and you just end up wasting time. Up here we get a lot of geckos in power supply units, a lot of the time the whole board is a write off.

November 11, 2013 | 03:33 PM - Posted by Rob (not verified)

> Whenever I see an external power brick I
> automatically assume they were cutting as many
> corners as possible when designing the product.

You do not explain WHY you assume that, nor do you explain why the Manufacturer MUST HAVE been wrong to have designed their Product that way; do you do a LOT of design, or a lot of ...

I would not want to assume about your decision process nor make too many assumptions about the design of something I was not involved in.

In addition to simplified replacement there is the consideration of weight distribution, thinness, ability to switch Power Supply Manufacturers easily (without affecting the portion of the Product where 'the Money' is), heat, interference, Electrical Standards in different Countries, etc. .

Repair may not be as cost effective whether the Power Supply were internal or external, UNLESS they could assume (like Computers) that the End User would most likely be able to do the repair / replacement themselves; which in the case of Monitors / TVs is NOT going to happen).

In addition, in the event that you are partially correct somehow, this is "one corner" that was cut; why assume they cut as many corners as possible. If you do not like a single decision they made in the design then you assume the rest is a bunch of junk.

May I suggest, without assuming for others, that is something not well understood, or perhaps well explained.

Rob - http://www.youtube.com/user/LowLightVideos

July 19, 2013 | 02:57 PM - Posted by brucek2

Thanks Ryan for an informative article that answered all of my most burning questions. As you mentioned there are more details to explore, but your top line findings are already enough for me to know that its probably best to hang back a bit before pressing that buy button (it sounds like they are going to get there though and that's great.)

July 20, 2013 | 06:57 AM - Posted by NLPsajeeth

Ryan that was an awesome article. Since I doubt anyone is going to risk opening a $3500 monitor I didn't think I'd ever find out what TCONs are driving the panel but you went above and beyond and got the info! Great Job!

It is still too bad NVIDIA refuses to support 2x1 and 2x2 surround for all monitors not just these tiled 4K displays. I was hoping that if a ton of these tiled displays hit the market, NVIDIA would be forced to add proper Surround support instead of white-listing monitors. But if what you said about the new standards comes to pass, they will probably be able to support these tiled monitors while still preventing generic 2x1 and 2x2 Surround. Too Bad.

DisplayPort 1.3 is under development. With all the delays I wonder what will get finalized first DP 1.3 or HDMI 2.0. I hope the DP 1.3 is as ahead of its time as DP 1.2 was in 2009. Here's hoping for 8K @ 120 Hz over a single cable. That would be awesome.

July 19, 2013 | 06:09 PM - Posted by Anonymous (not verified)

define "tiled" ?

July 20, 2013 | 08:30 AM - Posted by Ryan Shrout

Tiled means there are multiple heads in side the display.

July 19, 2013 | 06:27 PM - Posted by Anonymous (not verified)

Would be interesting to know if there's any additional early adopter problems by using two/three of these monitors. I suspect that drivers are full of bugs for this corner case. One obvious problem with Windows7 might be that the taskbar is stretched over both monitors because of Eyefinity/Surround must be enabled.

July 20, 2013 | 08:31 AM - Posted by Ryan Shrout

I would suspect that any issues that plague Eyefinity or Surround would be in play here, yes.

July 19, 2013 | 06:58 PM - Posted by EndTimeKeeper

Great Article, it makes me want one even more lol. I have a silly question how would this monitor work if you where to hook up something like a sony playstation 3 to it? The reason I'm curious is that I run both my PS3 and my computer on my monitor and was wondering if I could still run a setup like that if I was ever to get a monitor like this one?

July 20, 2013 | 08:28 AM - Posted by TMoney58 (not verified)

It works exactly like you'd expect. I have the PQ321 and I have my PS3 plugged into one of the HDMI ports while my PC uses the Display Port.

July 19, 2013 | 07:55 PM - Posted by Anonymous (not verified)

It's VESA DisplayPort 1.2 and VESA DisplayID 1.3 specs that enable this monitor. DisplayID is the next generation of the ancient VESA EDID standard. DisplayPort. 1.2 added MST support.

July 20, 2013 | 07:06 AM - Posted by NLPsajeeth

It was odd that Ryan was referring to DisplayPort 1.2 by name in the article and then started talk about "VESA 1.3". It seems that DisplayID v1.3 was indeed ratified by VESA in April or May. It certainly make sense for a monitor to advertise itself as tiled via DisplayID/EDID.

Since the spec was just finalized I'm guessing it will probably be a while before it makes its way into display controllers.

July 19, 2013 | 08:07 PM - Posted by Anonymous (not verified)

Thanks for the article. I'm one of those interested in knowing how the monitor scaled/pixel doubled from 1920 and that tiny bit of info about it being unsupported with MST saved me $3500. I guess I'll be waiting for the 4k/60hz TCON panels to come out.

July 19, 2013 | 08:26 PM - Posted by Sean Harlow (not verified)

"If you had the ability to run each side at 1920x2160 (taking up the entire left and right side of the screens) there might be more utility but ASUS told me that wasn't yet on the roadmap."

What's different about this than how the dual HDMI mode operates? Why couldn't you just put the monitor in that mode, but attach different devices to each port?

July 20, 2013 | 08:32 AM - Posted by Ryan Shrout

Perhaps nothing, but Dual HDMI assumes the two inputs are the same input.

July 19, 2013 | 11:13 PM - Posted by Branthog

I know it's lame, but even though this has such a high resolution, the 16:9 aspect really puts me off. Do we really enjoy swiveling our head left and right all the time at our desks?

July 21, 2013 | 09:05 PM - Posted by Anonymous (not verified)

Yes we do, that's a lot more enjoyable than looking up and down.

July 19, 2013 | 11:25 PM - Posted by Anonymous (not verified)

I'd be interested to see this display tested on different Apple hardware. I have a hard time finding information what for instance newer macbook (air/pro) handles. Also, no tests were done with Intel hardware, does that mean it's a no-go with Intel 4000 chipset?

July 20, 2013 | 07:16 AM - Posted by NLPsajeeth

Apple maintains strict control over video drivers on OS X so it is possible they may not support using 4K@60p with this monitor.

PCper did tests with Intel hardware in their unbox/preview video. Intel's Collage feature supports 2x1, 3x1, 4x1 and 4x4 monitor configurations and is available on Ivy Bridge (like the Intel HD 4000) and Haswell integrated GPUs.

July 19, 2013 | 11:28 PM - Posted by Anonymous (not verified)

Great article as always Ryan! There are a couple of things I would like to see going forward. First, 8x AA is way beyond overkill at 4k res to be sure. Let's look into exactly how much AA is need at 3840x2160. Show us more screenshots and in-game assets like that Skyrim tree at different modes and levels of AA. And let's also look into how much of a performance hit they will have. Though I imagine AA will cost about the same amount of performance it could possibly be different than what we are used to seeing with resolutions that are more commonly used today. Keep up the good work for us all!

July 20, 2013 | 02:18 AM - Posted by Hung Low (not verified)

correct me if I'm wrong. I have a rMBP, which has a GTX 650M and thunderbolt port. If I buy this monitor, I would be able to plug a display port cable into the TB port and drive the monitor?

July 20, 2013 | 07:10 AM - Posted by NLPsajeeth

Yes but 4K@60p may not work in OS X. NVIDIA had to put a special driver hack in the Windows driver to make the monitor work due to their artificially crippled drivers. Apple maintains tight control over the Mac OS X NVIDIA driver which is even more crippled than the Windows driver. It may not support tiled displays.

July 20, 2013 | 08:30 AM - Posted by Ryan Shrout

We hooked the MBA that Ken has up to this monitor through DisplayPort and could NOT go higher than 1920x1080.

August 9, 2013 | 11:06 PM - Posted by Luke Abell (not verified)

Ryan- have you tried hooking it up to a Macbook Pro Retina? Or the new 2012 iMac?

July 22, 2013 | 12:08 PM - Posted by mLocke

Will 120hz ever become mainstream? *sigh*

July 23, 2013 | 07:08 AM - Posted by AndyT (not verified)

"The first problem was an occasional ability to fail a cold boot"
+
"Another potential problem was the lack of ability to see the POST screen and BIOS during boot"
=
This is basically not consumer ready on nVidia hardware right now. Or is it just me that wouldn't want to put up with that level of crap from a $4500 GPU + display combination?

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

By submitting this form, you accept the Mollom privacy policy.