First PC build help

How does returning a MB usually work? Would I be returning to seller or the manufacturer?
Seller - are you within 28 days(?) - if so it will be a very simple process.

But, abuse that A2 socket first - i've had boards before like this that turned out to be fine (long term fine - not just get working for now fine)
 
Yes well within 28 days of purchase.

What do you mean by abuse exactly? :D I checked it's clear and not damaged and put the stick in a number of times now.
 
Yes well within 28 days of purchase.
It should be a straight swap out then.
What do you mean by abuse exactly? :D I checked it's clear and not damaged and put the stick in a number of times now.
Just reinserting a few times - exactly what you've been doing. You could try with the other stick too - reinserting - as not all things are made equal.

But, if it hasn't automagically sprung to life now it's not looking great for that slot...
 
Thanks, all around 70% now and still no errors.
It will be good if we can determine that both sticks are clear in slot B2 as we can then concentrate isolating A2 as the culprit. This will be when you test A1,1B1 - because if they pass then we can (almost) rule out the cooler as an indirect cause of slot failure.

Although, belt and braces would perhaps require a reseating. I can hear you shout for joy from here.
 
No errors returned on 100% with a single stick in slot B2. It's actually going past 100% but I assume this is sufficient? If so I'll swap it over and run the same test.

A2 is actually the slot furthest from the concrete slab of a cooler!
 
A2 is actually the slot furthest from the concrete slab of a cooler!
I know - but if the board has slight warping it will find the path of least resistance - which maybe around that slot (probably not, but we need to rule it out.).

You wont need to reapply thermal paste for a reseat, or completely remove the cooler - it will be a very simple process.
 
no.
as a rule of thumb, i'd run it to 400%.
reason is, if each run is only 90% effective at finding an error, then 4 error-free runs has a probability of 1:10,000 of missing an error.
Yes, missed that.

Be glad you're not running memtest86...

...or not depending on preference.
 
leave em hanging... lol
The master of the tantric build log. :D

But even Sting must eventually *cough*satisfy*cough* otherwise Trudie would have left him, with half his fortune, years ago...

...unless they're still going?!

@Colin_M - are you practising tantric memory testing - or is that stick still teasing? (ok, giving up on this laboured metaphor now - if the words stick and inserting start to crop up it could get awkward.)
 
Stick one ran past 400% on all 8 instances without error.

Swapped sticks over an one instance has given errors immediately so looking like the ram now?

What should I do next? Same test using A1/B1?
 
Yeah also crashed on the blue screen with memory errors again. Swapped sticks back for piece of mind and running memtest again with no errors does appear to be an issue with that one stick.

I will get onto them and return the set for a replacement.
 
@Colin_M
try the known good stick of ram in A2. does it work and pass 400% memtest?
at least it'll help us definitively pinpoint that it's a faulty stick of ram, and not 2 concurrent issues going on.
 
Back
Top Bottom