Word of warning to those considering the Bytecc ME-350U2

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
The unsuspecting customer would be wise to know that there are two models:
  • the older one which is the ME350U2, but also refered to as the ME-350U2 (see the url).
  • and the newer one which is ME-350....which they apparantly distinguish as the 355 (i.e see the url, ... also conveyed in Newegg's item number).
The new version is actually a Hotway HD4-U2, pretty similar to this HD6-U2 model. Of course the actual HD4-U2 is not listed on Hotway's website, but I know that it had been previously. The telltale signs are the different labels (Bytecc vs Smartdrive) and the base.

Buyer beware indeed Charles_A. Like several other people, it was discovered (post purchase) that Windows would not detect the drive on one certain desktop and laptop to which it was supposed to serve. Of course it seemed to work fine on my computer, but that matters little. Incompatibility problems anyone? Yes, indeed, the Genesyslogic GL881E-04 bridge chip seems to suck the big one --- but don't tell these guys, as they seem to think its honky dorry. Alas, with no firmware update available on the net, back she went.

Perdictably, the store wants to do a diagnostic, and just as perdictable as that, it predictably worked on their test machine. Nevertheless, given that it was purchased within the all so gracious 7-days-or-less-and-maybe-we-might-do-something period, and it was accompaigned by the sacred reciept, the tech wrote upon the reciept that we could exchange or return but a 15% restocking fee applied. The lame excuse for the restocking fee was that they couldn't sell it as new any more. Well, I got news for them, I've never seen them sell any opened merchandise before, so they dock the customer 15% and then send it back to the distributer for whatever type of RMA/return/refund. Nice.

Anyways, that just didn't sit too well, and so we complained to the manager (who seemed like a decent fellow, and whom even admitted to knowing about a compatibility issue with the chipset for that particular enclosure). After 3 minutes of badgering, he caved and allowed a swap without restocking. He explained that the problem he faced was pressure from head office -- and I completely believe him, but thats too bad -- Don't pass the buck. If there's a known problem with hardware, stop stocking it, and stop trying to stick it to the customer. Send it back to the manufacturer and say wake up! -- Anyways, the only other problem was that the model we swapped for was $10 cheaper. This took another 2 mins of badgering to ensure that we would receive the $10 refund and not have to purchase an equivalent $10 worth of stock to make up for the difference (personally I was quite willing to do this - to meet the guy half way - but I couldn't think of anything I needed at the moment, and my cousin just doesn't need anything). In the end, alls well that ends well.

Now as for the replacement (Bytecc ME-705U2), it worked straight out of the box without problem (a quick glance in the store [Mgr. opened it for me] indicated it has a Cypress chipset). And dispite being cheaper both cost and construction wise (plastic as opposed to aluminum), I liked this replacement model better for the fact that it has ventilation slats all along its sides -- whereas the tempermental ME-350 is/was a totally enclosed enclosure! I don't care if the thing was made of aluminum, that's just asking for baking a drive.
 

mubs

Storage? I am Storage!
Joined
Nov 22, 2002
Messages
4,908
Location
Somewhere in time.
Thanks for the warning, CityK. Bytecc generally seems to have a good rep., and it is surprising that they screwed up with this one.

The local Fry's Electronics had the ADS Dual Link Drive kit (DLX-185) for $60 - $20 rebate, and I bought it. 5.25 / 3.5", integrated worldwide power supply, ball bearing fan, metal chassis with plactic cover, USB2/FW. AFAIK, it uses Cypress/Oxford. I hope it doesn't turn out to be a lemon.
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
mubs, what kind of thoroughput are you seeing on the USB? By the sounds of it, my cousin is only seeing a horrid ~4-7MB/s. I won't get a chance to look into this until the weekend. However, it doesn't sound particularly promisiing -- I was reading a thread on CDFreaks about external enclosures and others who have the Cypress chipsets also reported dismal performance. Also, apparently some of the Cypress chips are nonprogrammable :( Hopefully this isn't the case in this case ! Er, this external case that is!
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
For shame mubs - leaving me hanging like that.

I got a hold of the enclosure last night. It employs a Cypress CY7C68300A chipset, which is indeed a "fixed function" IC (i.e non-programable...i.e firmware programers need not apply). However, there doesn't appear to be anything wrong with it.

Within XP, HDtach clocked it doing 25MB/s (in a theoretical 60MB/s zone) and my real world "copy and stop watch" (aka me counting out loud) tests concurred spot on with that figure. For perspective, the best user reported USB2.0 HDD rate I believe I've ever seen was 28MB/s.

Getting it to work in DOS was fun (sarcasim). If you enjoy hours of editing config.sys/autoexec files, creating boot images, rebooting and changing BIOS settings), then USB HDDs and DOS will likely find a special place in your heart. It appears the available USB mass storage DOS drivers are highly sensitive to memory managers and BIOS usb emulation settings. If anyone is actually interested in this particular subject, I now have a slew of links which should send you well on your way.

For comparitive purposes, Ghost was able to write to the disk at 6MB/s and was able read (via image integrity check) at ~18MB/s. My understanding is that the write speed is typical of other users experiences.

I thought maybe I could improve upon the writing rate by speeding things up on the internal disk side of the equation (internal disk --> mobo IDE controller --> mobo USB controller --> external bridge chipset --> external IDE disk) via implemention of a UDMA DOS driver, but the driver was failing to initialize and I expended all my time available. I suppose it might, although I'm not certain what mode DOS natively runs HDDs at (likely stated in the Reference Guide, but I'm not checking right now) or if that is a possible bottleneck at all. Might look at it later on.
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
CityK said:
I thought maybe I could improve upon the writing rate by speeding things up .... via implemention of a UDMA DOS driver, but the driver was failing to initialize
I took a look at this again last night. My previous night's difficulty with the udma driver was due purely to a clerical error which I was overlooking. I got the UDMA driver working fine but with just one exception. *
I suppose it might [ed]help[/ed]
Although I wasn't doing an apples to apples comparison (because I was using two completely, and vastly different boot images), the boot image that loaded the UDMA driver showed ~10% increase in Ghost image writing speed (i.e. things were flying at a breakneck pace of ~7MB/s) over the boot image that didn't load the udma driver.

If anyone is interested in said driver, you can find it here.

I'm not certain what mode DOS natively runs HDDs at (likely stated in the Reference Guide, but I'm not checking right now)
I couldn't find mention of DOS's native level of support for disk controllers.


* The one exception was when I used it in conjunction with the USB enclosure (some sort of conflict with the usb driver and the loading of the udma driver into high memory), which would cause some EDD error. Interesting to note is that using the usb driver required only using himem and not emm86. What is so interesting about that was that when I checked tonight for a link to this driver that I could post, I discovered a new version (v2.5) was released today. And, as you can read here, the authour notes that:
There was a math ERROR in all of the prior V2.4 drivers, found by Ladislav Lacina of the Czech Republic in running UDMA2 with no EMM386 driver. I seldom test with HIMEM but no EMM386; and I shall NOT omit such tests in the future! This error has been corrected, and the V2.5 drivers now run fine with HIMEM/EMM386, with only HIMEM, or with neither. My apologies to Ladislav and to all!"
Perhaps the EDD error being generated for me when testing with the usb enclosure/drivers was related. Unfortunately, I can't test that theory as I returned the enclosure back to my cousin.
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
PS - to satisfy my own curiousity about whether there is a performance increase, I'm going to do an apples to apples comparison later on (i.e I'll take a boot image that loads the driver vs the same boot image with the exception that I'll rem out the driver loading line).
 

CityK

Storage Freak Apprentice
Joined
Sep 2, 2002
Messages
1,719
CityK said:
to satisfy my own curiousity about whether there is a performance increase, I'm going to do an apples to apples comparison later on (i.e I'll take a boot image that loads the driver vs the same boot image with the exception that I'll rem out the driver loading line).
The apples to apples comparison revealed absolutely no performance improvements with the udma driver. In fact, a very slight performance hit was observed.

I have to (rightly or wrongly) conclude that the improvements I saw the other night were attributable solely to the vast differences between the two boot images (i.e different sets of drivers and memory loading & usage).
 

mubs

Storage? I am Storage!
Joined
Nov 22, 2002
Messages
4,908
Location
Somewhere in time.
On April 19 said:
For shame mubs - leaving me hanging like that.
Sorry, CityK! Was away from April 9 till a few days ago. I really haven't had a chance to use the enclosure, nor will I for quite a while. All my hardware related stuff has been on hold for a long time. When I do get to test it, I'll let you know.

FWIW, I bought a WD 120GB USB2/FW external last year. IIRC, I had already listed the speeds in some thread, but I'll do it agaion here:

Code:
597,241 KB file copy
How timed: right-click drag-and-drop; from the time "Copy Here" was clicked on the contxt menu to the time the "Copying.." explorer window disappeared

==========================

DESKTOP CONFIG

W2k Pro; 2xP3-900; 112 MHz bus; 4x256MB ram
Promise Ultra-100 disk controller
Adaptec USB2Connect 4000 (AUA-4000B); 4-port USB2 PCI controller
Adaptec FireConnect 4300 (AFW-4300B); 3-port Firewire-400 PCI controller

Seagate Barracuda ST3160023A-RK 160GB on Promise channel-1 (no slave device)
WD 1200JB on Promise channel-2 (no slave device)
External WD 120GB 2MB Cache FW-400 / USB2

----------------------------

Internal copy for frame of reference

Copy from Seagate on Promise channel-1 to WD on Promise channel-2
Source file was between the first 77.5 GB and first 90 GB on the source disk (NTFS)
File was copied to the space immediately after the first 5.1 GB on the destination disk (NTFS)
Destination partition was fully defragged before copy. Copied file did not get fragmented.
26.63 secs; 22,427 KB/sec

----------------------------

Copy to external HDD

Common Info:
Source file was between the first 5.1 GB and 5.6GB of the source disk (NTFS)
Destination file was copied to the beginning of the external disk (FAT-32)
Reboot before each copy; file deleted after each copy was timed.


Copy from Ramdisk to WD External using FW-400
24.21 secs; 24,670 KB/sec; 3.4% faster than Promise *******


Copy from WD on Promise channel-2 to WD External using FW-400
27.54 secs; 21,686 KB/sec; 3.4% slower than Promise


Copy from WD on Promise channel-2 to WD External using USB 2.0; Adaptec drivers (12/2002)
33.19 secs; 17,995 KB/sec; 24.6% slower than Promise; 20.5% slower than FW-400


Copy from WD on Promise channel-2 to WD External using USB 2.0; Microsoft Standard drivers (02/2003)
35.10 secs; 17,015 KB/sec; 31.8% slower than Promise; 27.5% slower than FW-400

==========================

LAPTOP CONFIG

W2k Pro; IBM T23 (26479KU); P!!!-1.13GHz; 133 MHz bus; 2x512MB; 48GB 5400 RPM HDD
Intel 830MP chipset
82801CAM ICH3-M I/O Controller Hub (ATA-100/66 EIDE, PCI bus, Low Pin Count interface, USB)
PCI bus 2.2 32-bit, 33MHz (TI PCI1420 CardBus, Mini PCI, external expansion)
2 Type I, 2 Type II, or 1 Type III / CardBus 32-bit / PCMCIA 2.1
Adaptec FireConnect 3-port Firewire Cardbus AFW-1430
Laptop was running on mains power and a fully charged battery

----------------------------


Copy from Laptop HDD to WD External using FW-400
39.32 secs

Copy from Laptop Ramdisk to WD External using FW-400
38.94
38.81; 15,388 KB/s
 
Top