No-latency overclockable (up to 1080p@~270Hz) board preorder

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.
cirthix
Posts: 27
Joined: 20 Dec 2013, 15:59

No-latency overclockable (up to 1080p@~270Hz) board preorder

Post by cirthix » 15 Mar 2016, 06:51

Hello fellow monitor overclockers!

A couple of years ago, I created a 240Hz 1080p monitor. I never really got past getting a couple units out to a companies for development uses. Now that I'm done with school, I have had some time to clean up the design and make it suitable for production.

If you are interested in modifying your monitor or TV with a really fast controller, check out the video here:

https://www.youtube.com/watch?v=66fW9PV ... e=youtu.be

AustinClark
Posts: 42
Joined: 14 Mar 2015, 00:03

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by AustinClark » 15 Mar 2016, 17:42

I remember watching your original video in the past. I was really excited then, and I still am now. I'm at work at the moment, so I can't listen to what is being said in the videos, but I have a few questions.

I've found korean 1080p TN monitors that can, out of the box, be overclocked to similar rates. They're actually quite cheap as well. Wouldn't it be possible to use the boards out of these monitors and use them on other panels? If so, is your board more flexible?

Does it cause the terrible gamma-shift like the overclocked korean monitors get?

What technology are you using, FPGA? If so, have you considered making future updates possible in the field?

Is hardware overdrive possible in the future, realistically?

How close can a software solution come to true overdrive?

You're still at the mercy of the LCD controller it seems. What are the limitations there? Could those chips be over-volted or overclocked to increase maximum refresh rates even FURTHER?

Are there long-term effects on the panels after using this control board?

cirthix
Posts: 27
Joined: 20 Dec 2013, 15:59

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by cirthix » 15 Mar 2016, 20:03

The boards should work with any four channel 8-bit lvds panel. If the panel is only 2ch, it probably isn't interesting anyway. For 10bit stuff, you'll have to wait for the displayport version. For 8Vx1 and 8ch lvds, I have adapters designed but not yet built.

The boards in most of those Korean monitors use scalars which adds latency and usually causes frame-dropping when overclocking. There are Korean ones that are dumb converters, and with work, could be reused.

Gamma-shift on OC is within the panel itself.

The simple boards that you see here are are a combination of video ASIC plus a microcontroller sitting on the side. The microcontroller handles the buttons, backlight (strobing), configuring the ASIC , and swapping of EDIDs.

These boards are nicely modular and designed with fpga-use in mind. If you check out my site, I also designed the zis7020 fpga module which is awesome, but too expensive for monitor upgrades (around $300 in small quantities). Something like this would be needed for overdrive. If there is demand, I'd be happy to follow through with it, but I don't think people are willing to spend so much for this.

Let's just say that there is a master plan behind all of this design work and LCD is not the target ;).

AustinClark
Posts: 42
Joined: 14 Mar 2015, 00:03

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by AustinClark » 15 Mar 2016, 20:56

cirthix wrote:The boards should work with any four channel 8-bit lvds panel. If the panel is only 2ch, it probably isn't interesting anyway. For 10bit stuff, you'll have to wait for the displayport version. For 8Vx1 and 8ch lvds, I have adapters designed but not yet built.

The boards in most of those Korean monitors use scalars which adds latency and usually causes frame-dropping when overclocking. There are Korean ones that are dumb converters, and with work, could be reused.

Gamma-shift on OC is within the panel itself.

The simple boards that you see here are are a combination of video ASIC plus a microcontroller sitting on the side. The microcontroller handles the buttons, backlight (strobing), configuring the ASIC , and swapping of EDIDs.

These boards are nicely modular and designed with fpga-use in mind. If you check out my site, I also designed the zis7020 fpga module which is awesome, but too expensive for monitor upgrades (around $300 in small quantities). Something like this would be needed for overdrive. If there is demand, I'd be happy to follow through with it, but I don't think people are willing to spend so much for this.

Let's just say that there is a master plan behind all of this design work and LCD is not the target ;).
I assume OLED? That would be incredible. Dunno how you'd source the raw OLED panels though.

Know of any way to undo the gamma shift? Windows calibration doesn't go far enough, and I'm not sure how colors will be affected even if it did work.

On those dumb korean monitors, the ultimate limitation is the random "white-out" you experience when overclocking at extremely high refresh rates. Do you have similar problems with your boards? I'm assuming it's the LCD controller that fails, so I'd guess not.

Thanks for answering! I may purchase one. I've got an FPGA that I've been thinking of using for the exact same purpose, but I've had a tough time learning how to use it and finding technical information on monitors.

cirthix
Posts: 27
Joined: 20 Dec 2013, 15:59

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by cirthix » 15 Mar 2016, 21:21

AustinClark wrote:
cirthix wrote:The boards should work with any four channel 8-bit lvds panel. If the panel is only 2ch, it probably isn't interesting anyway. For 10bit stuff, you'll have to wait for the displayport version. For 8Vx1 and 8ch lvds, I have adapters designed but not yet built.

The boards in most of those Korean monitors use scalars which adds latency and usually causes frame-dropping when overclocking. There are Korean ones that are dumb converters, and with work, could be reused.

Gamma-shift on OC is within the panel itself.

The simple boards that you see here are are a combination of video ASIC plus a microcontroller sitting on the side. The microcontroller handles the buttons, backlight (strobing), configuring the ASIC , and swapping of EDIDs.

These boards are nicely modular and designed with fpga-use in mind. If you check out my site, I also designed the zis7020 fpga module which is awesome, but too expensive for monitor upgrades (around $300 in small quantities). Something like this would be needed for overdrive. If there is demand, I'd be happy to follow through with it, but I don't think people are willing to spend so much for this.

Let's just say that there is a master plan behind all of this design work and LCD is not the target ;).
I assume OLED? That would be incredible. Dunno how you'd source the raw OLED panels though.

Know of any way to undo the gamma shift? Windows calibration doesn't go far enough, and I'm not sure how colors will be affected even if it did work.

On those dumb korean monitors, the ultimate limitation is the random "white-out" you experience when overclocking at extremely high refresh rates. Do you have similar problems with your boards? I'm assuming it's the LCD controller that fails, so I'd guess not.

Thanks for answering! I may purchase one. I've got an FPGA that I've been thinking of using for the exact same purpose, but I've had a tough time learning how to use it and finding technical information on monitors.

I'm guessing you have one of those 27" samsung PLS panels? No fix for it that I know of. It might be the panel itself and not even in the tcon. Dunno.

For the korean ones that you're talking about, you're probably talking about the popular M240HW01-V8 24" 144Hz panel. I have a beta-tester which has tested with this panel. He can do 240Hz, but on darker images, the panel stops operating and goes full-white. I suspect that this has to do with an overcurrent protection limit on some powersupply in the TCON (remember: TN panels are normally-transparent and you use some energy to display black) due to how it is triggered, but have not investigated TCON modifications to fix the issue yet.

FPGAs are not easy to use and you'd need to generate a correct overdrive table to have it look good. Out of curiosity, what do you have?

AustinClark
Posts: 42
Joined: 14 Mar 2015, 00:03

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by AustinClark » 15 Mar 2016, 22:06

cirthix wrote: I'm guessing you have one of those 27" samsung PLS panels? No fix for it that I know of. It might be the panel itself and not even in the tcon. Dunno.

For the korean ones that you're talking about, you're probably talking about the popular M240HW01-V8 24" 144Hz panel. I have a beta-tester which has tested with this panel. He can do 240Hz, but on darker images, the panel stops operating and goes full-white. I suspect that this has to do with an overcurrent protection limit on some powersupply in the TCON (remember: TN panels are normally-transparent and you use some energy to display black) due to how it is triggered, but have not investigated TCON modifications to fix the issue yet.

FPGAs are not easy to use and you'd need to generate a correct overdrive table to have it look good. Out of curiosity, what do you have?
I don't own any Samsung panels.

As for the Korean monitors, I suspected it was a power issue as well. I did a lot of testing and tinkering with them awhile back. http://forums.blurbusters.com/viewtopic ... 5&start=30

I have the nexys 4 development board. I bought it prior to this project, but I'm pretty sure it supports LVDS just fine, and it's pretty powerful, so that shouldn't be an issue.

cirthix
Posts: 27
Joined: 20 Dec 2013, 15:59

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by cirthix » 15 Mar 2016, 22:38

AustinClark wrote:
cirthix wrote: I'm guessing you have one of those 27" samsung PLS panels? No fix for it that I know of. It might be the panel itself and not even in the tcon. Dunno.

For the korean ones that you're talking about, you're probably talking about the popular M240HW01-V8 24" 144Hz panel. I have a beta-tester which has tested with this panel. He can do 240Hz, but on darker images, the panel stops operating and goes full-white. I suspect that this has to do with an overcurrent protection limit on some powersupply in the TCON (remember: TN panels are normally-transparent and you use some energy to display black) due to how it is triggered, but have not investigated TCON modifications to fix the issue yet.

FPGAs are not easy to use and you'd need to generate a correct overdrive table to have it look good. Out of curiosity, what do you have?
I don't own any Samsung panels.

As for the Korean monitors, I suspected it was a power issue as well. I did a lot of testing and tinkering with them awhile back. http://forums.blurbusters.com/viewtopic ... 5&start=30

I have the nexys 4 development board. I bought it prior to this project, but I'm pretty sure it supports LVDS just fine, and it's pretty powerful, so that shouldn't be an issue.
I'm sticking to my theory of overcurrent protection in one of the powersupplies in the tcon.

Your fpga board doesn't have enough i/o. You need 2(pairs)*4(channels)*5(ABCKD)*2(inputs+outputs)=80 pins available.

AustinClark
Posts: 42
Joined: 14 Mar 2015, 00:03

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by AustinClark » 15 Mar 2016, 23:13

cirthix wrote:
I'm sticking to my theory of overcurrent protection in one of the powersupplies in the tcon.

Your fpga board doesn't have enough i/o. You need 2(pairs)*4(channels)*5(ABCKD)*2(inputs+outputs)=80 pins available.

Overcurrent protection makes the most sense to me as well. Hadn't considered that back when I was toying with it.

Darn about the I/O though. I assumed I needed just a few LVDS pairs. I saw a video about that here: https://www.youtube.com/watch?v=DIIjVSmbtFc

How/why is your solution different? I could increase the I/O, but then timing might become a problem.

cirthix
Posts: 27
Joined: 20 Dec 2013, 15:59

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by cirthix » 16 Mar 2016, 10:48

AustinClark wrote:
cirthix wrote:
I'm sticking to my theory of overcurrent protection in one of the powersupplies in the tcon.

Your fpga board doesn't have enough i/o. You need 2(pairs)*4(channels)*5(ABCKD)*2(inputs+outputs)=80 pins available.

Overcurrent protection makes the most sense to me as well. Hadn't considered that back when I was toying with it.

Darn about the I/O though. I assumed I needed just a few LVDS pairs. I saw a video about that here: https://www.youtube.com/watch?v=DIIjVSmbtFc

How/why is your solution different? I could increase the I/O, but then timing might become a problem.
More bandwidth, more wires.

What solution?

User avatar
lexlazootin
Posts: 1251
Joined: 16 Dec 2014, 02:57

Re: No-latency overclockable (up to 1080p@~270Hz) board preo

Post by lexlazootin » 16 Mar 2016, 20:00

This looks dope, i'll defiantly get one.

But i'm curious on how 'clean' or 'clear' it will look without overdrive compared to a 144hz monitor?

hopefully it doesn't take too long before you get 150 orders :)

Edit: How much does the actual panel matter? will some outperform others?

Post Reply