Hi, after some time a friend of mine got the RROD, he asked me if I could fix it, I told him that I could try to fix it since I have acces to a BGA soldering station, so I can desolder and sold again the CPU and the GPU. After reading a while on the internet about the origin of the RROD I found out that the origin of the RROD is the cheap X-Clamp microsoft used in the console, so my question is, If I resoldered the CPU and GPU and replaced the X-Clamps with screws like the ones that are shown here will it solve the RROD? Thank you for your time.
Its the lead free solder to add to the problem. If your going to reball the gpu and cpu than yes it will fix your problem,just don't use lead free solder.
You know I was just about to start a topic on this very thing. Let me give you the full run down on the RROD issue and possible BGA resoldering issues. The RROD seems to arise from board flexing. The CPU seems to be quite alright by most standards and it appears solely to be a GPU and scaler problem (the ANA/HANA chip directly above the GPU on the board). When the GPU fails it causes the RROD, when the scaler fails it appears to cause the E74 error. This is all speculation at large as I have seen some mighty convincing proof from people with expensive equipment that I don't have. The main homebrew fix for the RROD is the x-clamp replacement fix. The method I typically employ requires one to remove the x-clamps and all related hardware, such as the posts, and replace them with 4 washers and a screw per hole on the heatsinks. It appears the pressure from this restores the broken BGA joints, however I have seen evidence after learning more about BGA soldering practices that this may not be the case as failures appear from the outside inwards. The very few professional examinations of this preliminarly show that there are no failing joints outwards in, but perhaps cold solder joints instead. I am not sure as I'm not an expert, I only solder modchips and replace basic components. Now the E74 problem seems to be caused by the ANA/HANA chip failure, although I have seen it get fixed via a basic RROD fix as mentioned above. I have had more "sucess" using a standard paint heatgun, Wagner 700/1000F from OSH. Many claim this melts the solder under the chips causing it to form proper bonds. Common sense dictates that heat has to travel through the chip to melt the solder and simple math shows that 700F is above the levels for many of the components to flat out fail due to warping. Personal experience shows that it works somehow, even if only for a month or three occasionally a year or more. I fixed one 360 today showing heavy artifacting in game by heating up the ram chips with said heatgun. Only time will tell how long it lasts. Can it melt solder? Definitely, I have removed broken chips off 360 mobos with it and have melted solder midair with it so it is theorectically possible if not dangerous. Now, as for BGA reballing I have only seen two real professional attempts at doing it/explaining it in the context of the XBox 360. One man said that as solder goes through heating/cooling cycles it's melting temperature goes up. I had never heard of this but I also cannot find formulas to figure out just how much this new melting point goes up. His point was that the temperature now needed to remove the GPU was so great that it would warp the GPU itself rendering it useless for reballing. His only suggested solutions were to have replacement GPUs on hand to solder on that were not subjected to high temperatures but good luck on getting a supplier for these short of removing them from other 360 motherboards. The other attempt was a Brazilian group that was mentioned on XBox-Scene by another who translated some of their documents. They made video showing the GPU being removed, reballed, and resoldered to the board and claimed it was a complete sucess. I have doubts and so do many others about the video being what they say it is since they clearly don't show every step of the process uncut. In my personal opinion I'd say it would permanently fix the problem as long as you make sure the x-clamps get replaced properly (couple different methods to this these days) and the cooling fans react properly to the rising temperatures experienced in a 360. Frankly I don't think they run fast enough for what they are doing but it would then make them quite noisy. If you really have access to such a station would you mind divulging some details about the process and how difficult it would be for someone such as myself to learn how to do this and acquire such a machine? As well as maybe pictures of the process when you do it sucess or failure? Again this is all my own PERSONAL understanding of the topic. Anyone feel free to correct me on any point. It's great to learn...sorry for the Tolstoy.
About the cost of the station, it is REALLY expensive, they use it to desolder chipsets from notebooks and repair them, a friend of mine who works there told me that the equipment costs about U$S 4000. I talked to him about reballing, he told me that it depends on the condition of the solder joints, if they are only broken it is much more easier to "rework" the solder than rebailing. About giving details, it depends on his boss, if he allows him to have visitors inside the lab where they have the equipment, I'm going to take lots of photos. PS: I made a mistake when writing the title of the topic I wrote "te" instead of "the".
I'm not convinced the x-clamp (will call it 'xc' in the rest of this post) is the real problem. The xc does two important things: it pulls the heatsink down onto the chip its cooling from the top and it pushes from the underside to equalize the pressure. With the replacement screws, you're actually causing warping by them pulling the board up on each side as the heatsink pushes against the processor. If you look closely at the xc studs originally in the heatsink, they're designed to push down on the board if the heatsink is seated too firmly, stopping the board from pulling upwards. The only problem might be the screws that pull the clamp down onto the chassis. Before replacing the xc, I would just remove those eight screws and see if that helps. As for the error messages, 3RROD is almost always the CPU. There was one instance where the machine screw xc replacement on the GPU side caused the heatsink to be pulled so far down onto the board, it would short out the components causing a 3RROD on command. To (unscientifically) check if the 3RROD is a CPU problem, when doing the overheating trick to fix a RROD, keep a cold compress on the GPU heatsink and allow the CPU side to overheat. If the RROD is gone, it's most likely a CPU problem. E74 is tricky error because it appears to be anything on the graphics side, but I think it's most likely the GPU. There are videos of people pushing down on the H(ANA) chipset and the video goes back to normal, but that could simply be the board warping and causing the GPU to make contact with the board again. When I had the xc replacement machine screws on the GPU side, if the screen started to pixelate and turn red / green, I could adjust the screws and bring the output back to normal. If not, when I shut the console down and started it back up, I would have the E74 error. Probably one of the most important things you can do is get some good thermal compound for the best possible heat transfer between process and heatsink. The stuff Microsoft's manufacturer used is horrible. I'd also recommend using the larger all aluminum CPU block heatsink over the smaller copper / aluminum variation. The block heatsink has much more surface area and better heat dissipation through the fins, in my opinion. The heatpipe is nice on the copper one, but impractical when the fins can't efficiently pull the heat out due to the passive cooling layout and small form factor of the unit. A note on the heat gun: with the self-overheating method, the console can rely on its onboard thermometers to safeguard against extreme temperatures. Unless you have access to an infrared thermometer or a controlled source of heat, I wouldn't use the heat gun method as you have no idea when you've exceeded the component's temperature range.
