bobbintb Posted May 7, 2014 Share Posted May 7, 2014 So, I have been trying to find the cause of this issue for a while now and I think I may have it. I had an older motherboard I used as a server and it appeared some of the SATA ports went bad. I have about 15 drives of various sizes and models. I did all the usual troubleshooting (yes, even swapping the power supply) and narrowed it down to the board as the cause of the issue. It was out of warranty and time to upgrade anyway. So I got an ASRock Z87 Extremem11/ac. When I got it I thought it was faulty as some of the SATA ports seemed to quit working after a while. The drives were all good. When I would swap two known good drives, one in a known good port and the other in a suspected bad port, the "bad" port consistently wouldn't work. It seemed I had a faulty board and RMAd it. The replacement came and I got the same, ports not working or drives being dropped. This is when I remembered the first motherboard. I thought it was just old and completely forgot the issue it was having but all three motherboards where exhibiting issue with drive being dropped or not having disks recognized by the BIOS. At times it seemed random, at other times consistent. I had ruled out the power supply as the cause because I swapped out power supplies initially and still had the issue. The current PSU I am using is a Corsair HX850. I have used several online PSU calculators and even when using generous numbers my recommended PSU should be 625W and this is 850W so I do not think it is not supplying enough power, correct me if I am wrong. I think the PSU cables might be the issue. I have been using a Norco 4220 case for all of this that has 20 HDD bays. They have backplanes that have 2 rows of 5 molex connectors to power the drives. To make cabling neater, I used the cables from older, non-funtional PSUs and made two cables that each have 5 molex plugs on the end. Both of these custom cables I plugged into one of the modular molex plugs of the HX850, and also the other PSU when I was using it. I do not remember the model of the other PSU. So, could this be the cause of my problem? I have a PSU tester and it says the PSU it good as well. Even though this PSU has enough wattage, could using only one of the modular 4-pin molex connectors to power all my drives be overloading the PSU? Link to comment Share on other sites More sharing options...
JKeifer Posted May 7, 2014 Share Posted May 7, 2014 The HX850 is a single rail power supply so I doubt that plugging all the drives into a single modular cable would be causing the problem unless there are latent issues with the custom Molex cables you made up. Is there any way to bypass those custom cables at least on a temporary basis? Link to comment Share on other sites More sharing options...
bobbintb Posted May 7, 2014 Author Share Posted May 7, 2014 That is what I was going to try today. I suppose the backplane might have issues as well and I will check that after I try the cables. Link to comment Share on other sites More sharing options...
JKeifer Posted May 7, 2014 Share Posted May 7, 2014 Yup that'd be next. Good luck! Link to comment Share on other sites More sharing options...
bobbintb Posted May 8, 2014 Author Share Posted May 8, 2014 Well, it seemed promising at first but after checking this morning a disk got dropped. I'll check the backplane but I am pretty sure that is not it. I tested the backplane with the original motherboard and still got the issue but I suppose it is possible that the original motherboard and backplane both had issues. If that's not it, then I have no idea, as I have tested every bit of hardware. I suppose I will have to look at software. Maybe Windows is responsible. It wouldn't be the first time. Link to comment Share on other sites More sharing options...
bobbintb Posted May 9, 2014 Author Share Posted May 9, 2014 I think the reason this has been so hard to diagnose is that there are multiple issues. Every once in a while I get a disk not showing up in the UEFI. I'm not too worried about that. But I think the bulk of my issue this entire time has been software related. I have ruled out every bit of hardware as the cause. The main issue is these disks disappear after being resumed from sleep and I can now reproduce the results. I didn't think of this before because, as I said earlier, the UEFI at times wouldn't detect it so I assumed it was hardware. Also, this issue happens with more than one instance of Windows, with Windows 8.1 and Windows Server 2012 R2 (same kernel). The solution for now is to not let it sleep. This isn't ideal but hopefully I can find a way to get it all working now that I (finally) know the cause. Thanks for the help everyone. As a side note, even though I was certain 850W was enough for that setup, I rigged a 500W to power the motherboard and GPU and the 850W to run the hard drives. I used a PSU tester to supply the 850W with a load since it was not connected to a motherboard. But as I was in the process of setting this up, it occurred to me that this test might not be valid. At that time my main theory was sleep mode was messing it up but I wanted to be sure it wasn't the power supply not having enough watts. I am not sure of the internal workings of sleep mode but I thought running the HDDs and motherboard on separate PSUs might mess things up and the HDDs might not actually shut down. That appears to be the case. Does that seem correct? Link to comment Share on other sites More sharing options...
Corsair Employees jonnyguru Posted May 9, 2014 Corsair Employees Share Posted May 9, 2014 It doesn't matter if the PSU is 500W or 1500W. The PSU is only going to put out as much as needed of it and if you're demanding too much power from the power supply, the power supply will shut down. You're running a backplane and custom cables. Sounds like a lot of resistance. And what does resistance do to voltage? Causes it to drop. And if voltage drops... the drives can drop. You can figure out this problem with a DMM. Measure the resistance between the first and last Molex connector of one of the cables that come with the PSU. Now measure the resistance between each end of your customer cable and between where you plug the backplane in and where the drive plugs in. If the resistance is 3x greater, you've got a problem. Another thing you can do with the DMM is to measure the voltages at the PSU, at the end of your custom cable and then where the drive plugs into the backplane. If you see anything close to a 5% drop in either the +5V and +12Vbetween the PSU and where the drive plugs in, then you've found your problem. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.