Question Are certain Ram slot combinations not booting my system/bios by design?

[praetoriann]

Are certain Ram slot combinations not booting my system/bios by design
or does it indicate a defective Ram slot?


Got 4 ram slots.
When i test with a single stick.
slot 1 = black screen
slot 2 = bios menu
slot 3 = black screen
slot 4 = bios menu

When i test with both sticks
slot 1-2 = bios menu "sees both sticks."
slot 1-3 = black screen
slot 2-3 = bios menu "sees both sticks."
slot 3-4 = bios menu "sees both sticks."
slot 2-4 = bios menu "sees both sticks."

Is 1-3 not working by design or are these defective slots?

 

[Lutfij]

Welcome to the forums, newcomer!

We're flying blind here, you should mention the make and model of your motherboard, it's BIOS version at this moment of time and the rams you're working with.

 

[praetoriann]

Welcome to the forums, newcomer!

We're flying blind here, you should mention the make and model of your motherboard, it's BIOS version at this moment of time and the rams you're working with.

Yes,
apologies for the lack of info.

The motherboard is a
Z790 aorus elite ax with ddr5 support.

The ramsticks are
Corsair vengeance 5200.
I checked the model number on the motherboard site, so they should be compatible.

Bios says.
UEFI AMI on the box. Not sure which version.

 

[Misgar]

Are certain Ram slot combinations not booting my system/bios by design

I don't think it's by design. More likely there's a problem with the RAM, DIMM sockets, CPU or CPU socket.

You don't say if you've got two DIMMs or four DIMMs.

Faulty DIMM sockets aren't that common. Check for bent contacts or debris in the socket.

Faulty RAM is more likely. Run MemTest86 on each DIMM (on its own). Check separately with the RAM speed set at stock JEDEC (4800MT/s?) and at XMP (5200MT/s) if used.

You might have one or more bent pins in your CPU socket that aren't making proper contact with the CPU.

There's a chance one of the two Integrated Memory Controllers in the CPU is damaged (associated with DIMM slots 1 and 3).

Sometimes excessive bending of the motherboard after tightening down the CPU cooler screws affects connections between the CPU and the DIMM sockets. Releasing the tension slightly on the cooler screws sometimes restores normal operation.

You might need an LGA1700 contact frame to reduce bending:
https://www.tomshardware.com/reviews/thermalright-lga1700-bcf-contact-frame

 

[praetoriann]

I don't think it's by design. More likely there's a problem with the RAM, DIMM sockets, CPU or CPU socket.

You don't say if you've got two DIMMs or four DIMMs.

Faulty DIMM sockets aren't that common. Check for bent contacts or debris in the socket.

Faulty RAM is more likely. Run MemTest86 on each DIMM (on its own). Check separately with the RAM speed set at stock JEDEC (4800MT/s?) and at XMP (5200MT/s) if used.

You might have one or more bent pins in your CPU socket that aren't making proper contact with the CPU.

There's a chance one of the two Integrated Memory Controllers in the CPU is damaged (associated with DIMM slots 1 and 3).

Sometimes excessive bending of the motherboard after tightening down the CPU cooler screws affects connections between the CPU and the DIMM sockets. Releasing the tension slightly on the cooler screws sometimes restores normal operation.

You might need an LGA1700 contact frame to reduce bending:
https://www.tomshardware.com/reviews/thermalright-lga1700-bcf-contact-frame

Sorry I'm a complete novice on some of the languge.
If I understood right I have 2 Dimms and 4 Dimm sockets.
I will see if I can loosen the screws on the cooler a bit.
I'm surprised because I just got these parts brandnew..
Thanks for the reply.

 

[Misgar]

I'm surprised because I just got these parts brandnew

It doesn't matter how new (or old) the board is. Motherboards bend as soon as you secure the cooler. All boards bend slightly when the cooler screws are tightened down., some more than others.

Most times the bending doesn't make any difference, but LGA1700 boards have a reputation for being more prone to problems. due to the length of the socket. Other CPUs are more "square" but LGA1700 CPUs are "rectangular".

As a consequence, in a few instances, LGA1700 sockets need additional bracing, in the form of a contact frame.

Of course the problem may have absolutely nothing to do with bending. If you cannot fix the problem, being new, you should be able to return the board and get a replacement.

Sorry I'm a complete novice on some of the languge.

Did you assemble the computer yourself or get someone else to do it?

The reason I ask is because it's quite easy to damage the delicate pins in a CPU socket if you remove the CPU and it slips out of your fingers, dropping back into the socket at an angle. This often means you have to buy a new motherboard, because bent contacts are not covered by the warranty. Ham fisted attempts to straighten bent pins using a sharp needle often makes things worse.

Loosening the four screws securing the cooler (or water block) isn't a difficult task. Be careful not to loosen them to the extent that the cooler no longer presses down on the CPU with sufficient force to maintain good thermal contact. One complete turn on each screw might be enough to reduce any bending to the point where the CPU makes good contact again with all the pins in the socket.

 

[praetoriann]

It doesn't matter how new (or old) the board is. Motherboards bend as soon as you secure the cooler. All boards bend slightly when the cooler screws are tightened down., some more than others.

Most times the bending doesn't make any difference, but LGA1700 boards have a reputation for being more prone to problems. due to the length of the socket. Other CPUs are more "square" but LGA1700 CPUs are "rectangular".

As a consequence, in a few instances, LGA1700 sockets need additional bracing, in the form of a contact frame.

Of course the problem may have absolutely nothing to do with bending. If you cannot fix the problem, being new, you should be able to return the board and get a replacement.


Did you assemble the computer yourself or get someone else to do it?

The reason I ask is because it's quite easy to damage the delicate pins in a CPU socket if you remove the CPU and it slips out of your fingers, dropping back into the socket at an angle. This often means you have to buy a new motherboard, because bent contacts are not covered by the warranty. Ham fisted attempts to straighten bent pins using a sharp needle often makes things worse.

Loosening the four screws securing the cooler (or water block) isn't a difficult task. Be careful not to loosen them to the extent that the cooler no longer presses down on the CPU with sufficient force to maintain good thermal contact. One complete turn on each screw might be enough to reduce any bending to the point where the CPU makes good contact again with all the pins in the socket.

I will try to loosen it a little bit. The cooler is quite big.

I choose the parts and assembled it mostly by myself but did ask alot of people mostly friends for advice on what I should pay attention to when it comes to compatibility, socket types, case size etc.
It took me around 10 hours to put it together because I was trying not to rush things and get everything right.
But after watching a couple of tutorials, I think I still didn't fully know what I was doing.
On techforums people also have different opinions, but the comments are often helpful.
I've messaged the manufacturer just in case. I don't need the other two extra dimm suckets, but its a downer to have new hardware with potentially defective components.

 

[Misgar]

But after watching a couple of tutorials, I think I still didn't fully know what I was doing.

I've been assembling PCs for decades and I still keep discovering new things.

On techforums people also have different opinions, but the comments are often helpful.

Even with the computer in front of you, it's often difficult to fix problems. You make an educated guess, change settings or parts and see what happens. If the first change fails, you think again and try something else. Unless you have an infinite parts bin, it can take a long time to resolve.

I spent several days fault finding an old system. After testing the RAM thoroughly with MemTest86 I was convinced the motherboard was faulty so I swapped it. Wrong! It turned out the CPU was bad. Not completely dead, but on its way out. After changing to a much slower CPU from my spares box, the system was fixed.

I don't need the other two extra dimm suckets,

That may be true. Two DIMMs often work faster than four, especially with XMP settings, but with a brand new board, something fundamental like faulty DIMM sockets is a cause for concern.

If the problem persists, or gets worse and affects the remaining "good" sockets, the warranty period may have expired if/when all four DIMM sockets fail.

Something to note is most desktop CPUs come with two Integrated Memory Controllers, connected to memory bus Channel A and Channel B.

Although you could fit two DIMMs in Channel A, leaving the Channel B slots empty, or fit two DIMMs in Channel B, leaving the Channel A slots empty, this is not ideal.

The best option is to utilze both memory channels by fitting a matched pair (2 DIMMs), with one memory module in each Channel. This is known as Dual Channel Mode, because you are using both channels, A and B.

If you fit two DIMMs in Channel A only, or two DIMMs in Channel B only, your CPU will be working in Single Channel Mode, which is roughly 5% slower than Dual Channel Mode.

If you fit four DIMMs, they will also be running in Dual Channel Mode. If you fit only one DIMM, it will be running in Single Channel Mode. If you fit three DIMMs, at least one of them will be running in Single Channel Mode.

From the above, I hope you can see that if you're constrained to running your memory in Single Channel Mode, because some of the slots are faulty, performance will be reduced slightly.

If you check the manual for your motherboard, you'll see the recommended configuration for two DIMMs is to fit one memory module in slot A2 (second slot away from CPU) and the other memory module in slot B2 (fourth slot away from CPU). A2/B2 is the standard configuration in many motherboards and often the most stable/fastest solution.

It's difficult to tell how your Slot 1, Slot 2, Slot 3 and Slot 4 terminology ties up with the manufacturer's Slot A1, Slot B1, Slot A2 and Slot B2 description (working away from the CPU socket towards the edge of the board).

Check the placement of your DIMMs and let us know.

I am concerned that your single DIMM tests indicate a problem with one of your CPUs memory channels, but I can't tell if it's the CPU that's faulty, or a contact problem in the CPU socket, or the pcb tracks up to the DIMM sockets, or the DIMM sockets themselves.

If you haven't tested your RAM sticks individually with MemTest86, they might be suspect too.
https://www.howtogeek.com/260813/how-to-test-your-computers-ram-for-problems/

If loosening the CPU cooler screws doesn't fix things, I recommend further investigation. Now is the time to act whilst you can still return components as potentially faulty. I know it's a pain, especially if it's your only computer, but if you've spent a lot of money, you deserve a fully working machine.

The cooler is quite big.

This is the NH-D15 cooler I use in one of my rigs. As you can see, the front fan overhangs the memory slots on some motherboards.

Bios says.
UEFI AMI on the box. Not sure which version.

You can usually determine the BIOS version when you are inside the BIOS. There are other tools running under Windows which can check the BIOS version. I use Aida64.

If your BIOS is more than one month old, it might be a good idea to check for new versions and update if necessary.

Your Intel CPU might need the new "microcode" fix and you'd be well advised to update immediately, before CPU degradation occurs.
https://www.overclockers.co.uk/blog/intel-cpu-microcode-update-what-you-need-to-know/

Updating the BIOS can be a bit nerve wracking the first time you do it, but it's vital if your Intel CPU suffers from the "overvoltage" bug.

In most instances, you download the BIOS file, copy it to a (preferably blank) USB memory stick (in the root directory), insert the memory stick into the rear panel, dive into the BIOS and find the update tool.

Locate the BIOS file in the list, select it and say Yes to start the update. Sit back for a minute or two and pray you don't get a power cut mid-update (a UPS is handy here). If the screen goes blank, wait patiently. Do not power off your machine mid update.

When the BIOS update is complete, you may have to go back in and make any changes required, e,g. boot drive selection, GPU/iGPU setting, XMP memory overclock.