View topic - PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Main » Forums home » Conventional hard drives

All times are UTC - 5 hours [ DST ]

PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Page 2 of 5

[ 81 posts ]

Go to page Previous 1, 2, 3, 4, 5 Next

Previous topic | Next topic

Author

Message

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 21st, 2019, 2:18

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

michael chiklis wrote:

DR-Kiev wrote:

There is a great chance it is spindel stuck issue

If pcb works good, in case of spindle stuck, should i be able to listen a bip sound with a hear next to the drive?

Could it be that a stuck spindle is the reason that the motor controller failed?

_________________
A backup a day keeps DR away.

Top

labtech

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 21st, 2019, 6:28

Joined: August 18th, 2010, 17:35
Posts: 3314
Location: Massachusetts, USA

We had a 16TB My Book Duo with 2 of those 8TB WD Helium HDDs in. Both no spin.

The techs at the IT company it went to first took the drives out and tested them. They said they were not spinning as well, but not sure if that happened naturally within the external enclosure itself or maybe it was a mess up during testing.
No visual physical signs of any burns on the PCBs.
Some brief preliminary voltage testing checked out fine. Nothing apparent being suspicious.

Good donor drives we ordered spun right up without issues in case a PUIS or anything weird was configured on them.
Replaced PCBs on each drive with ROM swap and got the patient drives working again.
https://www.datarecoveryguru.com/blog/2 ... ard-drives

_________________
Hard Disk Drive, SSD, USB Drive and RAID Data Recovery Specialist in Massachusetts

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 21st, 2019, 16:49

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

DATARECOVERY GURU wrote:

Well, for starters, both of these NASWare ready model hard drives strangely sustained electronic failures. Upon testing each HDD at the time for file recovery, neither drive was spinning when powering on.

With that in mind, parts were ordered to repair the PCB on each hard disk drive, as well as sourcing an entirely new WD casing.

Was a new enclosure really necessary? Did you test the original bridge PCB?

DATARECOVERY GURU wrote:

The next step in the data recovery process was to inspect the SMART attributes on each drive using advanced data recovery software.

There is no need for "advanced" software. Just about any freeware SMART tool will do this.

DATARECOVERY GURU wrote:

Finally, the data recovery process from this external drive went as expected, mounting the Mac volume on a MacBook Pro, access the 7.1TB worth of files and extract them out to a brand new Seagate USB external drive.

This type of file recovery requires advanced data recovery software and hardware knowledge and tools, including a substantial investment of over $600 in parts alone.

So you mounted the customer's data on a Mac OS, and then copied the files to an external drive? Where did the "advanced" software and "advanced" hardware knowledge and "advanced" tools come into the picture?

ISTM that this was potentially a very simple DIY job.

_________________
A backup a day keeps DR away.

Top

labtech

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 22nd, 2019, 6:09

Joined: August 18th, 2010, 17:35
Posts: 3314
Location: Massachusetts, USA

Frank, many times you waste your time criticizing things and facts when it is unnecessary.

ISTM that this was potentially a very simple DIY job.
Maybe to you theoretically speaking, given the more technical background. For the average people, even techs, in reality, not so much, as far as the plan and logistics to recover it and the costs involved.

Was a new enclosure really necessary? Did you test the original bridge PCB?
Maybe not, but for $50, it would make sense to buy a new working one compared to the time disassembling the old failed one (potentially ruining the chance of WD accepting the device for warranty replacement) and trying to find the needle in the haystack. These 8TB Helium drives in new condition are over $500 a piece. A customer is likely expected to be quite frustrated to have a $1000+ invested in a device that cannot be replaced by the manufacturer. Many times, it is better to work smarter, efficiently and with the customer's best interest at heart, rather than trying to be MacGyver for no reason.

"Advanced software" compared to the average person/computer shop/IT folks, who use basic tools "to do stuff", hence unsuccessful nor provide the desired results when trying to recover anything. It is witnessed every day when inquiring about "So what happened with this device? What does it do? What kind of tests did you run?", etc.

Here is an example of advanced hardware/software in ensuring the job is done correctly that the average repair/recovery techs don't have: EEPROM Programmer.

_________________
Hard Disk Drive, SSD, USB Drive and RAID Data Recovery Specialist in Massachusetts

Top

michael chiklis

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 22nd, 2019, 14:09

Joined: December 5th, 2011, 5:38
Posts: 1257
Location: Italy

I will receive donor pcb soon.
At the moment it's located in Milan, i'll let you know if pcb + rom swap will work.
Thank you.

Attachment:

shipping.jpg [ 186.35 KiB | Viewed 3440 times ]

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 22nd, 2019, 18:26

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

labtech wrote:

Frank, many times you waste your time criticizing things and facts when it is unnecessary.

AFAICT, you took a simple job (2 x PCB swaps + 2 x ROM transfers) and beat it up with irrelevant references to advanced data recovery tools. AIUI, no advanced tool is yet able to deal with WD/HGST helium models, at least not at the firmware level. Moreover, it would appear that you departed from the standard procedure of cloning the drives and then decrypting the clones, presumably because no advanced data recovery tool can do this. Am I correct?

And yes, I did reread your blog, and it still conveys the impression that advanced data recovery tools are necessary when they are not.

BTW, does PC3K correctly report the helium level (SMART attribute 22/0x16)?

_________________
A backup a day keeps DR away.

Top

labtech

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 22nd, 2019, 19:02

Joined: August 18th, 2010, 17:35
Posts: 3314
Location: Massachusetts, USA

fzabkar wrote:

labtech wrote:

Frank, many times you waste your time criticizing things and facts when it is unnecessary.

Moreover, it would appear that you departed from the standard procedure of cloning the drives and then decrypting the clones, presumably because no advanced data recovery tool can do this. Am I correct?

You are right, did not clone each drive, because it did not make sense to in this particular case. Taking too long, more resources and more money (which customers simply do not want to pay). Since it is an encrypted RAID0 with healthy drives scenario, it made more sense to clone out the entire decrypted logical drive.

Correct, I am not aware of any tool that can put together correctly an encrypted RAID0 from WD enclosure and decrypt the data by working with individual drives/images. So, advanced thinking is necessary. Otherwise, any Jo Mo could do it. But they cannot.

fzabkar wrote:

BTW, does PC3K correctly report the helium level (SMART attribute 22/0x16)?

Attachments:

Helium level.png [ 43.46 KiB | Viewed 3418 times ]

_________________
Hard Disk Drive, SSD, USB Drive and RAID Data Recovery Specialist in Massachusetts

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 22nd, 2019, 19:21

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

Thanks for the PC3K SMART screenshot. I prefer CrystalDiskInfo because it defaults to hexadecimal raw data. As they stand, some PC3K SMART attributes (eg temperature and spin-up time) are unintelligible until they are converted to hex.

_________________
A backup a day keeps DR away.

Top

michael chiklis

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 5:05

Joined: December 5th, 2011, 5:38
Posts: 1257
Location: Italy

I did backup patient ROM

Attachments:

W25Q80BW - WD80EZZX-11CSGA0 - PCB 0J35276 BA5145B.rar [519.11 KiB]
Downloaded 89 times

Top

michael chiklis

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 10:44

Joined: December 5th, 2011, 5:38
Posts: 1257
Location: Italy

I received donor pcb and swapped the rom.
Now the hard drive spins but it does beep sounds, very likely heads are damaged but it could be something else as helium that is vanishing, some sort of faulty on donor pcb... maybe?
I don't know nothing for sure, this is my first helium drive and all we don't know much how this type of drives behavies.

I want to remind you all that original pcb seem to be fine at the begin (DRD + DSC flags were on and i could misure voltages on motor chip) but motor didn't spin, now with donor pcb it spins but heads are producing beep sounds (i think this could be another kind of pcb issue).
I was thinking to swap motor chip from donor to patient pcb... i don't know if is a good idea.

Attachment:

audio hdd.rar [1.5 MiB]
Downloaded 89 times

Top

labtech

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 10:55

Joined: August 18th, 2010, 17:35
Posts: 3314
Location: Massachusetts, USA

fzabkar wrote:

michael chiklis wrote:

DR-Kiev wrote:

There is a great chance it is spindel stuck issue

If pcb works good, in case of spindle stuck, should i be able to listen a bip sound with a hear next to the drive?

Could it be that a stuck spindle is the reason that the motor controller failed?

Based on the beeping behavior, I am inclined now to think that spindle being stuck somehow may be the issue, as well.

_________________
Hard Disk Drive, SSD, USB Drive and RAID Data Recovery Specialist in Massachusetts

Top

michael chiklis

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 11:52

Joined: December 5th, 2011, 5:38
Posts: 1257
Location: Italy

I don't think spindle is stuck, i can hear clearly platters turning and vibrations.
I think most likely is head damage or something releated to helium deficiency.

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 17:25

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

Is the motor controller hot to touch?

Do you think that the VCM is being driven? You can measure the voltage across the VCM current sense resistor (R22D).

_________________
A backup a day keeps DR away.

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 19:29

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

michael chiklis wrote:

I did backup patient ROM

FWIW, I have extracted the ROM components and tested their checksums. There are big gaps in my understanding, and my terminology is probably incorrect, but I can't see any problem (as is to be expected).

Code:

Parsing ROM CP table ...

 Name       Offset  Size (bytes)   ID         unknown   Checksum
table/hdr           table / hdr    tabl/hdr
----------  ------  -------------  ---------  --------  --------
 STRT            0    5000         0080       00000020          
 GNRC         5000    1000         0081       00000020          
 CODF CODF    6000   B7000  B63CC  0082 0082  0000001E  AA55AA55 OK
.EPOS ....   BD000    1000         0085       00000001          
 FSVG FSVG   BE000    C000   A0F0  0086 0086  00000020  AA55AA55 OK
 HIN1 HIN1   CA000    1000   1000  0099 0099  00000020  AA55AA55 OK
 HIN2 HIN2   CB000    1000   1000  009A 009A  00000020  AA55AA55 OK
 FERP FERP   CC000    2000   12EC  009B 009B  00000020  AA55AA55 OK
 FSVS FSVS   CE000    7000   6460  0087 0087  00000020  AA55AA55 OK
 MINQ MINQ   D5000     200    148  0083 0083  00000020  AA55AA55 OK
 FCHN FCHN   D5200    2900   2900  0084 0084  00000023  AA55AA55 OK
 FAEP FAEP   D7B00     F00    F00  0088 0088  00000021  AA55AA55 OK
 FIMG FIMG   D8A00    1C00   1C00  0089 0089  00000022  AA55AA55 OK
 FLOG FLOG   DA600     300    214  008B 008B  00000020  AA55AA55 OK
 FCOD FCOD   DA900     300    138  008C 008C  00000023  AA55AA55 OK
 FINQ FINQ   DAC00     C00    AB0  008E 008E  00000020  AA55AA55 OK
 FDFT FDFT   DB800     C00    C00  008F 008F  00000020  AA55AA55 OK
 FCHG FCHG   DC400     C00    C00  0090 0090  00000020  AA55AA55 OK
 PRSV PRSV   DD000     600    568  0091 0091  00000020  AA55AA55 OK
 VOPT VOPT   DD600     500    500  0093 0093  00000021  AA55AA55 OK
.GNRC CMDT   DDB00     800    7C8  0094 0094  00000020  AA55AA55 OK
 CHDF CHDF   DE300     800    1A8  0096 0096  00000021  AA55AA55 OK
 EINQ EINQ   DEB00     100     88  0097 0097  00000020  AA55AA55 OK
.FSEC GNRC   DEC00     100    100  0098 0098  00000020  AA55AA55 OK
 FCRD FCRD   DED00     100    100  008A 008A  00000020  AA55AA55 OK
 GNRC GNRC   DEE00     300    300  008D 008D  00000020  AA55AA55 OK
 GNRC GNRC   DF100     100    100  009D 009D  00000001  AA55AA55 OK
 GNRC GNRC   DF200     100    100  009E 009E  00000001  AA55AA55 OK
 GNRC GNRC   DF300     100    100  009F 009F  00000001  AA55AA55 OK
 GNRC GNRC   DF400     C00    C00  0092 0092  00000020  AA55AA55 OK
.EPOD ....   E0000    2000         0095       00000001          
 GNRC GNRC   E2000    E000   E000  00A0 00A0  00000001  AA55AA55 OK
 FLLG FLLG   F0000    1000    DD0  009C 009C  00000022  AA55AA55 OK
 GNRC GNRC   F1000    F000   F000  00A1 00A1  00000001  AA55AA55 OK

Total CP count = 34

Parsing FSVG adaptives ...

              Size                 Checksum
Name  Offset (bytes) Unknown   expected  actual
----  ------  -----  --------  -----------------
SHD0     1E0   300C  00000000  75839576 75839576 OK
SHC0    31EC   3E00  00000000  5727446D 5727446D OK
OVLY    6FEC    800  00000000  8EAA7CAF 8EAA7CAF OK
HARM    77EC    1B0  00000000  428B624F 428B624F OK
LINK    799C    274  00000000  0F69D295 0F69D295 OK
WCSC    7C10   1000  00000000  5ECF4AC2 5ECF4AC2 OK
FSPN    8C10    39C  00000000  8E419E84 8E419E84 OK
AFC     8FAC    240  00000000  4224646E 4224646E OK
WCSD    91EC    6CC  00000000  B07F85C9 B07F85C9 OK
FLTG    98B8    800  00000000  6ECD71D9 6ECD71D9 OK
RIWI    A0B8     38  00000000  459F3CC0 459F3CC0 OK

Total FSVG adaptives = 11

Parsing FSVS adaptives ...

              Size                 Checksum
Name  Offset (bytes) Unknown   expected  actual
----  ------  -----  --------  -----------------
CYL      244     84  00000000  55A4A589 41CD67FD
KT       2C8     6C  00000000  858C1108 24C9994D exp + act = 0xAA55AA55
TPCR     334     AC  00000000  00010100 000F1E5C
MFGF     3E0      8  00000000  AA54D654 0000D401 exp + act = 0xAA55AA55
PES      3E8    8C4  00000000  15AACF6B EA66B095
RWGP     CAC     74  00000000  FC5D6984 FB536FD0
RW32     D20    3F4  00000000  0B30A378 8A528A9C
RWTC    1114    158  00000000  FFFECF2D FFFECF2D OK
SWTG    126C   150C  00000000  0FC064E0 0C556696
WIND    2778     18  00000000  00000001 D7F025B1
RRO2    2790   1C14  00000000  30221A11 32AE7A3E
ACDC    43A4    1C8  00000000  00000000 00000000 OK
MACT    456C    2A8  00000000  401D8A95 1FAC1D96
MAFF    4814    104  00000000  4900E2E6 4900E2E6 OK
RRDC    4918   1340  00000000  00000000 00000000 OK
FLTS    5C58    808  00000000  7BF81C00 7BF81C00 OK

Total FSVS adaptives = 16

Parsing CODF CP ...

  Offset     Size      Checksum
           (bytes)  expected / actual      unknown1  unknown2
--------  --------  --------  --------     --------  --------
     168      1BC4  753E6916  753E6916 OK  00025D88  00000000
    1D2C     13014  007E150E  007E150E OK  00000000  00000000
   14D40     9F9BC  16C28619  16C28619 OK  00280000  00000000
   B46FC       180  AFE13AA0  AFE13AA0 OK  00108B00  00000000
   B487C      1B50  537CBC92  537CBC92 OK  006B4000  00000000

Total CODF segments = 5


Parsing ROM resident CP table in CP 0x81 ...

Total ROM resident CPs in table = 34

96-bit hash                   offset
-------- -------- --------  --------
013028C2 0A3D0D4A 0CD065E6         0
FFFFFFFF FFFFFFFF FFFFFFFF      5000
13D30BAB 2E05DC23 EC8237D3      6000
B1566E63 0CD04043 BD862E20     BD000
C9A700D2 744CEDFA A62ADB07     BE000
8D13C720 6C9C74B3 769C0D86     CA000
186715B5 0939A70D 350FB8E1     CB000
2B53D282 C9840985 269C270A     CC000
FF7DF4C6 1AB6B56E 9DFC5C77     CE000
6B8A99B7 5596836C 924C5264     D5000
59DC5706 656835A6 9B806F23     D5200
F4DE79D8 5C732533 43F642C4     D7B00
A0A0C002 3ED7A263 32DDE7C6     D8A00
7276BFC2 80C599E8 D73A3383     DA600
10E78D4C C17EE205 D2FF131A     DA900
B5CF6FA7 19E2EA03 2A9B2C2B     DAC00
6D7DE0CA 46C0FA16 F17B1DF1     DB800
8FAE122F 06B8F4EB 5EF83DA1     DC400
1C0B77C9 DD949C22 C40EEE5E     DD000
E64CBC05 123C783B 54F796A1     DD600
8F622568 84CFE3CD D2F07DDE     DDB00
789C4B3F 202ECD04 7D279092     DE300
21BCD412 C3F7033B EBC2C18A     DEB00
B3C7C8C8 E6FD6158 716BBDC9     DEC00
EE698C8E 9B2322A8 67811847     DED00
9DBDAE85 0F770665 B69BBC53     DEE00
34798481 E29CBF4C F2B42F9E     DF100
86DF33FD 3696F070 9418D7DC     DF200
6235A9FB E0093344 A66D8EE8     DF300
5DA2A927 D1556E46 A8A6CDC8     DF400
7B47E63C 6DA22892 B3982199     E0000
B1701444 10E212CB 85C2ECE6     E2000
59BE66E7 2E1B1312 302C1C8A     F0000
91136E7E C998661F 7CDBC20A     F1000

_________________
A backup a day keeps DR away.

Top

michael chiklis

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 20:25

Joined: December 5th, 2011, 5:38
Posts: 1257
Location: Italy

Wait a second!!!
I was going to measure voltages on R22D as fzabkar suggested, when i've seen two capacitors that were missing. :twisted:

This is what we get when we buy from untrested sellers on ebay, spildit warned me.
You were right spildit after all!

Now is very late here in italy, but tomorrow i will try to swap those 2 capacitors from patient pcb to donor and i'll let you know if something changes.

Attachment:

missing components on donor pcb.jpg [ 456.27 KiB | Viewed 3308 times ]

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 23rd, 2019, 20:54

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

You might like to examine the motor terminals with an oscilloscope, if you have one.

_________________
A backup a day keeps DR away.

Top

michael chiklis

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 24th, 2019, 11:07

Joined: December 5th, 2011, 5:38
Posts: 1257
Location: Italy

Ok, i've soldered those two capacitors below the motor chip and now seem that drive is spinning properly, listen to audio file

Attachment:

audio hdd after capacitor swap.rar [869.41 KiB]
Downloaded 103 times

I've attached the hard drive to sata port of my 64bit Win7 OS, now it is identified in device management but only after "drivers installation" message, which takes about 1 minute since i power on hdd.

Attachment:

Driver installing.jpg [ 25.81 KiB | Viewed 3257 times ]

I don't know if is normal that it takes so long.

MRT doesn't support this drive, even for surface scan or simple smart checking.
HD Tune is too "stupid" to show the smart properly, crystaldisk is better but isn't able to show thesholds, Victoria is better because is able to show helium level which HD Tune and Crystal can't.

Attachment:

HD Tune smart.rar [2.24 MiB]
Downloaded 85 times

Attachment:

CrystalDisk Smart WD80EZZX-11CSGA0.jpg [ 199.28 KiB | Viewed 3257 times ]

Attachment:

Victoria Smart WD80EZZX-11CSGA0.jpg [ 199.65 KiB | Viewed 3257 times ]

Standing on what Victoria smart is showing, the drive seem to be fine but i haven't found yet a tool which is able to show me sectors on UA.
MRT ---> failed
R-Studio ---> failed
WinHex ---> failed
Victoria ---> failed

I don't know yet if is normal behaving like that when this encrypted type of hdd are attached direcly to sata port, the drive came here with his WD enclosure attached to sata/usb controller which has ASM1151W mcu.
Probably after i'll remount the drive to his usb controller i will be able to access to data, but first i've to find out if controller and power supply are ok. I don't want donor pcb get killed.

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 24th, 2019, 17:15

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

The ATA command for reporting SMART thresholds was rendered obsolete many years ago, but HDD manufacturers have continued to support it. I wonder if PC3K and Victoria now retrieve the threshold data some other way. Perhaps the thresholds are included with the main attribute data.

Could you show us CrystalDiskInfo's hexadecimal text dumps of the SMART data? There should be an option to display a hex dump of the output of the SMART Read Data and SMART Read Thresholds commands.

BTW, I'm relieved that I didn't cause you grief with that dodgy replacement PCB. Nice job.

_________________
A backup a day keeps DR away.

Top

fzabkar

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 24th, 2019, 18:03

Joined: September 8th, 2009, 18:21
Posts: 11621
Location: Australia

michael chiklis wrote:

AIUI, you couldn't clone the drive (without the bridge PCB) even if you wanted to. :o

ISTM that one may be able to exploit these enclosures by using a SATA protocol analyser to examine the SATA traffic for VSCs. This may then lead to a "solution" for this new firmware architecture.

_________________
A backup a day keeps DR away.

Top

labtech

Post subject: Re: PCB 0J35276 BA5145B for WD80EZZX-11CSGA0

Posted: May 24th, 2019, 18:18

Joined: August 18th, 2010, 17:35
Posts: 3314
Location: Massachusetts, USA

Michael, what model WD enclosure is it, please? Link from manufacturer with picture would be best.

_________________
Hard Disk Drive, SSD, USB Drive and RAID Data Recovery Specialist in Massachusetts

Top

Page 2 of 5

[ 81 posts ]

Go to page Previous 1, 2, 3, 4, 5 Next

Main » Forums home » Conventional hard drives

All times are UTC - 5 hours [ DST ]

Who is online

Users browsing this forum: Google [Bot], samick and 15 guests

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum