GuitarG Posted October 24, 2024 Author Posted October 24, 2024 On 9/17/2024 at 12:17 AM, panda_zzz said: Greetings. I'm following your project, although at one time I chose dell m4800 and old alienware as platforms for experimentation). Could you please provide a link regarding possible support for 4G Decoding and ReBar in Coreboot? Because it is a prerequisite to upgrade mxm gpu to Ampere generation. Here you go: https://www.phoronix.com/news/Coreboot-Resizable-BAR EDIT: Haswell may have problem with Coreboot because thanks to Intel (TM) Boot Guard. Your mileage may vary. I had already tested this alternative BIOS simplifed version, and it works! Now I need to compile updated version with corrected bugs and graphical interface. Never though I would be making my own bios hahaha!
panda_zzz Posted October 25, 2024 Posted October 25, 2024 20 hours ago, GuitarG said: I had already tested this alternative BIOS simplifed version, and it works! Now I need to compile updated version with corrected bugs and graphical interface. very interesting, will look forward to hearing news of your success and following your work
GuitarG Posted November 17, 2024 Author Posted November 17, 2024 Adapter for faster debuging on spare 8570w, I will make same later for main one. I plan to make mini board with dual bios, one coreboot and second will be original hp one. I can test much faster now. Right now I can boot linux, but I have no visible BIOS and fans don't work on my compiled coreboot. BTW, did you know different flash IC's loads time is varies? You can tell by how long it takes from power button to visible BIOS logo. Macronix is x2 faster than original winbond on my main notebook 🙂 On left main one vs test on right. Both with latest updated firmware. Coreboot is even faster (but I don't have video for now) D2P0Obbj.mp4 1
GuitarG Posted January 28, 2025 Author Posted January 28, 2025 (In professor's voice from Futurama) Bad news everyone, my 8570W lately started to malfunction. Motherboard needs to be repaired, I checked CPU and RTX GPU separately, and they are okay, but in my modded 8570w I'm getting black screen freezes when CPU uses more power. Swapping in weak CPU works, but that is not point. So hold your fingers crossed for me making my PC alive again, thanks. 1 1
loopster Posted January 29, 2025 Posted January 29, 2025 @GuitarG In case you can't fix it, pm me, I got a spare one lying around which I am willing to donate for free to a mad lad for the good cause. Postage costs only. 1
GuitarG Posted January 31, 2025 Author Posted January 31, 2025 After more than five years with me, it was his time to go to notebook heaven. RIP, sweet prince. I have spare mobo, but I won't use RTX this time, I will go back to P5000. I will do cable mod and BIOS thing, even one more DIY heatsink work. So everyone else could go my way and make his laptop better. Thank you loopster, and everyone, for support. Someday I will mod some RTX GPU, again.. 1
GuitarG Posted February 7, 2025 Author Posted February 7, 2025 Well, my spare 8570W suddenly doesn't work, too. It does sometimes turn on, boots into windows, but mostly it is not booting and sometimes bluescreen occurs "ACPI Error". I never had seen such problem, it seems to me I may have curse on me, because even my third laptop, clevo I used recently, have "died". Three laptops in such short notice.. Damn. I have no other logic explanation for these issues, despite being very good at electronics and modding.. I will give myself some time to rest to come back later. After I visit some voodoo doctor..
martin778 Posted February 7, 2025 Posted February 7, 2025 ESD? I always zap things when it's cold outside.
GuitarG Posted February 11, 2025 Author Posted February 11, 2025 Okay, now I'm testing it, but it seems my 3940XM died. I changed CPU to something I had spare sitting in my CPU storage bin and it works, for now. I wanted to buy Legion 5 Pro for 1800$ with 4070, but I was hesitating, fortunately. I will buy 3840QM this time, 3940XM died not because of mobo or not having enough cooling, but because I bought it from Chinese shop and it had factory defect seen under some angles with direct light on it. Old board should be working, I will test it too sometime. Nowadays throwing 6000$ won't get you same quality as back in the days. I am happy to continue my thread! Long live the fighters ! ^_^ EDIT: Below you can see silicon die from naked eye view and under 4x microscope. It may be uneven heatsink or CPU that didn't pass QC in factory and been sold in gray market. It had two big circles on it from start and I had feeling it could die someday.
martin778 Posted February 11, 2025 Posted February 11, 2025 To be fair in the past we had CPU's running with more damage than that. If the silicon is damaged, the CPU should be either working or dead, not degrading over time.
GuitarG Posted February 13, 2025 Author Posted February 13, 2025 Interesting setup spotted in Asus laptop - third fan. It is funny to see similar approach to temperature problem as my second fan blowing additional air to GPU cooler
martin778 Posted February 15, 2025 Posted February 15, 2025 Not a fan of such solution (pun intended) - gaming laptops in general have a short lifecycle and finding a proprietary replacement fan will be a major PITA. At least ASUS used to apply LM from factory so you'd never have to pull the heatsink assembly apart. 1
GuitarG Posted February 18, 2025 Author Posted February 18, 2025 New board, new approach. This time results satisfy me more Soldered two pieces of metal to original mounting points, holds very hard and screws RTX very nicely. Now I have to sort, dissasembly and test a lot of copper pipes. And lastly, solder whole heatsink assembly
GuitarG Posted March 1, 2025 Author Posted March 1, 2025 Secret for success is to align first, then push harder second time. After baking like this first, then its screwed harder and heatsink only for max efficiency Now time for final tests, touch and black paint 1 1
Myro Posted March 1, 2025 Posted March 1, 2025 On 3/1/2025 at 9:16 PM, GuitarG said: Secret for success is to align first, then push harder second time. After baking like this first, then its screwed harder and heatsink only for max efficiency Now time for final tests, touch and black paint Great job! How are the temperatures under load? And what is that PCB doing in the corner of the heatsink? Looking forward to see the final build fully painted!
GuitarG Posted March 10, 2025 Author Posted March 10, 2025 On 3/1/2025 at 10:38 PM, Myro said: Great job! How are the temperatures under load? And what is that PCB doing in the corner of the heatsink? Looking forward to see the final build fully painted! Thanks ! Still ironing heatsink out, I forgot about different thermal pads dimensions. It shame I don't have some nice CNC machine. I played for a while and temps were up to 90C turbo CPU and GPU 70-80C depending on load. VRM or VRAM overheated once after 30 minutes, so I'm making corrections What PCB do you have in mind ? If first photo, it is donor board+case cutted up for aligning heatsink with highest homemade precision 🙂 I use damaged CPU and GPU board, screw heatsink on and clamp heatpipes. 1
Myro Posted March 10, 2025 Posted March 10, 2025 On 3/10/2025 at 7:16 PM, GuitarG said: Thanks ! Still ironing heatsink out, I forgot about different thermal pads dimensions. It shame I don't have some nice CNC machine. I played for a while and temps were up to 90C turbo CPU and GPU 70-80C depending on load. VRM or VRAM overheated once after 30 minutes, so I'm making corrections What PCB do you have in mind ? If first photo, it is donor board+case cutted up for aligning heatsink with highest homemade precision 🙂 I use damaged CPU and GPU board, screw heatsink on and clamp heatpipes. What corrections are you referring to? In my opinion, the most suitable solution would have been a custom-made heatsink, preferably copper. However, the biggest downside is that obtaining one is much more challenging. As for using the pieces of the donor board, that’s impressive! I never thought of that 🧐 I look forward to new updates from you)
GuitarG Posted March 16, 2025 Author Posted March 16, 2025 On 3/11/2025 at 12:05 AM, Myro said: What corrections are you referring to? In my opinion, the most suitable solution would have been a custom-made heatsink, preferably copper. However, the biggest downside is that obtaining one is much more challenging. True, best and very costly way, I got quoted almost 150$ for one ! And then is even bigger problem, heatpipes. It wasnt possible for me to set them up like I imagined. That is because of laptop non-negotiable dimensions like battery or RAM slots. Sadly. I circled two troublesome VRM's protrusions, depth for coil cooling is also different Now hardest part comes - to solder it. And if works like intended - paint job. 1 1
GuitarG Posted March 18, 2025 Author Posted March 18, 2025 As good as it can get now, benchmarks after "another patch": 1 1
GuitarG Posted April 3, 2025 Author Posted April 3, 2025 It was windows update that corrupted my me region along with bios .... I used windows 10 ssd for using nvflashk for vBIOS flash (it was faster than learning how to do same in ubuntu), used it for like 4 minutes, after successful flash I restarted... "WAIT FOR UPDATES BEING DONE" During those few seconds before I pull plug, it made so much destruction. I had to use different flash chip and search for original bios dump!! ME region destroyed laptop entered manufacture mode! And I used Windows Update blocker on that drive (it was "emergency safe win10" external ssd). Now it is working again. But let it be warning for everyone =] TL;DR - using Win10 for vBIOS flash broke my laptop BIOS. GPU is fine 1 1
loopster Posted April 4, 2025 Posted April 4, 2025 That's a great turn of events. I just came here to express my condolences, and now there's no funeral at all 🙂 Your initial announcement scared me so much that upon reading it, I rechecked and replaced the thermal pads on the VRMs of my freshly built-in RTX 3000 (upgraded from P3000 in ZBOOK 17 G1) with thicker ones yesterday evening. You wrote earlier that your VRMs were overheating during early testing. How did you notice or figure that out?
GuitarG Posted April 5, 2025 Author Posted April 5, 2025 On 4/4/2025 at 3:24 PM, loopster said: You wrote earlier that your VRMs were overheating during early testing. How did you notice or figure that out? Pretty simple, test using game/program taking between 90 to 100% usage of GPU, check GPU core temp with Msi Afterburner and if core temp is lower than 90 celsius, any sudden shutdown means power delivery overheat. Not all GPUs, especially those in laptop, lets you check those. I checked height of original heatsink and compared that to mine.
SuperMG Posted May 4, 2025 Posted May 4, 2025 On 4/5/2025 at 6:57 PM, GuitarG said: Pretty simple, test using game/program taking between 90 to 100% usage of GPU, check GPU core temp with Msi Afterburner and if core temp is lower than 90 celsius, any sudden shutdown means power delivery overheat. Not all GPUs, especially those in laptop, lets you check those. I checked height of original heatsink and compared that to mine. Hello. Will you ever test RTX 4060/4070 MXM 3.0A on your laptop? They are from X-Vsion and cost below 450€ They don't require 4G decoding as @panda_zzzgot the 4060 work on Alienware M17X R4 with optimus. No one else tested the GPU on eDP without optimus 1
SuperMG Posted May 6, 2025 Posted May 6, 2025 Hello guys. I tested RTX 3060 from X-Vsion on the first slot of my laptop without OPTIMUS. Nothing on the main screen but does output on the HDMI screen. I cannot go far because the cooler wasn't compatible and I had the three LED blinds lol 1
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now