Got a Samung HD103UI (1TB EcoGreen) on Sunday

Silencing hard drives, optical drives and other storage devices

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Got a Samung HD103UI (1TB EcoGreen) on Sunday

Post by Terje » Wed Jun 18, 2008 9:26 am

My HTPC has proven to be allergic to the WD 1TB GB for some reason (if I connect it, I cannot so any other SATA devices) so I decided to go out and get a 1TB F1 to see if that worked any better.

No luck there (samsung drives has always been hard to find in Tokyo) but to my surprise, I found some shops carrying the new 1TB EcoGreen drive so I got one of those instead (darned cheap drives as well, just about 16000 JPY).

Now, I have not gone through extensive testing with this thing, and mind you that I got an older Pentium M Aopen motherboard here with a slightly sensitive SATA implementation.

What I can say is... wow. I got seagates, WD 5000 KS and AAKS and 1TB GP around on various computers in the house, but the HD103UI quickly had me going wow...

The reason for the wow was that I just connected the thing and had no time to actually test anything. I just dragged out a sata cable and put the drive flat down on the TV table.

This stunt usually gives me some vibration noise in the TV table, but this time, I head no such thing whatsoever.

There was some seconds of clicking at spinup, but otherwise the thing was just silent.

The next thing I noticed after the drive had been lying there for a while in a room that was ambient 29C was that it was very very cool. Not cold, but just that kind of warm which is actually comfortable to hold your hand at.

I have just taken my 2 termometers out and with 25.5C ambient, I am now measuring 32.5C on the upside of the drive and 33.9C on the underside (facing the table). This is after spinning 24x7 for 3 days with some light traffic.

This is on Linux and smartctl gives me a SMART temp of 29C.

I then tried to 6 processes to random writes to the disk at the same time for 15 minutes which caused the temp to rise to 34.5C on the upside and 30C according to SMART.

Please note that this is without any airflow whatsoever. Its lying on the TV table just behind my plasma. Disks normally get reasonably hot in that position if I leave them for a long time.

I had a bit hurry to move some data on it to replace a failing drive, so I have already copied 300GB of data to it, but running a sequential write of an 8GB file reports an average of 79.6 MB/sec (which is not the max of the drive, since its already 1/3 full and the outermost cylinders has the highest throughput) and reading the same 8GB file is done at 94.5MB/sec.

Not state of the art of course, but pretty good for a 5400 rpm drive.

This is all very unscientific testing (I know) but I dont really have much time to test on right now.

Unfortunately, I do not have a GP drive I can use for direct comparison either (the once I have are in use, but I am very tempted to tear up my P180 and take one out for direct comparison... the Raid should in theory allow me to take it out and put it back in rebuild after all ;)) but my memory tells me that the samsung is just a small bit cooler and more silent than the GP and even better, it seems to work flawlessy with the moody SATA controller on my HTPC.

Maybe I got a lucky sample, but I got a feeling we might have a new low noise/temp champ here.

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Re: Got a Samung HD103UI (1TB EcoGreen) on Sunday

Post by Terje » Wed Jun 18, 2008 9:47 am

Terje wrote:but otherwise the thing was just silent.
Just to comment on my own stuff here...

With "just silent" I mean that I could not hear it over ambient noise standing a little bit less than 1 meter away.

Yes, I can obviously hear it if I get close enough.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Wed Jun 18, 2008 10:03 am

Sounds really good. 16000 yen turns out to be US$148 right now, which is a great deal, indeed. The next cheapest 1TB drive I know of is the Samsung F1 at around $175. I did some shopping searches, and in Canada/US, there are just a couple of online shops with them in stock, all well over $200. I guess we'll have to wait. I know Samsung won't get us any samples...

oberbimbo
Posts: 65
Joined: Sun Dec 23, 2007 3:18 am

Post by oberbimbo » Wed Jun 18, 2008 11:42 pm

That's great news. I was debating getting another GreenPower but cheaper and perhaps even more silent makes me wait (ok, in fairness, the 3 platter 1TB GP should come RSN, maybe that one will be lower priced, too).

The F1 is widely available in Europe (bit cheaper than the GP, too) but it vibrates way too much to be of use in silent systems.

aztec
Posts: 443
Joined: Mon Dec 12, 2005 5:01 am
Location: Foster City, CA

Post by aztec » Thu Jun 19, 2008 1:13 am

My TivoHD wants this!

NextdayPC has it on their site for US$175 (not in stock though)

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Sat Jun 21, 2008 9:25 am

Well, I just could not stay away from comparing these drives a bit more (sorry, I got something of a storage fetish. I like hacking on drives, raids, SANs and NASes and filesystems).

Unfortunately, the only place I had to test this relatively easy was on my HTPC, and as already mentioned, the SATA controller there does not like the GP, so I had to connect the GP through a SATA<->USB dongle (no spare PCI sata card around) which means that a fair performance comparison was not possible.

Performance is clearly not why you buy these drives however.

Sorry that I do not have fancy measuring equipment around so everything here will be very subjective.

Idle Noise
=======
Both drives are very silent. However, the EG is the best.

Its not a big difference in the noise level, but what is more important is that I like the noise from the EG better than the GP. The GP has a noise that is higher in frequency vs the EG. Its almost like there is some metallic bi-noise in the "swoossshhh" sound from the platters rotating which the GP has, but is not there on the EG.

In a way, it reminds me of difference of the rumble of some old slow spinning 5 1/4" drives vs the whine of a modern 15k 3.5" drive.

The EG is a clear winner in my ears although the difference in terms of dB's is probably very small. With my fridge making some humming and light rain outside. I start hearing the GP 70-80cms away while I can clearly separate the sound of the EG from the ambient noise in the room about 60 cm away.

In terms of vibration, both drives are good, but the GP is slightly better. I can barely feel the idle vibration of the GP sample I have if I hold it in my hand, I can feel it on the EG.

I pulled out a WD5000KS for good measure and its not even a comparison. Both EG and GP drives are way more silent and produces less vibrations.

In terms of vibration, If the 5000KS is 10 (most vibration) and the GP is 1(best) I would put the EG I got around 3-4.

On my TV table, the 5000GS will put a hearable humming sound in the table itself (the table hums easier and more than my P180 cabinet).
Neither the EG or GP make vibrations strong enough to make the table hum (at a volume that can be heard with my ears anyway) so I suspect the difference will probably not matter for a machine with 1 or 2 drives, but might make a difference if you have many drives.

Seek Noise
=======
I had AAM enabled on the GP and no easy way to disable it so comparison will only be with AAM.

Without AAM the noise of the EG drive is of course (?) much louder than the GP. It takes a bit work to get the EG really noisy though.
I was making a small shell script using a bunch of dd's reading at different offsets on the drives for this, and frankly I needed a few more dd's reading at the same time to get noise out of the EG than I needed on the GP, but once pushed hard enough, the noice from the EG is quite loud without AAM.

I have never heard this during normal operation like streaming a TV recording or copying a few hundred MBs of files, but maybe if it was the OS drive I would have heard it better.

With AAM, the noise difference is more difficult to compare. The two drives actually sound a bit different. The GP seems to have a bit louder clicks (real chunk like sounds) but less of them. Maybe something about how the drives order the reads or maybe its just a result of the USB interface not being able to put as much load on the drive as the SATA interface can.

From what I hear, under light random access, EG is best, but under really heavy load, the GP seems better (could be due to the USB interface)

In terms of vibrations, you can feel the seeks on both drives in the hand, but the EG seems slightly better (again, might be the result of more aggressive reads there due to the SATA interface allowing more commands/sec which again could make things feel different).

Idle Temperature
=========
On idle, with ambient temperature of 25.5C in the area where the drives was:

Eco
33.4C - Top side
36C - Underside at the electronics

GP
32.7 - Top side
46C - Underside at the electronics

This might not be entirely fair to the Eco. It had been on 24x7 for 5 days
The GP only for 30 minutes. It could be that the GP was still warming up.

In any case, the heat generated by the two drives from the area around the platters/arms are very similar.

The GP does however have much hotter electronics. I am pretty sure that even if it gets a bit cooler around the platter, it will give more heat to the surrounding which again means more airflow needed.


Load Temperature
=============
Eco:
36.2C - Top side
39C - Side of drive
42C - Electronics under

GP:
37.2C - Top side.
42C - Side of drive
42-50C - Electronics under depending where I measured

Again, from holding the to drives in the hand, its pretty obvious that the GP is clearly making more heat than the EG.
Both are cool compared to any other drive I have had (the 5000KS is for instance way warmer), but the EG is the winner here.

From all I have seen so far, if I wanted 4 drives in a raid in a busy pc (might produce a lot of vibration and seeks), I might go for the GP but again, I am not so sure. In this case, I would maybe select the 24x7 server optimized GP version as well (which I have never tested).

For all other PC configs, I would take the EG.

I might be a bit less certain that the EG is a clear champ after comparing the drives directly, but I still think its in most cases the best drive for silent PC use you can get at the moment.

I am seriously considering going out to get a second of these drives for my HTPC just to be able to throw out all other drives I use there. It would be a nice cleanup of my HTPC mess.

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Sat Jun 21, 2008 9:56 am

Terje wrote: Load Temperature
=============
Eco:
36.2C - Top side
39C - Side of drive
42C - Electronics under

GP:
37.2C - Top side.
42C - Side of drive
42-50C - Electronics under depending where I measured
I let the stress test run for almost 2 hours after this and that had an interesting effect.

The EG now measured 37C on the top of the drive, the GP had reached 40.6C.

I do suspect a lot of the difference here comes from the fact that the drive is just laying straight on the table. The heat from the electronics has no way to escape except through the metal housing and it gradually heat it up.

Bottom line... the EG probably does quite a bit better in places with really tight mounting and low airflow.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sat Jun 21, 2008 10:51 am

Terje,

Your comparisons are certainly interesting, and I don't doubt that what you hear is really there -- at least for the samples you have.

I question the significance of the temperature differences you saw, however. The total heat of any drive can be summed up very simply in power dissipation. Any differences in temperature readings just have to do with how & where the heat is distributed.

Looking just at the published specs, the Samsung is 5W idle and 6.2W peak. The WD is 4W idle and 7.5W max. Would the 1.3W difference in peak result in big temp differences? Or the 1W difference in idle? I kind of doubt it.

I think what you're probably seeing is different parts of each drive getting hotter, but it doesn't change the basic fact that they are very similar. The biggest difference between them has to be the platter/head count -- the Samsung has 3 platters (they state 334GB max/disc) while the WD has 4 (AFAIK). You'd have to give the nod to the Samsung probably running a bit cooler... but as you noted after your longer run, the difference in temp was just 3.6C. To me, that's not highly significant; probably better if you took 20 temp readings all over each drive, and then averaged them all out.

Luminair
Posts: 223
Joined: Wed Mar 21, 2007 10:45 am

Post by Luminair » Sat Jun 21, 2008 12:47 pm

FYI mike someone (at ncix I think) said that the GP is now made with the larger 320gb platters... might be something to look into.

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Sat Jun 21, 2008 1:19 pm

Yes, 3.6 C does not matter much by itself, but it might matter when you sum it all up. 1-2W per drive in the computer, a few on the CPU and a few more on the GPU and the motherboard and you got 10-20W less to get rid off.

I will happily take any -C I get, but of course, I am probably part of the SPCR Asylum.

What I was however most interested in was actually the effect of the quite hot areas reaching 50C on the electronics on the GP, which might matter more in terms of how you mount it in the right settings (tiny HTPC enclosure for instance).

I have for instance seen drives that has been in silent drive enclosures with broken electronics most likely due to overheating. In such settings the drive get cooled sufficiently through direct contact with internal metal sufaces, but the electronics does not get cooled that way and there is absolutely no airflow.

I actually ended up putting down numbers from 3 locations on the drive just because I ended up checking many spots. This all started due to the difference in heat on the electronics, but the simple comparison I did was just to take both drives in my hands. That quickly reveals any hot areas that should be checked and the WD is hotter all over.

I just like the feeling of the samsung. I have been working a lot with storage devices of all kinds and sizes the last 20 years. From small usb devices to multi million dollar SANs and this is the first drive I have encountered that just gets "comfortably warm" with no hotspots at all. That was a wow! experience for me.

I like it when barriers like that get broken down. I want all storage devices to be cooler than myself even without active cooling. That makes things a lot easier :)

porkchop
Posts: 496
Joined: Thu May 25, 2006 1:19 am
Location: Australia

Post by porkchop » Sat Jun 21, 2008 10:17 pm

this is cool, its nice to have another option to the wd gp drives.

any idea if samsung plan to release any 2 or single platter versions :?:

Melluk
Friend of SPCR
Posts: 68
Joined: Wed Apr 30, 2003 2:02 pm
Location: The Netherlands

Post by Melluk » Sun Jun 22, 2008 5:59 am

Terje, is it possible for you to HD Tune test the EcoGreen?

The numbers for the 334GB/platter WD GreenPower are posted here
viewtopic.php?t=48796

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Sun Jun 22, 2008 9:01 am

If HDTune allow me to do the read test on a disk without creating an ntfs filesystem (that is, it can read from the raw device rather than just from the filesystem) I can probably arrange that.

But I did a very quick test on my linux box. Made a quick shell script that simply read 1GB of data of the raw disk in 1MB blocks with offsets increaseing by 100GB per iteration.

Probably reasonably comparable results with HDTune I think. A bit more work to make access time measurements that can be compared since I don't know how they measure it.

Seems like the disks perform quite similar, which I guess is expected.

This is the output (again, offset increase by 100GB per iteration):
0: 94.8 MB/s
1: 88.4 MB/s
2: 86.8 MB/s
3: 83.2 MB/s
4: 77.2 MB/s
5: 75.4 MB/s
6: 69.2 MB/s
7: 62.3 MB/s
8: 54.1 MB/s
9: 43.5 MB/s

This is the script
#!/bin/bash

a=0

while(($a<10))
do
let s=$a*1024*100
echo -n "$a: "
dd if=/dev/sdb of=/dev/null bs=1024k count=1024 skip=$s 2>&1- | grep copied | awk '{print $8 " " $9}'
let a=$a+1
done

Melluk
Friend of SPCR
Posts: 68
Joined: Wed Apr 30, 2003 2:02 pm
Location: The Netherlands

Post by Melluk » Sun Jun 22, 2008 10:56 am

Thanks! Looking good. :)

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Wed Sep 24, 2008 9:04 am

Toms Hardware has reviews of the new 1TB Samsungs.
http://www.tomshardware.com/reviews/ter ... ,2026.html

Seems like they have also updated the previews WD 1TB review tables with the new data making things easier to compare

http://www.tomshardware.com/reviews/hit ... 017-9.html

I believe that this shows Ecogreen vs. 4 platter 1TB GP, which is probably not entirely fair, but still, shows less power usage and lower temp on the ecogreen vs. the 4 platter GP.

It might be interesting to know that the ecogreen I bought originally has failed. Powercycled the PC and it failed to work afterwards. Seems like something in the initialization of heads fails on start. It spins up, it clicks for a while as head assembly moves, but it never shows up in the bios and the circuitry on it gets pretty warm before it just gives up.

I am waiting for a replacement.

Before it failed, I got over a special offer on this drive for sub 90 USD/unit, so I bought 2 more with plans to raid them.

I hope this was just a bad sample and it is not a quality issue with the eco green.

Aris
Posts: 2299
Joined: Mon Dec 15, 2003 10:29 am
Location: Bellevue, Nebraska
Contact:

Post by Aris » Wed Sep 24, 2008 10:58 am

I do like how samsung just comes out and says its a 5400rpm drive instead of how WD tries to bullsh** their customers and say its a variable 5400-7200rpm drive. I still meet people to this day that dont believe me when i tell them their WD drive is a 5400rpm drive.

tksh
Posts: 9
Joined: Sat Jun 09, 2007 6:15 am

Post by tksh » Sat Sep 27, 2008 8:01 pm

I too, am hoping that Terje just got a bad luck of the draw. I've been eyeing the EG for a while now but the general lack of release outside of the Japan and Australia has given us few reviews.

Seeing a review from Tom's Hardware gives me a bit of hope that we'll get these Samsungs before Christmas (and finally more reviews so I can decide whether to get these or the GP)

sneaker
Posts: 133
Joined: Tue Dec 24, 2002 8:37 pm
Location: Australia

Post by sneaker » Sat Oct 11, 2008 12:53 am

I bought a HD103UI yesteday and have a weird issue with it in my Vantec Nexstar 3 enclosure. You switch on the enclosure, and seemingly nothing happens for about 9 seconds. I say seemingly because when I put my ear against the enclosure, I hear what sounds like a seek noise. But it's not until that's finished that it begins spinning up.

The enclosure previously had a HD501LJ in it; no such problem with that drive.

I'm kinda disappointed about another problem I failed to consider - unlike the WD GPs I've been buying, on this drive AAM is off by default, and not having any way to connect it via SATA (we all use laptops), I can't switch it on. Like Terje said, its seek noise out of the box can be quiet loud.

Those annoyances aside, it runs cool and its idle noise/vibration is on a par with the GP.

edlight1
Posts: 36
Joined: Tue Mar 08, 2005 8:33 am

Post by edlight1 » Thu Oct 16, 2008 4:12 pm

Quite sad -- having to send back my new Eco Green. Samsung's HUTIL says "Ecc Errors" - "this drive has some defects."

Also, my Gigabyte nforce 3 board doesn't recognize it again on a warm reboot/Windows restart. I had the jumper on to defeat SATA 2, as the board doesn't have it.

I did run HD Tune's File Benchmark on it, though. It came with AAM off, and people say it has loud seeks that way, so I ran it that way and then with AAM on Fast. HD Tune couldn't set the AAM like it can on my other Samsung, so I did that with HUTIL (in bootable CD form - it only runs on the 1st CD drive detected if there are 2).

Results:

No AAM

Image

AAM set to Fast

Image

It ran very cold. 19C, as a matter of fact. Cold in here at that time, but my WD SE16 was at 27C.

Darn, that's a shame.

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Fri Oct 17, 2008 8:46 am

sneaker wrote:I bought a HD103UI yesteday and have a weird issue with it in my Vantec Nexstar 3 enclosure. You switch on the enclosure, and seemingly nothing happens for about 9 seconds. I say seemingly because when I put my ear against the enclosure, I hear what sounds like a seek noise. But it's not until that's finished that it begins spinning up.
Hm...
I would think that it spins up before you hear the seek noise.

The drives I have got all makes odd noises on startup, but they do spin up before the odd noises are made.

I suspect this is some sort of calibration its doing, although I have never experienced it on any other drive.

It does take a few seconds extra for that than normal as well.

It does not happen if I suspend the drive and start it again. At least not on Linux.

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Fri Oct 17, 2008 8:50 am

Terje wrote: I hope this was just a bad sample and it is not a quality issue with the eco green.
I total, I have bought 4 of these drives now.

The 2 first I bought, both failed. I have got replacement units and they look good so far.

The first one failing was the one I mentioned above which would not even show up in bios afterwards (although it did spin up), the other did not really fail, but as I filled it with data, I started noticing bad block errors.

A scan showed that a fairly large chunk of the middle area of the disk had bad blocks, so I took it back and got a new.

The two I bought a bit later so far runs without problems. I have put them in a raid1 to make sure I don't get any more surprises, but might convert all four into a raid 5 or 6 in the future.

Might be that very early batches had problems. Hope I don't experience more failures anyway.

Terje

sneaker
Posts: 133
Joined: Tue Dec 24, 2002 8:37 pm
Location: Australia

Post by sneaker » Tue Oct 21, 2008 1:16 am

I think the spin-up anomaly I described earlier is due to some compatibility issue between the HD103UI and the Sunplus chipset in the Nexstar 3. I have three Nexstar 3's with the following PCB types and chipsets:

PA-988 REV1.1 Sunplus SPIF215A
PA-988 REV2.0 Sunplus SPIF225A
PA-989 REV1.3 Oxford Semiconductor OXUF934DSA

I've just finished testing the HD103UI in all three. With both Sunplus's, there's that strange 9-second pause before spinup I described earlier. With the Oxford, the drive begins to spin up immediately upon switching on the enclosure. My HD501LJ and WD10EACS work with the Sunplus's without issue.

edlight1, I can't get HUTIL to run on my laptop at all without error, nor ES-Tool or Hitachi's Feature Tool. It's ridiculous they can't provide Windows (or Linux-based) utilities in this day and age.

Terje
Posts: 86
Joined: Wed Sep 08, 2004 4:50 am

Post by Terje » Wed Oct 22, 2008 12:41 am

sneaker wrote:My HD501LJ and WD10EACS work with the Sunplus's without issue.

edlight1, I can't get HUTIL to run on my laptop at all without error, nor ES-Tool or Hitachi's Feature Tool. It's ridiculous they can't provide Windows (or Linux-based) utilities in this day and age.
He, interesting, my GP does not work with the onboard sata of my Aopen based HTPC, the Eco does. :)

With regards to AAM and linux, I set that on the EcoGreen with hdparm without any problems when on SATA, but I could not manage to set that through the USB-SATA dongle.

aztec
Posts: 443
Joined: Mon Dec 12, 2005 5:01 am
Location: Foster City, CA

Post by aztec » Mon Nov 10, 2008 8:59 pm

my EcoGreen generates a fairly loud writing noise at writes. ;)

Is this helped by turning AAM ON? Or does AAM just help during seeks?

Thanks!

deaphix
Posts: 5
Joined: Thu Dec 11, 2008 8:17 pm
Location: tr

Post by deaphix » Thu Dec 11, 2008 9:08 pm

Quietness is normal for a 5400rpm drive but the problem is its too slow :( for anything even for burning dvds under vista, raid is a solution but not logical

bokko
Posts: 1
Joined: Thu May 13, 2010 6:44 pm
Location: Ottawa

PA-989 REV1.3 Oxford Semiconductor OXUF934DSA

Post by bokko » Thu May 13, 2010 7:10 pm

In case anyone else's Google search leads them here this chip is contained in a Vantec Nexstar 3 enclosure esata/firewire and USB most newer drives will not work properly unless you set a jumper for sata 1 speeds only on drive. I have found this resolves most quirky problems with older SATA chipsets.
On WDC drives it's usually a jumper on pin 5 and 6 double check WDC support site to confirm for your model.
Seagate see picture here
(couldn't post link)
If link is broken try to Google "seagate jumper sata 1" should bring you clickable link to picture of jumper settings on Seagate's site.
As for Low power drive that are a nice trade off between speed and silence look at Seagate LP drives speed and transfer rate beats WD green hands down. My LP sits in Antec MX-1 12" from me while I sit here typing BESR backup is running and wouldn't know it was occurring if the blue light wasn't blinking :-)

whiic
Posts: 575
Joined: Wed Sep 06, 2006 11:48 pm
Location: Finland

Post by whiic » Fri May 14, 2010 12:48 am

Image

Post Reply