Was all ready to go with my 8800GTX on LN2 tonight, but had a setback.
Ill give you some backround (as usual) but I think "a picture is worth 1000 words...." and the pic is below. Foxconn Mars mobo, Q6600 G0, 8800GTX, Corsair HX620W psu, 2X1gb PC8500 Mushkin ram.
Had to revoltmod the 8800GTX for most of it, but I did it successfully the first time, and this time I felt was easy. Resistors are MUCH bigger on the 8800 than on any of the mobos ive worked on, and I felt I did a good job. Need revoltmodding from things loosening up from fan removal and GPU pot installation, as well a hardmodded with a VR the Vmem point instead of the pencil mod I had there previously.
I used 1/2 inch armaflex for insulation, and in hindsight I should have used 1/4 inch. The 1/2 inch WAY too thick and it made it difficult to mount the card, but I got it done.
I used vaseline (a small amount) in the PCIe slot (actually IN the slot) as per Gautam. I was told this is an acceptable substitute.
Boot up first time and get a "good POST" beep and then this....I immediately powered down. Checked connections, etc.....Booted again....same thing.
Any suggestions on what the problem could be, how to determine for sure, and how to fix?
ADDL INFO ADDED FROM MY REPLY ON ANOTHER FORUM CUT-N-PASTE
Mods were set to max resistance on each, thereby producing stock volts.
Temp was about 0C. I only slightly cooled the pot since I was just booting for the first time.
I didnt insulate the GPU PCB hardly at all. I insulated the pot like there is no tomorrow.
Here is what I did on the 8800 PCB.
1. Vmod Vcore/Vcore measure/Vmem measure and pencil modded Vmem itself(was working well on stock cooler). Was like this running fine with no issues for like months. Prior to step 2 I performed the Vmem mod with a VR. I really thought all Vmods were REALLY good on this one but who knows.
2. Pulled stock cooler. Covered all necessary areas with tape (core, 6 pin pcie slots, mobo connection area, sli connectors, variable resistors glued on, etc..
3. Sprayed 2 coats fairly heavily with FINE-L-COT-SR (as per Maxi's protocol)
4. Let dry 2 days.
5. Pulled tape, alcohol cleaned mobo connection and core.
6. 1/4 inch armaflex square cutout over core. No other insulation on GPU other than the silicone spray in the end, but put gray pipe foam insulation square on back of GPU in 1st mount attempt. There was so much armaflex on the the Tek9 that the screws werent long enough to go all the way through without some pressure, and said pressure disrupted the Vmem/Vcore measure point mods located on SMDs almost directly behind the core on the backside of the card.
7. Reperformed the 2 measure point mods without cleaning the silicone. I could see exactly where they pulled from, and the wires were allready taped exactly in position, and these are BIG SMDs not like those tiny suckers on the mobos. Wasnt a big issue and went well. I figured since I was soldering the wires back in the exact place where they were located to begin with, there was no silicone there to worry about.
8. Mounted to Tek-9 without any backside insulation. Had to use a reasonable amount of pressure but the foam armaflex gave way nice, and there was no obvious bowing to the PCB of the GPU.
9. Vaseline in the PCIe slot (sm amt) and addl armaflex around the PCIe slot on the mobo.
10. Connect. GPU thermometer temp was about 23C ambient, and I pulled the Tek9 down to 0C with LN2 prior to first boot.
Booted and saw that screen. I would spam del hoping to get into BIOS, and that screenshot I posted would change, but change to another fuzzy/nonsense something-or-another.
ANY ideas at all on what I can do to try and diagnose the problem or am I just dead in the water without any options? PCB and all surface mounts (SMDs, resistors, etc...) on the GPU looks outwardly fine to me. Other than the glossy silicone spray of course, and marks in non important areas with no apparent circuitry from gllue remnants, tape remnants, etc......
Maybe I got some silicone under my tape on the mobo connection area and a good GOOD alcohol SCRUBBING is in order? Its about the only thing I can think of....
Can I measure resistance across the Vcore mod point to ground and Vmem mod point to ground and compare against a functioning card to gather information?
Any of these ideas reasonable? Any others?
Here is the pic of what the monitor looks like
Ill give you some backround (as usual) but I think "a picture is worth 1000 words...." and the pic is below. Foxconn Mars mobo, Q6600 G0, 8800GTX, Corsair HX620W psu, 2X1gb PC8500 Mushkin ram.
Had to revoltmod the 8800GTX for most of it, but I did it successfully the first time, and this time I felt was easy. Resistors are MUCH bigger on the 8800 than on any of the mobos ive worked on, and I felt I did a good job. Need revoltmodding from things loosening up from fan removal and GPU pot installation, as well a hardmodded with a VR the Vmem point instead of the pencil mod I had there previously.
I used 1/2 inch armaflex for insulation, and in hindsight I should have used 1/4 inch. The 1/2 inch WAY too thick and it made it difficult to mount the card, but I got it done.
I used vaseline (a small amount) in the PCIe slot (actually IN the slot) as per Gautam. I was told this is an acceptable substitute.
Boot up first time and get a "good POST" beep and then this....I immediately powered down. Checked connections, etc.....Booted again....same thing.
Any suggestions on what the problem could be, how to determine for sure, and how to fix?
ADDL INFO ADDED FROM MY REPLY ON ANOTHER FORUM CUT-N-PASTE
Mods were set to max resistance on each, thereby producing stock volts.
Temp was about 0C. I only slightly cooled the pot since I was just booting for the first time.
I didnt insulate the GPU PCB hardly at all. I insulated the pot like there is no tomorrow.
Here is what I did on the 8800 PCB.
1. Vmod Vcore/Vcore measure/Vmem measure and pencil modded Vmem itself(was working well on stock cooler). Was like this running fine with no issues for like months. Prior to step 2 I performed the Vmem mod with a VR. I really thought all Vmods were REALLY good on this one but who knows.
2. Pulled stock cooler. Covered all necessary areas with tape (core, 6 pin pcie slots, mobo connection area, sli connectors, variable resistors glued on, etc..
3. Sprayed 2 coats fairly heavily with FINE-L-COT-SR (as per Maxi's protocol)
4. Let dry 2 days.
5. Pulled tape, alcohol cleaned mobo connection and core.
6. 1/4 inch armaflex square cutout over core. No other insulation on GPU other than the silicone spray in the end, but put gray pipe foam insulation square on back of GPU in 1st mount attempt. There was so much armaflex on the the Tek9 that the screws werent long enough to go all the way through without some pressure, and said pressure disrupted the Vmem/Vcore measure point mods located on SMDs almost directly behind the core on the backside of the card.
7. Reperformed the 2 measure point mods without cleaning the silicone. I could see exactly where they pulled from, and the wires were allready taped exactly in position, and these are BIG SMDs not like those tiny suckers on the mobos. Wasnt a big issue and went well. I figured since I was soldering the wires back in the exact place where they were located to begin with, there was no silicone there to worry about.
8. Mounted to Tek-9 without any backside insulation. Had to use a reasonable amount of pressure but the foam armaflex gave way nice, and there was no obvious bowing to the PCB of the GPU.
9. Vaseline in the PCIe slot (sm amt) and addl armaflex around the PCIe slot on the mobo.
10. Connect. GPU thermometer temp was about 23C ambient, and I pulled the Tek9 down to 0C with LN2 prior to first boot.
Booted and saw that screen. I would spam del hoping to get into BIOS, and that screenshot I posted would change, but change to another fuzzy/nonsense something-or-another.
ANY ideas at all on what I can do to try and diagnose the problem or am I just dead in the water without any options? PCB and all surface mounts (SMDs, resistors, etc...) on the GPU looks outwardly fine to me. Other than the glossy silicone spray of course, and marks in non important areas with no apparent circuitry from gllue remnants, tape remnants, etc......
Maybe I got some silicone under my tape on the mobo connection area and a good GOOD alcohol SCRUBBING is in order? Its about the only thing I can think of....
Can I measure resistance across the Vcore mod point to ground and Vmem mod point to ground and compare against a functioning card to gather information?
Any of these ideas reasonable? Any others?
Here is the pic of what the monitor looks like
