Accessory Zisworks X28 DIY 4K 120Hz Monitor

[ChainedHope]

Source


Would you buy this? Their website currently has a price of $815 USD shipped worldwide for the full kit (29 inch) or you can buy kits without panels for $470 USD (28 inch) or $390 USD (39 inch) and would need to then find a TN or MA panel and make a cable for connecting the panel to their controllers. The upside to the kits is that you can get these for cheaper and find a cheaper/better panel online and you can do your own debugging on the controllers (ships with a usb to serial adapter) to better suit your needs.

website: http://zisworks.com/

 

[Kmpkt]

Seems like an awful lot of trouble to save a few hundred bucks unless this sort of thing is a hobby. Also seeing the way they ship the panels, handwritten instructions, etc. it seems like a VERY ghetto operation. I think where this could have more merit in the future is once larger platform 3D printing becomes mainstream and you can just print your enclosure from a supplied file.

 

[darksidecookie]

this seems really interesting, but i'll wait untill people have experimented enough with it and know what combinations of panels and controllers work(and the price has come down). After that it would be awesome if you could just pick you monitor/contoller pack and a custom designed enclosure like in the custom mechanical keyboard community(ramaworks design wise). And we would finely have a compelling gaming/work monitor that doesn't look like a spaceship.

 

[Ionrent]

 

 

[lhl]

For those interested in learning a bit more about the project, the designer has been hosting an ongoing AMA about the project, some of the answers which address @ceski's feeling about how monitor mfgr's are prioritizing developments (less about crippling and more about optimizing for pricing and profitability).

I dropped an email to Zis and asking about how closed/open the software is - basically the FPGA is closed, but the other microcontroller (basically Arduino-like) parts are open. Things open for hacking include custom EDIDs, modifying the backlight config, but probably not the actual LCD addressing.

Interesting bits:

In the x28 and x39 displays, the column driver is capable of around 276 thousand lines per second. At 4k, that's 276000/2160=127.7, roughly 120Hz. For lower resolutions, the image is scaled horizontally, but the vertical scaling is done by tricking the row drivers to pulse more than one row at a time. The column drivers are doing the exact same job, but instead of 276000/2160, you get 276000/1080=255Hz. That's why the system can't do 4k@480Hz. For higher line rates, you need either faster digital to analog converters or to split the panel into top/bottom or even rows odd rows and double the number of DAC channels that can be used. This is not done for cost reasons. The last display that I've seen this technique used on was the IBM T221 which was a 3840*2400 display from 2001 which came out with a ~35K usd pricetag iirc. It was an interesting display, truly pushing the boundaries of lcd tech at the time.

Input lag is 2 lines on the DP chips and 4 lines in the tcon+panel (plus up to three more in dual input mode). This results in typically 22~33 microseconds of lag (basically zero, most monitors have input lag measurements in the tens of milliseconds). 1080p mode for games looks better than a native 1080p panel due to the reduced screendoor effect. Sure, the scaling is fine for 540p, but its still 540p, its not going to look anywhere as good as the 2k or 4k modes.

Scanning is available at 4k120. Push one button to enable or disable it. Scanning is similar to strobing, but does not require the long vertical blanking periods. Blurbusters will have good coverage of the feature, I've been working to add additional flexibility to the firmware to make their testing easier.

The input boards on the x28/x39 do not implement overdrive. Scanning does a decent job of hiding most of the transition time on the 28" without adding much extra cost. Scanning on the x39 helps, but pixel transition time on the 39" is much longer, so it is mostly persistence of vision effect being improved.

The idea that the pixels must fully transition all possible g2g transitions before the next frame or additional refreshrate is wasted is a common misconception. Most of the pixels make it most of the way.

Hopefully the LTT video and the surrounding coverage (I hadn't heard about the project until seeing that video) bring up some sales. I think the project is awesome, although I just plunked down for a 32UD99-W and I just sort of have no need for that screen (still considering putting an order in, just because it's such a cool project). A couple notes: the shipping versions will have a case covering the electronics. There may be an upgrade at a later point on the DP2LVDS board at some point to DP1.3/1.4 that might also give adaptive sync support as well.

 

[ChainedHope]

I agree with most of the points you guys have made. I put in an order for the X29 w/ Panel and a few more panels to try from eBay. Interestingly the FPGA they are using I just happen to have a dev board and driver kit for... Might be able to hack into it depending on how they have secured it. All else fails I can probably pull the flashed memory off the board and get a look at how its operating even if I cant read the actual bit programming using some hardware tools I have lying around from my job 2 summers ago.