"Disk Failed", then "Disk Unplugged" errors

Questions about SNMP, Power, System, Logs, disk, & RAID.
Post Reply
dkarhu
New here
Posts: 3
Joined: Sat Jan 23, 2021 3:20 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by dkarhu »

I am having these disk unplugged issues on a fairly new 12-bay rack mount unit - TS-1232XU-RP. 3 of the drives showed "unplugged" one day. Hopefully warranty will cover.
Metria
Getting the hang of things
Posts: 92
Joined: Thu May 28, 2015 2:38 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by Metria »

So just wondering why, if this is a hardware problem, that when a stand alone drive is placed in the "bad" slots, instead of a raid drive, they so far appear to work fine?
xtreme
Starting out
Posts: 39
Joined: Sun Aug 07, 2011 6:49 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by xtreme »

My QNAP faced the backplane failure after 2½ years of use. Customer support offered to swap the backplane for $360 + shipping costs as long as I ship the whole device back to USA (I'm from Europe). Obviously this wasn't an option for me.

The problem I was facing
xtreme wrote: Fri Jun 19, 2020 4:46 am I have QNAP TVS-463 (bought late 2017) with BP TS-463 V1.00. I use to use it with only Slots 1 & 2 occupied and had no idea of this backplane design flaw.
About year later (12/2018) I occupied the Slot 3. About six months from that (2019) I added the 4th drive and the 3th drive started getting ejected constantly. On Slot 3 and 4 there was a HGST 12TB drive.
Based on what I've learnt from you guys was that the MOSFETs (4957AGM) most probably overheated constantly and then broke.
After testing the drives for several months with external PSU I confirmed that the problem was only power delivery related.

On December 16th of the 2020 I promised to proceed to change the MOSFETs and to build a better cooling solution.
I've practised to solder better and I have now replaced the MOSFETs. I had to find magnifier and a new soldering iron too.

TVS-463 Backplane MOSFET fix - TS-463 BP V1.0 (MOSFETs located on the same side as the SATA connectors)
TS-463_BP_V1_0_unmodified.jpg
TS-463_BP_V1_0_modified.jpg
I found MOSFETs from online:
APEC MOSFET AP4957GM SOP8
from ebay
$9.62 (incl. shipping) 2pcs

About the cooling solution part... I will reveal it soon... stay tuned.

EDIT (March, 15th) removed broken link
You do not have the required permissions to view the files attached to this post.
Last edited by xtreme on Mon Mar 15, 2021 9:45 pm, edited 1 time in total.
Beddhist
Getting the hang of things
Posts: 92
Joined: Fri Dec 29, 2017 5:36 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by Beddhist »

Congratulations and thanks for sharing.

Since I found that just turning up the fan speed made the problem go away I would imagine just gluing a U-shaped alu profile onto the chips may fix the problem.
ctkelvin
New here
Posts: 2
Joined: Mon Jan 25, 2021 4:41 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by ctkelvin »

kommisar wrote: Tue Jan 08, 2019 5:05 am After almost a year in service my TVS-473 NAS suddenly dropped disk 3 and the slot refused to work since when. Sounds familiar? :)

Before continuing I would like to emphasize, following IS NOT approved, confirmed or authorized by QNAP! Use it with caution!

First of all don't panic your data as well as the disk itself are fine and you only need a way to backup it. To do that:
1) Power Off your NAS
2) Take the "bad" disk out
3) Connect it to the NAS backplane with male-to-female SATA data extension cable
4) Use any external power source with SATA power connector to provide power to the "bad" disk. It can be standalone PC PSU but you need to figure out how to turn it on.
5) !!!!!!!!!!!!! IT IS IMPORTANT TO MAKE ALL THE CONNECTIONS WITHOUT POWER TO NAS OR EXTERNAL SATA POWER SOURCE !!!!!!!!!!!!!!!!!!!!
6) Power on the external SATA PSU first and NAS after that
7) Wait for NAS to start, RAID to synchronize and backup all your data to external storage.

Now, when the data is safe, it is the time to contact QNAP support, and RMA your faulty NAS. It will take time and in some case money but you will get fixed NAS and keep it under the warranty.
But it is also possible to try and fix the issue by yourself. The cost - penny and 15 minutes with soldering station. !!!!!!!!!! BUT YOU WILL VOID WARRANTY AND ANY UNAUTHORIZED/UNPROFESSIONAL MODIFICATION CAN DAMAGE YOUR NAS. IT ALSO REQUIRES SOLDERING SKILLS AND KNOWLEDGE HOW TO PROPERLY HANDLE DELICATE ELECTRONIC COMPONENTS !!!!!!!!!!!!!!!!

In my opinion the only option to go is to contact QNAP. QNAP support team is helpful and professional. The issue is known to QNAP and RMA is issued on spot. And you need really good reasons to try and fix the NAS by yourself.

If you brave enough to take the second option anyway, lets start from the issue background. It looks like the devices affected are TVS-x73 manufactured before late fall of 2017. On those devices disks 3-8 SATA power is controlled by a chain of MOSFETs. High level on the Gate of control N-channel MOSFET opens high-power P-channel MOSFET and the corresponded SATA slot gets power. And it looks like the control part of the power circuit is responsible for the issue in subject. For some reason control circuit disables high-power MOSFET for a moment and it starts chain of events leading to the disk to be unplugged. Sudden loss of the power also makes HDD to produce loud "click" mentioned earlier in this thread. Another evidence is the following error in kernel log:

ata6: SError: { RecovComm PHYRdyChg 10B8B Dispar } -- Count:x
ata6: SError: { RecovComm PHYRdyChg CommWake 10B8B Dispar DevExch } -- Count:x

It is possible to modify the control circuit to lock the high-power MOSFET in open state permanently and fix the issue for good. I will not provide exact instructions what to do to avoid unnecessary casualties. But if you have right set of skills, info above is more than enough to figure out offending part and right fix procedure.

To summarize, the issue is HW related. At this stage there is no SW/FW fix. If your "Disk Unplugged", your NAS is on warranty and you see symptoms described above - contact QNAP. If it is not possible just follow the provided clues.
Hi kommisar, I have owned a TVS-873 which encountered the problem exactly same as above description.

Could you send me the fixing procedure to my email ctkelvin at yahoo.com.hk? Thanks.
Beddhist
Getting the hang of things
Posts: 92
Joined: Fri Dec 29, 2017 5:36 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by Beddhist »

The fixing procedure is explained in this topic, somewhere above the post you quoted. Remove your backplane, identify the MOSFET(S) and either bridge them or replace them.
ctkelvin
New here
Posts: 2
Joined: Mon Jan 25, 2021 4:41 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by ctkelvin »

Beddhist wrote: Mon Jan 25, 2021 5:05 pm The fixing procedure is explained in this topic, somewhere above the post you quoted. Remove your backplane, identify the MOSFET(S) and either bridge them or replace them.
Yes, I know either the bridging or replace the chipset.
However, the bridging needs to identify correct pin to do so. Is it just using multimeter to short the SATA power connector on the board with the chipset pin to find out?
Or other ways to identify which pins they are.
Beddhist
Getting the hang of things
Posts: 92
Joined: Fri Dec 29, 2017 5:36 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by Beddhist »

My electronics knowledge is near zero, so I can't help you with that. I was lucky that somebody posted the fix for my model above somewhere.

One thing you could do is check whether the part number of your MOSFET matches that of another one, perhaps in one of the posted pictures. If it does, then there is your answer. If not, I would take one of the diagrams to an electronics tech, perhaps with the datasheet for your chip and they should be able to work it out for you. Might be easier to just replace them, though. If you can buy them...
xtreme
Starting out
Posts: 39
Joined: Sun Aug 07, 2011 6:49 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by xtreme »

TVS-463 Backplane MOSFET Cooler - TS-463 BP V1.0 (MOSFETs located on the same as the SATA connectors)

Here I present my cooling solution. First I had to buy a bunch of things to experiment with.
Assembled4.jpg
Assembled3.jpg
https://imgur.com/a/FeFGlb0

Spare parts can be found online:

Code: Select all

Copper Heat Pipe (2x80mm)
https://www.ebay.com/itm/172734208323
$6.53 (incl. shipping)

Blower Fan 12V 140mAh (50x15mm)
https://www.aliexpress.com/item/4000146133937.html
$2.80 (free shipping)

Copper Plate (1x100x100mm)
https://www.aliexpress.com/item/32807343300.html
$9.55 (incl. shipping)

Aluminum Heat Sink (40x40x11mm)
https://www.aliexpress.com/item/32859563728.html
$5.49 (incl. shipping) 2pcs

Double Side Thermal Conductive Tape for Heatsink (20mm x 5m)
https://www.aliexpress.com/item/32833827745.html
$6.99 (incl. shipping)

Conductive Golden Heat Sink Glue
https://www.aliexpress.com/item/4000315251116.html
$5.97 (incl. shipping)

Noctua NT-H1 3.5g, Pro-Grade Thermal Compound Paste
https://noctua.at/en/nt-h1-3-5g?tab=buy

2mm Copper Heat Sink with Thermal Conductive Adhesive (2x67x18mm)
https://www.aliexpress.com/item/32935308547.html
$8.87 (incl. shipping)

Transparent Acrylic Plastic (PMMA)
Buy it grom your local hardware store or from AliExpress or eBay or similar.
I used ones with 2mm and 4mm thickness.

Plastic Hose for Blower Fan (Diameter = 7mm)
I bought from local hardware store. Needs to be flexible enough.

12VDC ON/OFF Thermostat
https://www.aliexpress.com/item/32883724008.html
$7.08  (incl. shipping)
Parts1.jpg
My solution has two elements: 1st the Passive cooler + 2nd the Active cooler
The passive cooler is based in heat pipe and copper plates which leads the heat to aluminium heat sinks. The copper parts are connected with Noctua Thermal Paste. The aluminium parts are connected with Golden Heat Sink Glue.
The parts are tied with small recycled screws. Any small screws are fine. The heat sink unit is held in place with acrylic plastic. Extra layer is glued with epoxy.
ModifiedParts1.jpg
ModifiedParts3.jpg
ModifiedParts2.jpg
The Active cooler is based on two plastic hoses and a Cooling Fan. The fan is controlled by ON/OFF thermostat which has it's sensor connected under the heat sink. The hoses are aligned to blow air straight to the MOSFETs when they reach 45°C/113°F degrees.
Fan settings: Fan is set to blow at Medium speed on 45°C/113°F and with Maximum speed when the temperature reaches 50°C/122°F.
ModifiedParts4.jpg
ModifiedParts6.jpg
ModifiedParts7.jpg
ModifiedParts5.jpg
Future plan (if summer gets super hot): Add a 2nd Active cooler on top of the heat sink and power it with FAN-port of the QNAP motherboard.

Test results:
After running for 24h and QNAP in standby mode (no media usage) and while the room temperature was 22.5°C/72.32°F:
QNAP System temperature was 40°C/104°F and heated up to 43°C/109.4°F on the 2nd day. ** QNAP Auto Fan adjusted from 45°C/113°F to 46°C/114.8°F on the 2nd day. ** MOSFET heat sink was around 35°C/95°F and went up to 40°C/104°F where it has stayed since even with a moderate use.
You do not have the required permissions to view the files attached to this post.
bruce_miranda
Getting the hang of things
Posts: 94
Joined: Mon Jul 03, 2017 6:53 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by bruce_miranda »

Add me to the list. Even made a new post here, not realising that it was the same issue.

viewtopic.php?f=25&t=159161

Has anyone ever done a repair on the TVS-471 (or even TVS-671 or TVS-871)
NAS : TVS-471
CPU : i7-4790T
RAM : 16GB
HDD : 2 x 10TB + 1 x 6TB + 1 x 4TB
SSD : 2 x 500GB (used for SSD Cache)
infernalaanger
New here
Posts: 2
Joined: Sat Apr 04, 2020 4:09 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by infernalaanger »

Just had the SATA 2 port call it quits on a TS-251+ purchased in 2016. Contacted QNAP $220 for an RMA. This is ridiculous. A new backplane can't cost more than $60 and 15 minutes in labor. $220? Seriously thinking about shopping a new NAS through a competitor. Worst customer service experience I've had in a long time.
lazlogogolak
First post
Posts: 1
Joined: Thu Feb 11, 2021 4:13 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by lazlogogolak »

xtreme wrote: Mon Jan 25, 2021 5:54 am My QNAP faced the backplane failure after 2½ years of use. Customer support offered to swap the backplane for $360 + shipping costs as long as I ship the whole device back to USA (I'm from Europe). Obviously this wasn't an option for me.

The problem I was facing
xtreme wrote: Fri Jun 19, 2020 4:46 am I have QNAP TVS-463 (bought late 2017) with BP TS-463 V1.00. I use to use it with only Slots 1 & 2 occupied and had no idea of this backplane design flaw.
About year later (12/2018) I occupied the Slot 3. About six months from that (2019) I added the 4th drive and the 3th drive started getting ejected constantly. On Slot 3 and 4 there was a HGST 12TB drive.
Based on what I've learnt from you guys was that the MOSFETs (4957AGM) most probably overheated constantly and then broke.
After testing the drives for several months with external PSU I confirmed that the problem was only power delivery related.

On December 16th of the 2020 I promised to proceed to change the MOSFETs and to build a better cooling solution.
I've practised to solder better and I have now replaced the MOSFETs. I had to find magnifier and a new soldering iron too.

TVS-463 Backplane MOSFET fix - TS-463 BP V1.0 (MOSFETs located on the same side as the SATA connectors)

TS-463_BP_V1_0_unmodified.jpg TS-463_BP_V1_0_modified.jpg
I found MOSFETs from online:

Code: Select all

APEC MOSFET AP4957GM SOP8
https://www.ebay.com/itm/APEC-AP4957GM-SOP8-P-CHANNEL-ENHANCEMENT-MODE-POWER/401706169653
$9.62 (incl. shipping) 2pcs
About the cooling solution part... I will reveal it soon... stay tuned.
Thank you for this advice, you saved my ..., I owe you one.

Just an FYI , that in my case a TS-451A was the failing device, but with the very same error with disk 3.
In this device the failing piece was a AP4957AGM and since it would have taken months to get the exact same MOSFET, I started looking for an alternative.
I found Si4925DDY https://www.vishay.com/docs/68969/si4925dd.pdf as an alternative. I replaced the AP4957AGM with Si4925DDY and seems to be working fine at a first glance.
I need some more time to see how it behaves on the longer run, but definitely worked better after an hour of runtime than before.
flyt100
First post
Posts: 1
Joined: Thu Aug 10, 2017 12:39 pm

Re: "Disk Failed", then "Disk Unplugged" errors

Post by flyt100 »

I just had a similar thing with my TS-451+ and tracked it down to the PFET switch for the 5V rail on bay 4 (On Semi NTMD6P02). Can't get my hands on that, so I'll use an irf7324...
xtreme
Starting out
Posts: 39
Joined: Sun Aug 07, 2011 6:49 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by xtreme »

lazlogogolak wrote: Thu Feb 11, 2021 9:50 pm Thank you for this advice, you saved my ..., I owe you one.

Just an FYI , that in my case a TS-451A was the failing device, but with the very same error with disk 3.
In this device the failing piece was a AP4957AGM and since it would have taken months to get the exact same MOSFET, I started looking for an alternative.
I found Si4925DDY https://www.vishay.com/docs/68969/si4925dd.pdf as an alternative. I replaced the AP4957AGM with Si4925DDY and seems to be working fine at a first glance.
I need some more time to see how it behaves on the longer run, but definitely worked better after an hour of runtime than before.
No problem. This is why these forums exist.
Thanks for sharing that alternative MOSFET (Si4925DDY) info. I hope it lasts for longer than the original.
I'm interested to hear from you after longer run so please keep us informed.
xtreme
Starting out
Posts: 39
Joined: Sun Aug 07, 2011 6:49 am

Re: "Disk Failed", then "Disk Unplugged" errors

Post by xtreme »

Marvell heatsinks
I added 2mm thick copper heatsinks to Marvell chips as they seem to get quite a bit warm.
These can be found from AliExpress for example.

Code: Select all

https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20210315060040&SearchText=Copper+Heatsink+Cooler+Heat+sink+2mm)
Marvell_heatsinks.jpg

QNAP has 2x PWM fan connectors
On the other side of the backplane board I added a fan. Fan gets the power straight from the QNAP motherboard. As my fan doesn't support PWM control I'm also using a separate temperature based voltage control for the fan. I'll swap it to a PWM fan when I get one in my hands.
Note: On the right side of the fan you need to leave space so that you can tighten the bottom screw of the backplane holder.
QNAP_fan2.jpg
You do not have the required permissions to view the files attached to this post.
Post Reply

Return to “System & Disk Volume Management”