Discussion in 'HP' started by 2.0, Sep 3, 2010.
Ah i see, nice, do you have pics to share of what it looks like without the issue?
Sure, the last picture in this post http://forum.notebookreview.com/thr...-lounge-part-2.515212/page-1039#post-11038121
Seems to look fine to me. So in person would you say that totally fixed it, it looks nice?
Just give me a couple more days before I figure it out. I assume I messed up with video drivers.
Sorry for the delay guys. It is taking me longer because some artifacts have showed up. This is because of the increased refresh rate set up for the LCD. I just need to do some more testing.
Can I ask you to make another mod? It seems like I'm not getting artifacts at 92 Hz. I kept the Dragon on for the whole day and nothing showed up, so I think it's the safest top frequency.
Also, is it possible to put N201J4-L01 back? Instead of using one of empty slots for another frequency. Let's just have 60 Hz and 92 Hz and full name of the screen displayed as CMO N201J4-L01.
After decreasing by 1 Hz down to 91 Hz, the artifacts are gone from desktop, but still there on shutdown and startup screens. Feels like they are getting more resources compared to the desktop picture.
Ooops, just gotten the same artifact again.
See the dark stuff in the left and right corners. I keep looking for the safest frequency.
Any idea why this is happening? LCD CPU overheat or what? It just takes longer or shorter time for this to show up at different refresh rates. The higher the frequency is, the quicker the artifacts will show up.
It shows up more on startup and shutdown screens whereas it's less visible on desktop
This is on desktop. It's very small in the corner and there are some smaller parts of it between the folder and the corner
CMO2000 - N201J4-L01 (stock and mod).7z
Artifacting could be due to the controller chip overheating. The image on the display is drawn pixel by pixel, starting at the top left. Setting a higher refresh rate means the pixel change needs to happen faster in order to reach the bottom-right in time to start again at the top. This determines the necessary pixel clock and it depends on the sturdiness of the controller pcb and, later on, also on whether the actual lcd crystals can switch fast enough.
Mostly the artifact is in the form of a quickly traversing 'blip' across the screen. This being the pcb dropping behind the requested pixel redraw rate. Your 'dark corners' are a bit mysterious in that regard. Could you unclip the bezel and see whether that makes a difference? Perhaps it's just a bit of pressure on the backlight that only becomes visible at higher frequencies.
Might also try a few solid-colour backgrounds and verify it happens with all colours or only with red, green or blue.
I see. Well, it's not a big deal I guess since we can just use safer frequencies. 90 Hz seems to be working fine. Can you make another mod please?
The chip gets quite hot at around 110 and higher, I could start feeling it after keeping my finger on the chip for some time.
Are you talking about the metal mounts around the LCD on the 3 sides here?
This is at 120 Hz
I forgot to tell you that even though you put 2 different 105Hz options in EDID, I could see just 1 of them on the list. Is it possible to have both displayed?
One more thing, you forgot to put 0x06 at 0x79 in the last EDID
It's necessary for the vBIOS to accept the LCD panel
Ah, yes. So it's certainly not the clip; it's following the Windows design nicely.
You may want to check Blur Buster's Frame Skipping test.
Haha, yeah that figures .
Should've known that would happen; there's tons of edids around with multiple copies of the same refresh rate (and timing parameters), but only one is invisible to the user. Even though we use different timings it might still be considered the same effective frequency to the driver/OS so it will simply lump them together. They'd have to have a 1Hz in difference then.
Yes, but you'll have to choose; 0x79 is in the range of the last text string and 'CMO N201J4-L01' is too long to fit in a single text block, so resorted to 'CMO' for the first and 'N201J4-L01' for the second in order to get them both fitted (0x06 corresponds to a non-character).
Don't really see the purpose of the text blocks anyway, tbh. The 'CMO2000' is always present and is a unique identifier for a particular panel (its plug and play id), so the text blocks ought to be superfluous. With Samsung panels that wouldn't hold water, but that's because of either their designers' laziness, sloppiness or incompetence (or all three).
Hmm, well ... thinking about this ... you might have your cake and eat it too.
Most laptop panel edids are 128 bytes since they have a limited set of possible settings to drive them. Desktop panels need quite a bit more information in order to accommodate a larger set of scenarios. The edid specification therefor allow you to use extension blocks, each another 128 bytes extra. We could add the timing sets to a second 128 byte block instead:
CMO2000 - N201J4-L01 (edid v1.4, 8-bit, CMO, 60, 85, 89 and 90 Hz).bin
When using the Linux software tools this 256 byte edid needs the 'write-edid-256.sh' script rather than the normal 'write-edid.sh'.
Passed well at 108 Hz. Remember there were 2 different frequencies for 105 Hz earlier? I guess I used the one with more consuming set up. I flashed the 91 Hz mod and raised it up to 108 Hz and it seems to be working ok now. But I need to keep the laptop on for the whole day again to figure out if it's safe.
I can't get why the OS is showing 2 options of 59 Hz and 60 Hz to me instead of just 60 Hz? What's the reason for this? So instead of just 60 and 91, I have 59, 60 and 91 Hz on the list.
The one you've sent me earlier is good. Having just N201J4-L01 is enough
No need I guess. The standard block size would suffice as it would be enough having just 1 or 2 options. But your mod is working fine.
Let's try to make another mod for 108 Hz with the standard 127 byte block size this time if it doesn't take you much time because I've not tested it properly yet. I just kept the screen on for 5 hours this time, no artifacts showed up on it yet.
Can you make a 108 Hz mod based on this file please? It seems to be working fine. I left the laptop on for the whole day and got no artifacts. Worth trying
CMO2000 - N201J4-L01 (stock and mod).7z
If you look at the timing parameters then the refresh rate is hardly ever an exact integer value. The '59 Hz' and '60 Hz' values are more like 59.9x Hz (and a lot more decimals). So apparently the '59' is simply the integer value (which is always truncated down) and the '60' is the rounded value, so upwards in this case.
Another possibility is that the OS shows you the legacy NTSC refresh rate (~59.94 Hz) and, again, shows that one as the truncated '59' rather than the rounded '60' in order to have the distinction visible to the user. They could've just shown the decimals instead, but that might've been considered as further muddying the issue.
Separate names with a comma.