BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Talk about overclocking displays at a higher refresh rate. This includes homebrew, 165Hz, QNIX, Catleap, Overlord Tempest, SEIKI displays, certain HDTVs, and other overclockable displays.
User avatar
hleV
Posts: 105
Joined: 27 May 2017, 10:31

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by hleV » 16 Jun 2020, 08:25

sosuyike wrote:
16 Jun 2020, 05:01
And on CRU?
Identical to first post, just refresh rate set to 180.
BenQ XL2546K @ 240Hz (DyAC+) • ROCCAT Kone Pro Air @ 1000Hz • HyperX Alloy Origins • CORSAIR MM350 PRO Premium • HyperX Cloud Revolver • NVIDIA GeForce RTX 3070 @ 2130MHz/8000MHz • Intel Core i7-8700K @ 4.8GHz • G.SKILL RipjawsV 16GB (2x8GB) 3000MHz CL15

User avatar
exdablju
Posts: 13
Joined: 07 Mar 2020, 08:54

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by exdablju » 19 Jun 2020, 12:58

I was using AntiOutOfRange - OOR Fixer with no problem but like few days ago it did not launch with windows (i had it added on startup), so i launched it but no icon was showing up on the tray. Now it's only working for one resolution change. If i minimize my game or something OOR is popping up and ALT+Home or whatever the combination is (i tried everything) is not working anymore :( Like the app is launching only for once and then it's closing? Can someone help? :(

nnnn20
Posts: 11
Joined: 27 Aug 2019, 19:53

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by nnnn20 » 21 Jun 2020, 15:03

IMPORTANT: Older T-Series Will NOT work (all models older then XL2430T)
How come is this inaccurate information still being provided? I posted proof a while ago of a xl2420t running at 244hz without any frameskipping issues. As long as you keep the pixel clock below 330MHz (i believe this is a limitation of the monitor itself as there is no difference between the ports, with or without the pixel clock patcher) and keep the horizontal total below a certain value (around 1500 pixels), it seems to work just fine with a few visual quirks here and there. This is still very useful for people who play CS at lower and/or stretched resolutions. For instance, I can get 1350x1080p at 210hz and it still looks pretty decent for older games.

The limit appears to be at about ~264hz (directly on windows) and about 268hz on the OSD. It seems to vary randomly between 262hz-268hz on the OSD when setting it to 264hz. Up to 260hz it does not frameskip, but I'm not sure about higher. I tried setting firefox manually to 268hz (which was what the OSD was showing), but IDK if that is the correct thing to do in a case like this.

Image

general information:
- Anything above 330MHz starts to look glitchy with artifacts, even with DP or patched DVI-DL

- For whatever reason, most resolutions only work above 200hz, which might be why people did not realize it worked on these older monitors. A few resolutions work below this value, I don't really remember which, I think the resolutions that do not bug the scaler work below 200hz, but the others glitch out the screen.

- Up until ~230hz it works fine, with the only bug being that the screen gets slightly "misplaced", with the last vertical lines of the right side appearing on the left side of the monitor. This does not seem to happen at certain resolutions such as 640x480 and 800x600, but it happens for most of the other ones (probably something to do with the natively supported resolutions by the scaler?).

- above ~230hz there is a chance the screen will go full white requiring a reset of the monitor. this seems to happen when there is too much black on the screen at once for more than a few seconds. The higher the hz, the lesser the tolerance in time. (apparently a protection against overvoltage? i recall reading something like this)

- above ~244hz a black vertical line starts appearing in the middle of screen

- above 264hz (on windows) the monitor seems to just glitch out completely

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by Chief Blur Buster » 21 Jun 2020, 17:09

nnnn20 wrote:
21 Jun 2020, 15:03
How come is this inaccurate information still being provided? I posted proof a while ago of a xl2420t running at 244hz without any frameskipping issues. As long as you keep the pixel clock below 330MHz
I missed that when I was refrehing the OP. Done.

The mistake was because the XL2420T is an NVIDIA LightBoost model, while everything else since is BenQ-branded motion blur reduction mode. Blur Busters Strobe Utility doesn't work on XL2420T as a result, unlike all the other models mentioned.

Interesting exception. Thanks for that photo, that 260 Hz overclock is quite crazy high -- almost a 2:1 overlock margin! I wonder if minor adjustments to Horizontal numbers creates any change to those artifacts.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Margusx
Posts: 10
Joined: 11 Nov 2016, 10:44

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by Margusx » 23 Jun 2020, 05:47

My ZOWIE 2536 does only 200hz whay everyone ignoring this model or ppl not buy because not doing strobing at 60hz
For this monitor on internet is so little information nobody say nothing about this model. Its seems that monitor not exsist it does perfect less issues then 2411z. 2411z is pretty much dead Because DVI you can extend its life by buying active plug these cost lot. In new videocards have only DP and hdmi

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by Chief Blur Buster » 23 Jun 2020, 22:29

Margusx wrote:
23 Jun 2020, 05:47
My ZOWIE 2536 does only 200hz whay everyone ignoring this model or ppl not buy because not doing strobing at 60hz
For this monitor on internet is so little information nobody say nothing about this model. Its seems that monitor not exsist it does perfect less issues then 2411z. 2411z is pretty much dead Because DVI you can extend its life by buying active plug these cost lot. In new videocards have only DP and hdmi
Thanks for the success report of the 2536!

This be true, and that's why I steer most people to the XL2411P (it's like an XL2411Z but with DisplayPort, and still has 60Hz single-strobe).
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

nnnn20
Posts: 11
Joined: 27 Aug 2019, 19:53

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by nnnn20 » 24 Jun 2020, 10:22

Interesting exception. Thanks for that photo, that 260 Hz overclock is quite crazy high -- almost a 2:1 overlock margin! I wonder if minor adjustments to Horizontal numbers creates any change to those artifacts.
If you have any suggestions I'd happily try it >:)

Even with the glitches it is so awesome to use for CSGO and TF2, definitely worth it! Almost makes the desire of going back to CRT go away. Somehow the screen is overall very stable when it is at 268hz on the OSD, and the black line even works as a very sketchy crosshair :P. As long as there are no very dark screens in any game that get the screen to crash, it is perfectly usable.

It is a shame that I'm forced to use scaling to get it to work, but I believe it is not adding that much lag considering my times on humanbenchmark have even improved slightly, normally averaging below 130ms, often with clicks below 120ms. I still dream with averaging below 120ms consistently someday, perhaps on a lower latency mouse + CRT it would be possible.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by Chief Blur Buster » 24 Jun 2020, 11:31

nnnn20 wrote:
24 Jun 2020, 10:22
If you have any suggestions I'd happily try it >:)
First before understanding how to tweak... Look at high speed videos, www.blurbusters.com/scanout ... Notice how screens refresh top to bottom?

Cable signals are delivered left-to-right, top-to-bottom. Like a book or calendar. We've stuck to this raster sequence for a whopping century, from the first 1930s analog televisions, to the latest 2020s DisplayPort signals. The raster structure is the same in analog and digital!

There's actually also a left-to-right signal scanout behaviour, but it happens much faster than vertical scanout (one pixel every 1/330,000,000th of a second at 330 MHz Pixel Clock), so not possible to capture in high speed video, and besides, it's usually micropacketized as full pixel rows on modern digital signals nowadays anyways.

Now, the sequential serialization of a 2D picture into a 1D series of picture elements (aka "pixel") over a wire, still persists for 100 years, though terminology of pixel is muddy in analog era, the serialization concept of the raster scan is identical.

Since pixels are delivered left-to-right in a pixel row, the monitor processing may be too slow to receive the first pixels at the left edge. So a bigger Back Porch helps. Usually, you want a bigger Back Porch than a Front Porch, when you're overclocking a display, in order to give more guard delay for monitor processing to catch up.

To understand how numbers are mapped from CRU to a monitor screen, look at this diagram:

Image

Porches are overscan which is simply dummy black pixels beyond edges of screen, normally to compensate for curvature of CRT tubes, but it's still in digital signals as dummy black pixels as simple guard delays to allow the monitor to synchronize.

Now, in some monitor processors, some monitors process the screen in two halves (left half and right half). That thin black lines is highly suggestive of this monitor processing behavior, and sometimes cable transmit two separate signals representing left half, right half (dual links).

Artifacts related to dual links sometimes can result in a vertical line down the screen middle. The photos is a big clue. What it seems like is the monitor electronics is a little slow to keep up with synchronizing to the first pixels for each half, I'm assuming you're also probably missing a few pixels at the left edge too.

Back Porch is usually displayed in the software, but it's calculated as subtracting all numbers from Total.
Back Porch = (Total Pixels) - (Active Pixels) - (Front Porch) - (Sync Width)

So to increase Horizontal Back Porch, just increase Horizontal Total without changing other numbers.

Pixel Clock equals (Horizontal Total) x (Vertical Total) x (Refresh Rate). That's the bandwidth cap you have to work under, it's sort of like the speed of light: Cable transceivers (HDMI, DVI, DisplayPort), GPUs and monitor motherboard have their own absolute maximum (sometimes adjustable/hackable, but usually nonadjustable)..

Most monitors don't like horizontals not divisible by 8, so try to keep horizontals divisible by 8. So as a practical rule of thumb, your minimum possible Horizontal Total for full resolution without scaling is usually 1944. (1920 + 8 + 8 + 8) but sometimes you can "cheat" by reducing horizontal resolution, aka 1900 + 16 + 8 + 8 ... Now this may not apply here as you are stuck with scaling (fewer pixels per row).

This prevents you from exceeding your Pixel Clock maximum budget (330 MHz, I believe), while regenerating more headroom for back porches to allow the monitor to synchronize glitch-free. Now, you say you are forced to use scaling, so you may not be able to use full resolution for now, but adding more Horizontal Back Porch is still the trick that sometimes fixes vertical line issue of missed pixels at left edge (and vertical line down middle).

Analog history (if you're interested)
Syncs controlled the electron beam position on an old CRT tube. Horizontal Sync reset the beam to left edge, and Vertical Sync reset the beam to top edge. Porches were overscan (padding between sync & visible signal), to compensate for odd shapes of CRT tubes. You can see the old 1940s TV signal diagram at this web page. And you know those old NTSC TV bars patterns (the vertical color bars)? Here's some images, of what NTSC color bars looks like on an oscilloscope for one pixel row (analog scan line). Although there were some changes for color TV versus monochrome, like embedding the NTSC colorburst signal in the analog Horizontal Back Porch. But you see, those porches were there too, and still used in digital signals even to this day. Porches were simply an overscan area to pad around the signal rectangle that is sometimes cropped/underscanned on a circular TV tube, and this persisted even up to today. All computer signals (even variable refresh rate) still have these porches today, invented approximately 100 years ago as overscan. (Amazing how some things are unchanged, eh?)

Applied Knowledge
The great thing is that understanding this, can help a person how to adjust/fix overclock artifacts by transferring signal delays around to areas where monitor processing is most needed.

In that method, you MIGHT be able to make your black lines disappear. Since your lines are only approximately 2 pixels thick (4 total if you've got 2 pixel thick removals at both one side edge and the center), you might only need about a 8 pixel addition to Horizontal Back Porch to make that specific glitch disappear, giving your monitor more processing time between the end of the Horizontal Sync, and the beginning of refreshing the Horizontal Active Pixels.

Try Making Horizontal Back Porch Slightly Bigger
Try increasing Horizontal Back Porch size by 8 or 16. This adds more safety margin to the left edge. If Horizontal Back Porch is an 8 and Horizontal Front Porch is 16, try swapping the numbers, although you may need to leave Front Porch unchanged -- usually Front Porch doesn't help overclocking so you can keep it at minimum if there's no glitches from it (you can test changes to it though). If you can't do this without exceeding your Pixel Clock maximum (in this case, 330 MHz), try reducing refresh rate or reducing horizontal resolution.

Try Testing Larger Horizontal Resolutions Too, But With Bigger Horizontal Back Porch
Now if you succeed in removing the black lines, try testing out full or near full horizontal resolution but with a larger Horizontal Back Porch. This may not work since the scaler memory may not be fast enough to line-process that many pixels, especially if pixel count overhead exceeds scaling overhead during overclocked operation. Monitor motherboards are often a mix of microprocessors/FPGAs and/or ASICs, and overclocking weak links vary a lot between panels.

The fewer pixels per row, the easier it can process an overclocked horizontal scan rate (the "Refresh Rate" under "Horizontal" column) since that's less processing time available per pixel row; so the fewer pixels per horizontal row it needs to process, the easier it is to overclocks. But there's also a "preparation time" in a monitor processor to begin a new pixel row.... And that's where Horizontal Back Porch can help you overclock better with fewer glitches. It inserts a delay into the beginning of pixel row, to help monitor processing catch up.

Hope this helps!
You say you are forced to use scaling, but maybe you can find a way to skip scaling
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

nnnn20
Posts: 11
Joined: 27 Aug 2019, 19:53

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by nnnn20 » 24 Jun 2020, 14:21

Thank you very much for the detailed write up. I'll read it properly and give it a try later. I have been following your site ever since 2013 (which is what made me buy this monitor at the time actually!) so I'm somewhat familiar with the concepts, but never really had to tweak an overclock to this extent before, so just knowing the basics had been good enough.

I was testing it again a bit earlier, with different values, and surprising the monitor boots up at very high refresh rates without skipping frames, however with severe "scaling" bugs, but it is consistently bugged the same way up to around 480hz or more. NVIDIA didn't want to save my 480hz resolution though, so i can't test it :lol:

Image

440hz got saved though 8-)
Chief Blur Buster wrote:EDIT to add zoomed crop:
440 Hz Overclock of 144 Hz BenQ XL2420T!
(Very glitched, malfunctioning scaler, low resolution, but frameskipping free!)
Image

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: BenQ UNIVERSAL 144Hz->220Hz OVERCLOCK for 1080p 144 Hz

Post by Chief Blur Buster » 25 Jun 2020, 14:52

nnnn20 wrote:
24 Jun 2020, 14:21
440hz got saved though 8-)
Holy overclock!
Even though it massively glitched, it didn't frameskip!

Here's a zoomed crop with histogram adjustments:

440 Hz Overclock of 144 Hz BenQ XL2420T!
(Very glitched, malfunctioning scaler, low resolution, but frameskipping free!)

Image

May I see your 440 Hz CRU screenshot?

Although highly glitchy and practically unusuable, I wonder if there's any special scientific applications possible for a cheap low-resolution ultrahigh-Hz screen. (Vision research).
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Post Reply