And it is that many users are processing RMA directly after verifying that their refrigerators do not touch with the IHS, something that is not unheard of, but when practically the coldplate is on the air … Something happens. Who is guilty? Why is this happening now and not on other sockets?
The LGA 1700 socket problem: badly designed?
New data shows a reality that many will not like and it is nothing more than a double problem. Yes, it is double, because the motherboards (some) and the CPUs themselves have it, so the solution to the problem is going to be very complicated and we explain ourselves.
As we well know, both in Intel and AMD the IHS end up being mostly concave because the weld between it and the die makes it not perfect by a few microns, which results in contact problems with totally straight surfaces. Manufacturers have alleviated this by correcting their coldplates and making them convex in the center, providing an almost total solution to the problem.
What happens now so that almost no possible refrigeration touches the IHS properly? Well, here is the novelty, since it has been detected that when installing the CPU in its socket, the motherboard sags due to a very curious problem: the excessive pushing force and tightening of the socket for the CPU.
That a motherboard bends is normal?
We are talking about something tangible, already in millimeters as can be seen and that curiously has always happened, except that until now it was the fault of the user or manufacturer for exerting too much pressure on the heatsink or water block. Curiously, the solution goes through a backplate of greater dimensions and rigidity than usual that prevents the motherboard from bending or curving in the socket area.
As a detail and curiosity, even if a complete solid one is used, the backplate and its hardware must first be completely installed and then the CPU is installed and not the other way around. This almost completely corrects the problem and makes the contact very good and also protects the board, but as always there are several “buts”.
The first is that this happens on motherboards with a Poor quality PCB and with a lower number of layers, where they are usually seen in the entry range and in some mid-range. Secondly, almost no manufacturer provides quality solid and metal backplates, so everything is complicated and finally and although said backplate meets the requirements, it must be the case that can be screwed in before installing the block and not through.
This which seems obvious is not so because there are systems that need to install the block at the same time as it is screwed through the backplate, this being a mere fixer so as not to damage the plate and to provide a supposedly improved rigidity to the assembly.
If you are wondering why this happens in this socket, it is simple: the change in the type of power distribution and the high consumption are a determining factor that requires a greater contact between the CPU and the pins.
We assume Intel has taken this into account and informed motherboard and cooling manufacturers, but it seems that very few have taken this seriously in light of all the problems we’ve seen in the last month. So choosing plate and dissipation has become a key issue to consider.
