EEVblog Electronics Community Forum
Electronics => Repair => Topic started by: phlegeton on August 08, 2024, 04:58:18 pm
-
I tried to find info about these caps, but I couldn't find any info. I looked for "47 C5N" Which didn't give any result, and I look for "S7 JQ" and "S7 IQ". I could find that "S7" could be 47uF?
What for packages are these caps? Are these Tantaal-polymeer caps?
Hope that someone can help me
-
My guess is tantalum
-
thanks!. Do you know how I can read what values they are? and do you know what package these caps are?
-
yes tantalums
47 would or could be 47 uf, i do think 47 uf but some web pages say 4.7 uf used with multipliers ???? for voltages measure between the pins ex : 12vdc would be a 16vdc etc ...
the other ... searching
would be from here : http://www.iequalscdvdt.com/Markings_and_Codes.html (http://www.iequalscdvdt.com/Markings_and_Codes.html)
47 uf indeed
S =4.7
7 is the multiplier 10 exp7
see following answer for more
-
Not sure if this is the correct datasheet, but the manufacturer looks like Panasonic:
https://docs.rs-online.com/cd03/0900766b8124ab01.pdf (https://docs.rs-online.com/cd03/0900766b8124ab01.pdf)
https://industrial.panasonic.com/cdbs/www-data/pdf/AAA8000/AAA8000COL106.pdf (https://industrial.panasonic.com/cdbs/www-data/pdf/AAA8000/AAA8000COL106.pdf)
"S7" is 47uF. If that's a "j" rather than an "l", then the voltage is 6.3V.
Conductive Polymer Tantalum Solid Capacitors
Surface Mount Type
TPE series B size
-
See also https://www.jauch.com/downloadfile/58afeb6aded9b_11cb0b44f8a2e66625ae/ntpg.pdf (https://www.jauch.com/downloadfile/58afeb6aded9b_11cb0b44f8a2e66625ae/ntpg.pdf)
The "47 C5N" might be voltage code 'C' (16V). 47uF seems reasonable for the size.
I don't think there's a clear / universally followed body or color rule regarding chip electrolytics; they could be tantalum (MnO2 or polymer), or aluminum polymer. I think the latter are uncommon or absent in these exact styles, but various chip versions do exist. (Aluminum are generally more common in "can" formats, but are not exclusive. Aluminum electrolytic I think are always can.) In any case, they are likely tantalum something or other.
Tim
-
Thanks all for the information! I de-soldered two of the caps from the board, and the both measure 50uF. So that means that they are both 47uF. I tried to find on which voltage rails the caps are attached to, but I only could find that the smaller cap is attached to the 5V rail.
(some background: This board is a backplane of a QNAP REXP-1620U-RP. A 16 bay disk drive expansion unit. Suddenly all the low 8 disk drives failed. So my first idea was, maybe there are some bad caps, which could cause a voltage rail to misbehave (have seen that before). Measuring the voltage on the backplane, is very had, if impossible at all, since a fan is in the way, and I don't want to run the chassis without fan, which could the control/motherboard to overheat. And even without fan, it's hard to get to measure.
-
Thanks all for the information! I de-soldered two of the caps from the board, and the both measure 50uF. So that means that they are both 47uF. I tried to find on which voltage rails the caps are attached to, but I only could find that the smaller cap is attached to the 5V rail.
(some background: This board is a backplane of a QNAP REXP-1620U-RP. A 16 bay disk drive expansion unit. Suddenly all the low 8 disk drives failed. So my first idea was, maybe there are some bad caps, which could cause a voltage rail to misbehave (have seen that before). Measuring the voltage on the backplane, is very had, if impossible at all, since a fan is in the way, and I don't want to run the chassis without fan, which could the control/motherboard to overheat. And even without fan, it's hard to get to measure.
So you are trying to randomly replace tantalum capacitors??? STOP!!!
Tantalum caps fail rarely and when they do, it is usually with a bang and fire. Those are not the usual electrolytic caps in power supplies that dry out or leak and have high ESR over time, causing problems!
You need to find a way how to measure the voltages before you attempt any repair. That is the first step to diagnosing the problem. Only then it makes sense to even start thinking about how to fix it.
Replacing parts at random is useless - esp. given that there are probably few hundreds of various decoupling capacitors - and any of them could be shorting a rail, for example. Or you spend a week replacing all of them, at a large expense - only to discover that it didn't fix anything because the problem was something else, such as broken wire, dodgy contact,blown FET in some supply, etc. And in the process you have likely also lifted some tracks and damaged the board. Moreover, that is only one possible issue - the drives could be failing for other reasons too - maybe one channel of the disk controller has failed?
-
I don't like to randomly replace capacitors. So I try yo find out if they are tantalum caps, or if the are a polymer type. And I would like to measure voltages, but that is almost impossible. I tried to figure out how the power rails are layout, and how it is all connected. But with a multilayer board, where not all traces are visible, and with no schematics, it's almost impossible (And I don't have the time to spend weeks on it to fix it).
I have a years of experience with storage. And I know that drives can fail in the most strange ways, and unexpectedly. But 8 drives, which all fails at once is possible but unlikely. There is one scenario that I can think of, and that one drive has developed some kind of short on it's power rail, and caused other drives to fail. But if that happened, why not all 16 drives ? (and I didn't find that drive yet, and with the chassis now taken apart, it's not possible either to test the drives further for now |O).
To make it more interesting, I let a drive scan for bad blocks (just before I took the chassis apart), and that ended without issues. Of course, it's not the same load if I tried to scan all 8 drives at once. The problem seems to be intermittent, or it's with a certain load.
In this case a pragmatic approach could be to replace the caps, if the cost aren't to high, then at least I (hopefully) eliminated one possible cause. Or I could replace the caps on one port, and see if that changes behaviour. Again, without schematics, and the intermittent behaviour this is impossible (at least for me). But if I can save this from ending as e-waste, I'll try.
-
I have a years of experience with storage. And I know that drives can fail in the most strange ways, and unexpectedly. But 8 drives, which all fails at once is possible but unlikely. There is one scenario that I can think of, and that one drive has developed some kind of short on it's power rail, and caused other drives to fail.
That makes no sense. The failure of one drive cannot be responsible for the failure of others. The only plausible reason for multiple failures is a failure in the common power supply which has then overvolted your HDD array. If this is the problem, then it will be easy to prove. Just examine and test the protection components on each drive. These typically consist of TVS diodes for each of the 5V and 12V rails, and/or smt fuses or e-fuses.
TVS Diode FAQ:
http://www.hddoracle.com/viewtopic.php?f=100&t=86 (http://www.hddoracle.com/viewtopic.php?f=100&t=86)
PCB protection devices:
https://www.hddoracle.com/viewtopic.php?f=100&t=1615 (https://www.hddoracle.com/viewtopic.php?f=100&t=1615)
-
Okay thanks for the informative articles! So that scenario is not possible and can take it of the list :-)
-
If a drive doesn't spin up, I can help you to narrow down the fault. Just upload a photo of the PCB. Otherwise, here is another tutorial:
HDD / SSD power supplies:
https://www.hddoracle.com/viewtopic.php?f=59&t=231 (https://www.hddoracle.com/viewtopic.php?f=59&t=231)
-
I still have to test all the failed drives. But when the drives failed, they were all spinning. But if a drives fails to spin, I now know where to look for. So thx again!
-
If they're spinning but cannot be detected, that's usually an internal fault. Typically it's a bad head or a damaged firmware module, or both. It's not something that would occur simultaneously on multiple drives. It could be that you have a RAID which was running in a degraded state and only became noticeable when the last drive failed.
-
What happened is that suddenly 8 drives in a RAID6 array failed (slots 1,2,3,4,5,6,7,8) While these 8 drives failed, the other 8 drives in the another RAID6 array (but in the same storage pool) just kept working. (of course the storage pool itself was not) The 8 drives failed with a unrecoverable error. But the drives where still spinning, and I could look at the S.M.A.R.T status etc.
*edit* Both the arrays were healthy before the drives failed.
-
It sounds like degraded heads. I suspect that these drives are Seagates.
-
How likely is it that the heads fail at the same time? But the disks in question are: HGST Ultrastar He10 10TB SATA
-
HGST is considered to be the most reliable brand at this time. I can't explain your failures. Very strange.
-
HGST is considered to be the most reliable brand at this time. I can't explain your failures. Very strange.
yep Seagate died on me many times, some FW updates where issued, but doing so made you lost all the content :palm:
unless you had some power spikes, or some controller failures, it's kinda hard to explain ? data corruption ?? virus ?
-
I can't explain it either. It's not something I've seen before that so many drive fails at once. I have witnessed several times, that during a rebuild of an array another drive failing, being caused by the heavy operations on the drive during rebuild (depend in the type of array that is).
A virus seems unlikely. Data corruption, depends on what data corruption. If a drive is faulty due to bad blocks for example, then it's like fzabkar already pointed out: The array will be in degraded mode. Sure.. in a RAID 6 array, when more then 2 drives fail, the whole array fails (However the remaining disks itself should still be ok). In this case, the storage pools are used for storing backups. So if a a file is corrupted, and is written to disk, that won't harm the disk itself (but the question is then: how got the file corrupted in the first place?)
Spike on a voltage rail, or a dip in the voltage rail, that could cause some strange behaviour. But what I now understand from fzabkar, is that this is not so easy, since components prevent at least at disk level. But what happens if a voltage rails has a large dip, or the voltage rail is very noisy. That has then to be on the main voltage rail I guess. That could be the power supplies. In this case, the power supply are redundant. And if there is a something wrong with those, I expect to see a alert, that a power supply failed (or both) in which case, the cause will be very obvious.
-
The preamp on the headstack needs two supply voltages. Typically these were +5V and -5V. The former was derived directly form the 5V input, usually via a low pass LC filter. I don't know if the newer HGST drives down-regulate the +5V supply. If not, they would be exposed to any noise or dips on this rail.
-
That is interesting, is that easy to determine?
-
That is interesting, is that easy to determine?
Yes. Remove the PCB, upload a photo, and I'll show you where to test.
-
See attached photo. Hope it's sharp enough.
-
Your PCB has two stages of protection. The first level has two TVS diodes and two smt fuses. The second level has an IC that incorporates a Vcore regulator plus 5V and 12V e-fuses.
https://www.hddoracle.com/viewtopic.php?p=22339#p22339 (https://www.hddoracle.com/viewtopic.php?p=22339#p22339)
I have marked what I think are the preamp supply test points on the HDA connector. Measure those voltages. I suspect that the negative supply will be -3V. If the positive supply is +5V, then power down and test for continuity with the 5V pins at the SATA power connector, and the +5V e-fuse.
-
Thanks for this. Maybe a stupid question, how can I measure the voltage. When the drive is in chassis, I can't reach it.
-
Attach the PCB to the PSU in a desktop PC?
-
I only have to power up the pcb, and the drive doesn't have to be attached? In that case, I could also use my lab power supply. I have a SATA power cable, I attach picture of it. If I can use that one, that would make things much easier.
-
Yes, all we're doing is confirming whether the 5V supply is being cleaned up before it gets to the preamp. In fact, just do a continuity check. That should be enough. No need to power up the drive.
-
I wasn;t sure if the drive needs to be connected for load. I checked, and yes, there is continuity between the +5V and the 5V e-fuse.
-
I wasn;t sure if the drive needs to be connected for load. I checked, and yes, there is continuity between the +5V and the 5V e-fuse.
I'm wanting to know if there is continuity between the output side of the e-fuse and the HDA connector. That will tell us if the preamp's 5V supply is being protected (and filtered?) by the IC.
-
Yes, there is continuity between the output side of the e-fuse and the HDA connector.
-
At this point I don't know any more than anyone else. Your drive has more than the usual protections, so it's hard to see how a bad PSU could be responsible for all those uncorrectable errors. But I'm no expert. :-??
-
Well, you do know more then anyone I ever met about the electronic part of hard drives. And I'm very thankful for your help and patience. And we can now rule out power supply issues. The question is then.. was this just a fluke? Or the OS which runs on the NAS, had a mind of it's own? Or the SAS controller? When I think about the SAS controller, I start to think about the lanes. That would be very hard to debug, or troubleshoot. I'm not sure if I want to dive into that rabbit hole.
-
When you have a lot of drives in a rack, their mutual vibration can cause tracking problems. However, your PCB has a shock sensor at two corners (the white angled components), so the firmware would have rotational vibration compensation.
https://www.hddoracle.com/viewtopic.php?p=19489#p19489 (https://www.hddoracle.com/viewtopic.php?p=19489#p19489)
I really have no explanation for your problems. That said, if you can retrieve a SMART report, then this means that the drives can still reach the firmware zone (System Area) on the platters. In fact, the SMART logs may have more detailed error information. You could use smartctl (Linux) or GSmartControl (Windows) to dump these logs.
-
I put everything together, and connected it to my NAS. I decided, not to check for bad blocks on the disk, but just to rebuild the arrays. If something is wrong with the disks, it will show during build. Since this is a Qnap NAS, I had to figure out how to clear the error status on all the disk. As it turns out, the error status is written to a file. By removing the lines which marked the disk as "error" I could restart the whole array build process. Rebuilding the arrays (from scratch) went without any problem. The last step was to write data to the raid arrays, which also went without any problems.
I monitored the disk for a couple of weeks, and everything is running just fine. So why 8 disks failed at the same time, not sure.
To all of you: Thank you for the help, and I also learned a lot. Thanks!!