How to upgrade the Phoenix (Clevo P870DM-G)from FHD to 4K (PICTURE GUIDE)

Might be, check with Prema if there is a way to check the hardware values.

 

here is the file if anyone wants to give it a try.... not even my Plasma TV does it... my Envy23, old Sager, iPad, iMac, the two Eizo monitors I have one with 15,000 hours of use, display this defect

https://www.dropbox.com/s/rygrp8cwhcowv12/254vs2544.psd?dl=0

 

Of course is a bad thing, it means that the whole top tier of the gamma curve is incorrect, anything you do will at some point have white surrounding it, any text, a subtle shadow, how are you going to know that what you are seeing is a glitch of the screen or a real issue on the image you are delivering to your client? Like on this computer where I'm writing this now, which is a $700 machine, looks perfectly fine. And certainly not what you expect of an almost $700 upgrade screen.

Lets hope is either a glitch of my pannel, which can be replaced, or something it can be fixed by software. Have to see what the OP have to say once open the file I created. Let me know if you need it in JPG.

 

I'm going to test it with CalMAN tomorrow. I've read some reviews by others an no one mention anything, of course no one has review it so deeply as me, and I also I found that there are two versions of this panel, a version .01 and a version .03 with a few months of difference.

 

I havent seen the .03 anywhere tbh. Last i heard it was out of production. But it was supposed to be the better one.

 

Discontinued? Really?

 

Specific panel versions can only have very short runs.

 

Meaning? Sorry not sure I understand.

 

I got an idea from the guy that made DisplayCal of making a profile for 130 cdm with the monitor set to actually 140 cdm so the top part is not used.

This is the exact answer. :
It's not uncommon for LCD panels to behave weirdly near their channel maxima, but usually the electronics that drive the panel make sure that those are not reached. Maybe in your case the electronics weren't properly adjusted for the panel. There's a way around it though: Set a calibration white level that's slightly below the actual panel brightness. This will adjust the video card gamma tables so that the problematic zone near the LCD panel's limits is no longer used.

 

Here is the reason, this is the uncalibrated tone response of this screen based on the input, you can clearly see how red and green curves clip at the end, this means that the screen when receives 95% of red is already clipping to full red, when receiving about 98% of green is clippin gon the greens already, and blue never makes it over 228 when receiving 100%. This is uncalibrated tone response curve , so something is wrong with how the hardware of the panel is mapping the input signal value to the respective pixels. Clearly an error on the gamma curve at the end.

Probably the advise I was given could fix this? I still have to try, but that will mean I will be crushing the dynamic range and probably creating banding since I will be actually using values from 0 to 240 to cover 0 to 255

this needs to be passed to the manufacturer to fix it

 

This is barely visible, i know what im looking for, so i saw it. But my idiot friend couldn't see anything.

But i do think, that part of this is alleviated because of the pixel density. But then again, for a $700 investment, the results are darn good for that panel. I mean my EIZO costed me $3000, and during my masters i was using a $600 12 bit (Emulated) Dell Ultrasharp, which was OK by all means , but it couldn't even come close to the 16 bit panel. Point being, the value for money was pretty good, considering how good the Dell was with its < 1 Delta and 99.8% Coverage.

 

What did you looked at? My screenshot or the Photoshop file I posted? I thought you didn't have this panel but the regular HD one? You won't see the problem on that panel, only on the 4K one. And is because of the curve issue I posted. Believe me, my screenshot does not show how bad it is. That's why I was hoping the OP of this thread could open the file and tell me what he see, or the guys at Eurocom, but no luck yet.

 

In here, it is so vissible you won't believe it. And as is not flat, have you seen the curved I posted? The onion rings you see on a circular grading are white in the center, next ring in blue, next green, next red, and then starts to get better. But any GUI of a program for example with subtle gradient, will be all weird looking. Believe me. I'm not exaggerating. That's why I need someone else with this same panel to open this file and see if have the same issue.

 

Nono i checked this (the PSD) on a friends P870DM-G with the 4K AUO panel. But im guessing the difference is because his LCD is calibrate using one of the entry level cyclometers. (Colormunki Smile if i remember correctly)

 

You mean something like this:



Wasn't exactly sure what you meant but here's an image of an all white background I did in adobe elements if it helps identify the problem.

 

Not exactly, basically you need a Pure white - 255,255,255 as background, and a box on top of it (smaller size) with 254,254,254

 

download this, open it and tell me if you easily see the differences or ifg they are barely visible?

https://www.dropbox.com/s/rygrp8cwhcowv12/254vs2544.psd?dl=0

 

good news mixed with no so good news.

The recommendation I was given to set the screen at 140 for example but calibrate to 130 to avoid using the top part of the spectrum works, the bad news is there is a 400 Kelvin difference between 254,254,254 and 255,255,255 which still is a problem since it kind of kills the adjustment of the white point of the screen, and crushes the dynamic range by a big chunk.

 

There is a chance that he calibrated the screen by adjusting the screen by using the hotkey, and not the slider of the OS, with the slider and the BIOS I have now, I can adjust to the exact brightness I want, so that's why I got a dynamic range or like 1150:1 , but that's is using the whole range of the screen, if I skip the top reducing the dynamic range and not using it, then the defect gets highly dissimulated....

Still this should be fixed by the manufacturer and bad panels replaced if they can't be fixed. I have a year or warranty, so I hope Eurocom honors it and replace my screen if a new model without this issue comes out or if the manufacturer comes with a solution

 

Yes, its clearly visible on 4k, especially evening time. The clearest one in background, smaller square in center, smallest rectangle in mid top and all those numbers are different colours I checked morning, afternoon, evening and after various tasks like book reading, video editing, intensive gaming, excel charting, wth i thought this was a bug with eyes, now for sure its a hardware or calibration issue.

 

It's definitely a hardware issue. As I measured it and definitely the gamma curve of the Non calibrated monitor right at the end the green and red have a spike up in the last bit ending clipped and the blue channel have a dip ending non fully saturated. Which means that if you calibrate the monitor with the white point "as measured" and leave the monitor at full bright, first, you will end with a flawed calibration that will have weird colors on the gray close to white and that your white will be a bit red. You will see this issue even if you delete or turn off your calibration.

The panel also have a change in temperature caused by this of about 600 kelvin between 20% gray and 100% white measured by Spyder giving it a tone response of only 1 star. This is because this things are measured before calibration and the natural gamma response is all over the place.

The good thing is that it can be sort of fixed by calibrating in a certain way, still this elevated the overall delta error to a mediocre standard panel. Still looks good, better than the original HD monitor.

But this confirms to me that this panel have issues and that they should be fixed by the manufacturer if this is even a possibility.

 

FYI, I want to make very clear since in this form of communication is impossible to see the face of the person the other side of the screen, that I have no intention whatsoever of bashing, making the product a bad product, which is muuuuuch better than the HD one, and for what I tested so far way better than the average laptop screen. That's why I'm so surprised to see this glitch on the uncalibrated and correctly calibrated panel. It is hard to catch at the end because when you are watching things that you didn't make yourself you might think, OK, that line in the gradient might be part of the design, and just ignore it.

I was able last night to get something I believe is quite decent, but don't want to report it exactly until I have done a few more tests so I'm absolutely sure I'm not reporting something incorrect.

I do appreciate very much the effort of Prema, and Eurocom of getting this going, it's business and everyone is supposed to make money with what they sell and make it worth doing it.

I do recognize too that my standard for checking screens is way over the average user, and it comes from a specific need. Most of my clients requires me of the same product 2 photos, one over a gradient background and one over white. And even though I have no control over the monitor they will be using to show it, I do need to have control over mine and be able to trust it to defend my job if quality is questioned. Sound stupid but is really critical on Photograpy over white to check for halos or rings around the object which are usually not visible on most monitors because they clip the last few values, but they show up later on a print or iPads or new generation iMacs or Mac laptops with the new wide gamut screens that they are using now. Specially the 5K ones.

Please let me know too if this needs to be moved to a new thread.

 

FYI, the defect on the lower portion is even worst. uncalibrated and calibrated....

Here is the PSD

https://www.dropbox.com/s/o3unx2hrsok9ysx/000vs001.psd?dl=0

 

wow 005 looks more like 15-20% GRAY

 

in fact it is, the lower part is clipped at 10% and the gamma curve is all over the place. The color curve is gamma 2.2, the black like is the native gamma of this display, which is impressive if it was not clipped at 10%

 

I think one can think of this whole situation, as the following:

You have an amazing internet connection to listening to the best music ever mastered, you have the best home theather amplifier, and amazing equilizer (the gamma curve inthis case) of 256 bands (8 bits unfortunatelly thats the LUT table of the 980) and you first take the first 10 bands, lets say from 10 HZ, you can't hear it but you can feel it, my subwoofer by HSU RESEARCH and my chest and guts can testify it..... to 100 HZ and you just cut them out... you just set those channels to -96 DB, zero it.... so while you still have some base, you will feel no explosions, you will feel no sensation at all when that earthquake start, that subwoofer that weight 110 LBs, and can create a sound pressure equivalent to a jet engine at 100' is tunned out, then you take the top 5 bands, lets say from 13K to 15K and you put them at -20 DB, so every time a note hit them they will be muted, still there but muted, and then you take the 16K band an you set it to +96 DB, so it will be always distortion between 13K and up, but a bit mutted, and then the bass will be absolutelly clipped out...

So to fix this, you take an 8 bits frequency compressor, kind of like an autotune, to compress all the range to fix from the 10 HZ to 20,000 HZ to 100HZ to 13,000HZ and you add an easy in and an easy out curve to smooth the differnces, but your compressor is only 8 bits, so as you cut the 10 lower values and the 3 top values, you remap the whole thing to start ar 100 HZ and finish at 13,000 HZ, so what was 10 originally now is 100, what was 20 is now 115, so you will be missing 5 steps there, and not only that, it will be out of tune, this is what the LUT table does once loaded on the GPU, it relocate values and compress them to look more natural, but as their values are remapped with only 8 bits, instead of having a nice curve you have a serrated curve and most of the notes are out of tune becuase are shifted just a bit to fit.

Add to this that you have a compressor that takes care of level the power along the spectrum, (gamma curve = input value vs represented value) and that has a totally not flat response, it actually raises the volume on the already clipped bass, and lowers it on the mids, and then raise it back on the highs, with a difference of about 10% between extremes and middle....

the rest on the middle, it has a super nice tone, pleasant if you do not pay much attention to it, but still renders a nice sound that you like it, just do not have bass and the trebble is distorted clipping, making cricks and plops sounds every time that cimbal is hit... you just do not know what you are missing...

That is what those graphics I posted means, if this post makes any sense at all.

Lets hope AUO can change our equilizer and make the music sound pure without missing notes.

 

Can you explain those a bit easier, those numbers dont tell much to me and graphic is more like avc curve or nvda since 2016
https://www.tradingview.com/chart/?symbol=NASDAQ:NVDA

 

The ;-) means you are kidding or you really do not understand it? I have no problem to try to explain it if you really dont get it. I know is a bit complex.

 

Somone mentioned me that this could be fixed by an new EDID for the pannel, while the EDID contains a description of the gamma curve, it is just the gamma curve "perfect wold" value, it does not have any description of the response of the panel. in fact, I think there is no way yo fix that without propper hardware to process a LUT table, which these panels dont have, no panel have as far as I know, is extra hardware only on high end monitors. And can only be calibrated unit by unit. That's why monitors have a minimum black point and a max white point and the gamma curve takes care or linearize the response of the panel to the human vision. I'm starting to lose my hopes.... someone prove me wrong please? I will never be happier of being wrong than now....

 

Why do I see you as a tempxxxxxxxxx user? I could swear I saw your user name a few minutes ago.... you posted more than 7 times Glitch?

 

Those are implicit costs, forum policy burden, autobots correction .. whatever, doesnt bother me much.
I mean if its hardware level mistake for these 4k's do you think its possible to claim for a swap or reimbursement from reseller?

 

About the name, don't take me wrong. I was just curious because I thought you were someone else I've being chatting with.

About the hardware mistake, all the panels have errors, especially cheap panels mass produced. Unless you have an HP Dreamcolor 10 bits panel which cost about $1K for real, not like this one that cost about $200 or less in origin according to what others have posted in this thread. I have lived with this errors for years, and is just now that machines like the new Mac Laptops are coming with monitors that are wide gamut, which are also a proble to profile and color correct because most hardware is just calorimeters, not read spectrometers, ColorMunki Photo is a real Spectrometer.

So I don't know what to expect, I'm sure there is no bad intention from anyone involved in this and the panels are just fine for gamers in this case. What more you want to play games than a monitor that uncalibrated looks great (with errors but great) and that at maximum brightness have a contrast ratio of like 1200:1 !!! It's almost like being looking at an HDR monitor.

But for me or anyone like me expecting to do work with it, not knowing if the delivered product will have a black hole on the lower black or a halo on the higher white, it is a real issue. Especially the tone response curve, if I adjust to have a nice skin color on the shadow I will have a blue skin on the highlight. How you work with that?

It's a big dilemma, becuse on a scale of 1 to 100 this 4K being a 70 the regular HD one is a mere 10.

So I don't know what to expect. I wish I had more time and more money to purchase a better spectrometer like the one from Calman that cost like $700 but would allow me to make a much better read and may be even create a LUT curve to correct this display. That spectrometer make readings at like 10 times the speed of a ColorMunki. Each time I measure this is like 30 minutes

And if OP or anyone is getting tired of me talking about this and posting my findings on this thread please let me know. This is a fantastic community and I don't want to become a non grata persona here.

Thanks for reading to everyone.

 

@victorwol

EDIT:::I read the entire thing, and understood it

Have you tried discussing your findings with the producing company?

What if you got a bad unit of the display?

Did you try to talk with @Prema and see if it can be solved from within Bios settings? Maybe it is something that a small tweak will fix? I remember nVidia having very custom settings for Gamma ramps, and they caused a bit of problem in madVR usage too, if used un-properly.

Also, had you tried contacting the selling company and see if they are willing to change it, or do something about it?

 

Oh they are all participating and helping. Just to be clear, most users won't notice this issues. You will if you create your own content and you know how it should look or if you have an expensive monitor like I do to compare to. Visually I was not able to find this issues on my first tests. It all started to show up by a test I usually don't do. Because I've never seen it. Usually monitors behave the other way around. They wash out the blacks and they clip out the whites.

But be sure. If you are going to use it to play videos, play games, browse the Internet. You won't see it. And is way better than the stock one.

 

Thanks!

That is the most important aspect at the moment.

On the other hand, recently I have been working as one of those visual novels creator, and I actually must draw entire thing, master it, edit and master sound, and all. I ought to have the most true image possible. But then, every single day of working on these projects, I stop and think that 99% of players will have bad displays, and I should probably make very vivid and saturated and bright images to compensate for lower end quality of their own displays, making their experience better in the end. Too many variables here.

After studying the gamma curves you posed a few pages back, I understand that people at notebookcheck saying greens are a bit saturated, they actually might had observed that green band clipping a bit more.

From your perspective, is there any other downside with the 4K display, besides the white having that strange behavior.

 

I'm a bit busy at the moment but I will come back tonight with your answer and more.

 

So you actually have this display? That's what your signature says...

Being a creator and working on a wide gamut display like this one, you have to be very careful with that you do, unless you work with programs that understand you are on a wide gamut display, like Photoshop for example, and even with Photoshop you have to check that you are working in the right color space, in your case, if this is going to end being played on computers?? Then you have to work aiming to sRGB, which is a very poor crappy horrendous and vomitive color space :-( but is what 90% of the people will look at these days. Wide gamut devices are making their way into homes, but still pretty young on that accessible range for the masses.

If you are not careful, you could either see things on your computer that looks beautiful and colorful and amazing and looks dull and boring on others people computers. Specially if they are using an Internet browser that is not aware of color profiles embedded into the graphics. Or the other way around, things could look normal on your machine and when you look at them on a regular computer they are super saturated.

There is no magic on the color profiles, all the formats if we talk about 8 bits graphics to make it easier, have 256 possible values on each channel, so when you think how can 16M colors not being enough, just think is actually only 256 shades or R G and Bl make that monochrome and you have an image with only 256 shades, and that is if you use it all!! Think if you only go from 50% gray to 100% white, then you not only have 128 shades... Think how this work then on a 6 bits panel... They dither to make you think you see more colors, but you are not...

So the difference between profiles is how things are mapped, and there is when the importance of working on 16 bits files comes into play, even if you are going to deliver only 8 bits sRGB files.

At the end full red on any color profile will measure 255,0,0 when measured with the color picker, but on profiles like Prophoto, which goes beyond the human visible spectrum, that full red could be 100,0,0 in the RAW data, the software knowing this, will show you 255,0,0 and you will see a full red on an sRGB monitor, and a MUCH stronger red on a wide gamut display, but if you open this file with a software that does not read the color profile tag, or you save it with the incorrect one, and open this ProPhoto file then the default profile assigned to untagged images is sRGB, this red will look super dull and dark, (I'm exaggerating with the values here to make a case)

I, as a photographer, work everything from start to my end, not clients end, in ProPhoto 16 bits, because it allows me to keep in the file information that is darker than full black, and whiter than full white and bring it back at any time if I need to, this is not possible with AdobeRGB or sRGB.

Lightroom and photoshop are fully aware of this, and as the file have the right profile assigned, and the same profile is picked by these programs and compensated against the profile of the calibrated screen, then I can see on my wide gamut monitor the same you will see on your sRGB screen, and the reason why these programs have color proofing and gamut warning clippings. Because I can go if I want waaaaaaaaaay beyond what you can see on your sRGB monitor.

It all comes to how you convert later this image to what you are going to send to your client. If is destined to go to the net, you need to convert to sRGB, if is going to go to print, you could leave it as it is or convert it to AdobeRGB. You have to select the correct color rendering intent to covert though, because you can clip what is out of gamut, or you can compress the whole image to make everything stay inside gamut without clipping anything, being the last one the default one, and usually what looks the best.

Right now this monitor is clipping both ends, leaving you to work only with let's say 230 shades instead of the 256, being the clipping at the bottom much worst.. Again, it sounds a lot, but if you have a soft gradient, a subtle shadow, you have to resource to dithering or grain or noise to smooth out the gradients. And as the tone response of this monitor (gamma curve) but tone response is the right term, since gamma curve is what is applied to compensate actually, not a characteristic of the monitor, is wrong, the full white either because of the filters of each color are wrong, or the LED of the backlight is of bad quality, the white is not white, the R and B channel are like 15% stronger than the G channel at full power and it gets worst as you lower the power of the backlight to a point where there is almost a 50% of difference between the R B and the G channel, I've checked this in real time with DisplayCAL interactive calibration tool. This creates what I like to call rainbows on the gradients, as information needs to be mapped onto whatever is left when you color correct the monitor and apply a LUT table. And to make it worst, it seems that the look up tables of these Gforce cards are only 8 bits, not even 10 or 12 as cheap monitors, only 8. So you want to try to get a monitor that needs very little correction. And that is why when I tested this at full power before having the chance to lower the brightness I got such a good scores, because the response of the monitor is more linear at full power, but who can really work or use a monitor as 350 Lumens? Only if you are working outside at the beach? Used at 120 lumens as recommend by any Profesional or expert, the tone response of this screen is really bad, and tone response is not the same as color accuracy, the color accuracy is still acceptable, the tone response on the other hand is not. In fact the regular HD screen have a much better tone response and color accuracy at 120 lumens than this one, the only problem that kills the regular HD one, is that it have a few colors that are very bad, and that kills the screen, so you know, any delta error bigger than 3 can be discerned by a normal human vision, some more some less, as pretty much everyone has some degree of color blindness. The HD screen have pretty much all under delta 3 and some few tones with a delta of 8 or 9 which is another color altogether. Unfixable on a 6 bits pannel, even on an 8 bits pannel. And the UHD screen at 120 lumens have all of the range between delta 2 which is not discernible and delta 4.5 which makes an average of about 2.5, which is still considered good, but for this price should be almost all close to 1 or under.

Wow... This is a lot of unedited words, so I hope I'm really saying something

 

Thanks a lot for the very detailed response. I will try to expand as much as I can.

I am creating the visual novel from scratch, it is for play from Ren'Py engine, so for PC / MAC / Android, through a dedicated program.

I do not understand color spaces well at the moment, and never had to chose them, I started recently using photoshop CC 2015. I was using the free tools Gimp and Paint dot net before, and they worked great for my needs.

I am creating a background photo, which probably will be created in reality as oil painting, or photo and then heavy editing through more than one program.

I kind of understand what is going on, the other displat, LG display which I have at the moment, has minor black levels way too close, and most dark scenes do come out a bit too dark, there being too little difference between 0, 1, 2, 3, 4, 5 and 6.

I love the deep blacks though.



About already having the display, I am ordering it these days, and will have it, I am 99% decided to go with it, though I do not have it already.

 

Color management and quality output by Tom P Ashe is a good book to read.

Deep blacks are nice. But remember you are designing for people that might be running stuff that is totally off and either can see a deeper black or see a gray patch with noise and compression because there is one of the places where compression affects the most. And also see color bands around the dark areas.

Aim to work in AdobeRGB at 16 bits and deliver in sRGB at 8 bits would be my recommendation.

 

Thanks!

I am sure to check trying to work like that.

Should I make two versions of the game, one for sRGB and one for adobeRGB? Like, enabling the best performance for every person.

Or even better, save a version with sRGB art, and one with adobeRGB, detect monitor or display ID then apply depending on the display the correct art. Though this would consume way too much time that could be invested in creating a sequel or another game entirely.

 

DHL Just dropped this off :


Close up of the label, I saw that the sticker has something under it and was curious , have i been duped??? :



Nope , Exactly the same panel we have seen on ebay ( for $625 - http://www.ebay.com/itm/AUO-B173ZAN...039402?hash=item1a0d678b6a:g:xAEAAOSwKtlWsSGF)




@victorwol

Let me know which colorimeter / spectrometer you are using, i can try and see if i can get my hands on it and run some tests.

 

Mostly ColorMunki and some Spyder 4. ColorMunki is a spectrometer then it can profile any kind of regular or wide gamut displays since it does not use color filters, the Spyder is a colorimeter so it needs to have either the right filters to the right kind of monitor, or o be adjusted in its first read by a spectrometer and then apply the offsets by software. Usually done by factory and included on the software.

I'm pretty much convinced the issues I see are caused by a poor quality backlight. You see, this monitor uses a white LED, which is actually not white, it's blue covered by a yellow phosphore, rendering a white close to real white, but there but there is no thing as real white . All depends on the environment, but usually D65 is the standard for web design and Internet and D50 for printing, but D50 looks too red, although your eyes will get used to any color temperature in about 3 minutes and see white as white unless you have a brighter source to compare. Well, if the LED on the back of this panel is not good, which in my opinion is not, then you will have a super poor tone curve, which is that graphic I posted where the color temperature changes along the brightness scale in this case like over 500 kelvin, which is a lot. Color correction of the panel to flatten that line just causes solarization, so you either live with one thing or the other.

Also quality depends on the filters used on each pixel, and the linearity response they have, at 100 brightness this panel have a white point where RGB is almost acceptable, for a $200 panel but not acceptable at all for a $600 panel, at 140 CDM the RB channels are like 40% stronger than the green one, landing the panel on a 1 of 5 stars quality on Datacolor charts. Since it needs a lot of compensation.

Since others have confirmed my findings, I can only wish you good luck and that you got a better improved version, it would be good to see if we have the same revision or when your panel was manufactured.

 

Anyone working at Eurocom, does those stickers mean that is a version 2 of the panel with a firmware 2 ?

 

I wish I could by a spectracal C6 but unless someone donate the money, I can't spend more into this effort.....

 

First of all I need to start with a big thank you!

I've jut finished replacing the G-SYNC 1080p panel with the new 4K one using your guide. You probably saved my computer It was of great help.

Second, I can confirm that the stock bios for the GSYNC version does not want to run properly on the new panel no matter what. Windows won't allow any kind of driver to be forcefully or normally installed, while linux seems to load the drivers but any application crashes at any kind of GPU accelerated request.

Now if I got my bearing right there are two options. Flash the p870dm bios over the p870dm-g one or use the prema mod. For the first, I'll give it a shot in two minutes, for the second I can't seem to find a download link for the mod. Is it not public yet? Can anyone share it with me?

Thank you very much. Again awesome post and guide.

 

did you guys got problems installing GPU driver without mod on a DM-G machines after upgrading to 4k?

 

Yes, at least my unit did. I got stuck with this:



I've yet to find a way to fix it. Even more odd, I can't seem to create a usb stick with rufus to boot in dos and try to flash the non-G bios.

 

A full day of tinkering and what I learned the most is "read all the posts and every line". Took me half a day to switch off UEFI boot to get rufus to work. The public bios-es from clevo are for the motherboard, is there any similar link for the vbioses for the 980? ( https://biosmods.wordpress.com/stock/ only has 980M)

On the same line, how bad would be to flash 980M over 980 just to get it working until I manage to find the 980 non-g?

Victorwol, if I got it right you were using the desktop 980 too and performed the swap. Did you manage to get the GPU installed and run correctly with a @Prema mod? The twitter photo from his channel is about the 980M. Can you give any tips on whether it's 100% functional and definitely worth the wait? Also, G-SYNC is supposed to be linked to a capable display panel, you have it running on the 4k one?

Thanks

 

If you force flash a 980M vBIOS to a 980 then the system is dead.
To get g-sync you need a g-sync card and the Mod BIOS.
To be able to use a g-sync card with 4K even without g-sync you also need yet another Mod BIOS.

 

Here you will find the stock BIOS's :

http://forum.notebookreview.com/threads/clevo-p870dm-dm-g-drivers-mirror.786094/

But do not flash the 980m with the 980 BIOS and vice versa.

There you go. The lord himself answered it.

Victorwol got it installed by Eurocom i think, and they took care of everything from the Installation to the BIOS.