The laptop is not booting up at all with it. I’ll try the second method after figuring out what the problem is. Just need to get to my multimeter first. But one thing’s for sure - I don’t have a dump for the 4 GB stick (Hynix chips). I had 2 different sticks installed, the other one is Corsair). Probably, I decided to not flash the other (now dead) one because the laptop was using the lower timings from the first stick. I hope it would be ok, if I flash SPD from the first one. But luckily, I got 2 sticks of PC3200 1GB one of which is not recognized, so I can clone the other one. Just need to figure out if the SPD chip is alive.
I feel alright with hotswap of RAM as long as your experience was positive. I’ll give it a go tomorrow.
About that wire, that’s very interesting… what is it made of? Is it metal? Cause if it’s fabric I really wonder: what is it there for?
It looks completely made of copper, just a buch of thin copper wires. What do you think?
Also, I’m interested in knowing where you made the incision on the pipe to pour the mercury in and how you sealed it.
I cut off the small end on the side of the CPU fan. After pouring the mercury in, I just used thermal glue to close the hole from the inside. When I was trying to remove the glue from the inside of the pipe, I came across the wire and just took it out with leftovers of the glue.
Ok, tomorrow I will send you this one modified for 1GB, and will also provide you different ones. I will see if I can find a generic 1GB Kingston DDR2.
I’ll keep you updated on this. Let’s see if I can sort this out myself with the other 1GB stick
I’m back with some findings.
The adapter I was using for the programmer wasn’t made well. The guy who made it forgot to solder the ground pin on the RAM connector and left in floating which is why the programmer wasn’t recognizing the SPD. I’ve fixed that and now it’s working.
Now, the sad part is that the RAM is actually dead because both SPD chips are alive and readable. I don’t have a hardware RAM tester to figure out which chips are dead, but I’m still wondering how they could be killed. Static electricity? I doubt it would happen twice in a row considering the random nature of this event and both sticks died one after another. My only guess is increased voltage difference between VCC on the SPD chip and GND. It shouldn’t be sharing any common pins with RAM chips, but the floating GND pin was the cause I guess.
I just picked up one of these absolute units for $100 on Ebay. Just has a little chassis damage, but came complete in box, and I mean complete. Seeing its a rare find, I too want to dabble in modding and upgrading it. It was something of a dream of mine in college but who could afford a $5000 “laptop?”
So anyway, I have an old MSI laptop laying around that’s about for parts, and its got an MXM Radeon HD 8970M, and I was dabbling into the possibility of swapping it in to replace the 2600XT, but seeing that its BIOS locked had me coming here. Looks like its not possible to upgrade the GPU from what I see so far, so time to upgrade everything else I guess haha.
He managed to make it work on 800MHz FSB by configuring motherboard straps to override the limitation of PM965. Then OC-ed CPU bus to 266 to gain the genuine perfomance of the processor. It also required to overclock RAM.
The chipset is thinking it’s working on 800Mhz FSB, but the real clock is 1066. I broke BSEL1 line for PM965 by removing one end of R520 and connected the nearest VCCP to it. After installing X9100 the RAM will get overclocked automatically as a result, so we need to increase RAM voltage for stability, otherwise you get errors during RAM tests. Something around 1.96v should be enough. If I recall correctly, I only changed the resistance of R148. I also increased VCCP because PM965 would probably need more power as well.
The biggest problem with quads is insertion of the APIC table which is bigger than the original one. I have asked for help from Maxinator500 from the biosmods forum, but he said he will be able to get to work from summer.
R148 should be around 75K and R147 around 47K. Remember to always measure exact resistance before installing the resistors. Run memtest and complete at least 1 pass to see if you get any errors. If you do, need to increase voltage using the following equation where R1 is R148 and R2 is R147, Vout = 1.96:
I was able to get as high as 4.25GHz with X9100, but I wasn’t even able to run a test because of BSODs and freezes. At max voltage 1.5V the CPU got as hot as 80 under no load even though I installed a huge copper heatsink with fans (but without heat pipes). I assume X9000 will behave in the same way.
You can adjust voltage through ThrottleStop. The connection between the CPU and the power controller is direct, so you can’t break it without cutting the traces if you want to install hardware switches for voltage control. Another option is to use a solder iron to remove the voltage control pins from the CPU. You will then be able to use testpoints on the motherboard to connect hardware switches. I did so on my Dell M6400, so I have one QX9300 with some pins removed
with ThrottleStop , I am limited to 1.2v for the CPU voltage I can not go above, do you know why?? I tried with ThrottleStop 5.0 and ThrottleStop 9.5. I would like to set a x9000 to 3.8ghz at least
I can get my hands on an x9100 for very cheap, does it boots and works straight on a stock HDX dragon 20.1?? my idea is to keep the stock 800mhz FSB and increase the multiplier with ThrottleStop.
Were you able to sort everything out your regarding ThrottleStop? I have your deleted messages saved in the email
X9100 should start at 800, but I don’t understand why not to install X9000 instead if you are not going to increase FSB
I have 2 hdx dragon, one has a x9000 es , the other has a t8100.
I want to swap the t8100 with an x9100 if possible…
the 1066mhz fsb operation seems risky.I need stability. I have 8gb of ram
is it possible to overclock the FSB via software??
yes , I solved the ThrottleStop mess! the UNLK button appears only if the cpu is an unlocked cpu(x9000, etc…). so t8100 could not .
Also the x9100 is theoretically more stable at higher clock than the x9000
I would like 800mhz fsb x 19 mutiplier to get 3800mhz
by the way what is your system+driver, are you on win7??
do you have some freeze sometimes due to driver compatibility issues?
1066 is risky? This is the config I’ve been working with for a couple of years and it’s working great. I only have some troubles with the touch panel behaving randomly, so I just turned it off (now I have more free space for resting my hands; profit ). But I can’t claim the FSB mod is the reason because I also did some other stuff to the motherboard like ATI-to-NVIDIA mod. X9100 is definitely a better option than X9000.
I’m using win7 x64. Which drivers are you asking about? I don’t experience any freezes.
I hardy believe you will be able to work continuously at 3.8GHz because of the fans spinning always at max and the temps going as high as 80 or more depending on the load. What thermal paste are you going to use for such high temps? Liquid metal would be the best option I think, but I have not tried it yet.
do you use a specific type of ram to get 1066mhz??
I might do the mod later because of you!!
first, I need to secure my purchase of x9100, I need to check its compatibility with stock 800mhz fsb.
so according to you x9100 can run at 800mhz and boot stock without mods?
nothing fancy, I am just using mx-4 thermal paste, my x9000 ES is doing 3400mhz with less than 55degree.
I have 8800m GTS, but sometimes the screen freeze or turn black when I plug the external monitor on the HDMI and vga port. I did not have any freeze when I uninstall the graphic driver on win8.1.
freezes are random. I do not know the exact prob.
I love soo much the dragon due to its design, ports and big screen.
if the MXM graphic card could be upgraded, that will be great!
the quad-core cpu would be an amazing upgrade as well!