WxMan1

The AHA-39160 game rule-book states:

If you connect a SCSI hard disk drive to the Adaptec SCSI Card 39160 that was previously connected to a different SCSI card, you must low-level format the drive before you can use it.

The drives I've been using have been formatted NTFS (Win2003) while attached to an AHA-2940UW host adapter. Does that constitute a "different" SCSI card?

Would that formatting have some bearing on the 39160 detecting - assumed to be U320 - HDD as synch 40, i.e., UW, instead of synch 160?

I bought these drives at eBay auction back in 2007 - advertised as Fujitsu U320 HDD - however they show up as ModusLnk HDDs. At the time the U160 host adapters were going for lots $'s for used SCSI U160 host controller. At the time my only option was a used AHA-2940UW. Recently I obtained an in-the-unopened-anti-static-bag AHA-39160 for $20 (w/shipping). That is a dual-channel 64-bit PCI (32-bit compatible ) SCSI host-adapter. To my horror these alleged Fujitsu U320 HDD's are only recognized as synch 40, i.e. UW (LVD). I've tried accessing them on both channel A & B; it matters neither if either or both are connected to the bus.

I'm using 68-pin LVD ribbon cable to 80-pin SCA w/active termination.

Please tell me - without lying - that formatting the drives with the 39160 SCSISelect tool will allow the 39160 to detect my HDD's at least at U160 (predicated that they are the U320's I always assumed them to be).

I'd hate to have to format the drives. Although I could restore a Ghost image to 39160 host adapter reformatted HDD's - not losing any data - it'd still a pain in the behind. What would grate my goat into curd is realizing these alleged U320 HDD's are actually UW not only in deed but fact..

Tyrsonswood

I don't think that card goes any faster than 160, but not positive of that (I always thought the 160 in the model number designated that.... an AHA29160 was a 160 card and the AHA29320 was a 320 card)

Now... about the 40 sync, I want to know this too as I have several U320 drives that won't run at 320 on a good U320 card. It runs that speed with other drives, but 40 only on these few. Go figure..... (not the same card you have but a 64bit LSI RAID card in a 64bit slot) So it will be interesting to hear some options as to the cause......

You are hooking up only these drives, correct?

__________________

WxMan1

AHA-39160 is limited to theoretical limit of 160 MB/sec x-fer rate. Any U320 HDD should operate according to U160 rules. That is what I expected. That is not what I see w/my alleged U320 HDDs. Hence the excerpt from the 39160 rule-book.

Its interesting that you mention reputed U320 HDD that don't; i.e., they run UW at best...

Tyrsonswood

The thing that got me about your post is that yours are "stuck" at 40 sync, same as mine..... I just set them aside for the time being. They "should" run at least U160 and I don't know why they don't...... (actually with my setup they should run 320 because my card supports it)

I don't know if this thread will produce an answer or not,(hopefully it does) but you are not alone on this.

__________________

WxMan1

Were yours new HDD's (or were they used)? Did you try to format your U320 HDD's w/your U320 host-adapter's SCSISelect? If not, any chance you could try that and see if they get recognized as something other than UW?

In my case they were s'poda be MAN, MAP or MAT Fujitsu HDD's (I don't recall what exactly). Thing is that they show up as ModusLnk brand. The thing is that the guy selling these on eBay got into trouble selling crap.

The first drive I bought, a 15k Fujitsu something or another was labeled Qualitas and ended up with serious issues, i.e., bad sectors. It ended up being unuseable. It wasn't worth it to send it back for the refund.

Per the last post on this thread:

How to find firmware for SCSI Fujitsu drive ?? - Hard-Disks - Storage

Check your SCSI cables and termination. The cable must be one of those twisted-pair U320 compliant cables, and the terminator has to be U320 compliant as well.

Also check the BIOS/setup of the Adaptec card, you may have to change some settings in there to get it to negotiate U320.

BTW - GreatDeals00 is the eBay seller I bought the HDD's from. There were some problems with stuff he sold. Just search for ModusLnk / Fujitsu on the interwebs; you'll find a URL where some guy lays out all the dirty laundry on that.

Does any of that last post at the above link hold any water? Or can I infer that the SCSI cables and SCA cards you're using allow some HDD's to be recognized as U320 (but not those problem-child UW only HDD's)? I'm really unclear if special U160/U320 cables or SCA adapters are necessary for SCSI-III or beyond.

WxMan1

Found these links:

PC hardware - General discussion - SCSI u160 cable vs u320

Essentially, the cable is(wire part) is the same, the major difference is in termination. Many u320 cables include the correct terminator on a T connector.

SCSI bus speeds and drive performance

It would appear the problem is related to either the cable, the SCA terminator (or both). I fail to see any difference in the non-terminated SCA related to SCSI synch speed.

Edit:

And this:

http://groups.google.com/group/comp....a197a582d7c804

Tyrsonswood

They are used drives but I trust the source and I know it's not cable related in my case as I added/swapped them into a chain of working U320's....... Same cables, same terminator. When I get to it I'll try the format, if available. I'm using an LSI card, not Adaptec so BIOS has different options but format "should" be there. At the moment that SCSI system is not "running", I need to set it up again is all. Will let you know what happens.

I have also tried one of these drives on a 29160 (expecting U160 out of it) and it was doing the same thing.(40 SYNC) I don't recall if I tried to format or not...... Low level format would take awhile on a 73gb drive...... (It will take even longer on the 148gb)

The drives I do have working at U320 are all Seagate 36gb..... I've got 5 of them on that LSI card, 2 internal and 3 external. (I got them from the same source as the others)

BTW, they scream in RAID 0.

__________________

WxMan1

Roger that. I appreciate your input.

When I can afford the $20 for an U320 active-terminating resistor on my $7.20 / hr - 20 hr / wk - for the U160 SCSI bus I guess I'll try that. When that don't work I'll try U320 rated SCSI cables (another $20 or so). IF THAT don't work, I'll buy a U160 rated HDD - ANOTHER $20 - and see if if works with all the foregoing.

And when NONE of that works, we'll get an IDE HDD of 150 GB or so - in order to image both of my existing SCSI HDD's - and then format the existing SCSI HDD's so that they MAY run at synch 40.

Oh, well. That I even HAVE a PIII 14000 computer that works is the best that somebody that DEPENDS on foodstamps can hope for.

Thankx for your help.

Tyrsonswood

Not that it proves anything at all but I checked some things yesterday... My 140gig ultra320 only runs at 40...... I knew that already. With active terminators, either the "auto" one in the Sun enclosure, or a Plug style it won't go past SE. The plug, cables and LSI card will do U320 with the seagates........ The Sun enclosure could be getting in the way, I don't know. Still, it's an LVD enclosure so it shouldn't be the issue. Setup in the LSI BIOS says 320, but when scanning drives during boot it finds it to be 40.

The LSI card has no low level format in the bios..... Strike that idea.

I do have a couple of U160 cards that I do know have format routines...... One is a 29160...... Sitting on a shelf or in a box, somewhere???.....

I'll dig one of those out and see what happens. I don't get the "format needed" to run at U320 though...... Gotta be something else. The 5 Seagates I have were plug and play other than address selections.

__________________

WxMan1

It appears that LVD buss termination is the culprit here. I don't know this conclusively (yet), but given the hardware I presently have: that's my best guess.

The LVD SCSI buss includes the entire length of the SCSI cable implemented and needs to be terminated with an active terminating resistor sufficiently fast for the speed being negotiated at both ends of the SCSI buss.

This means the dangling end of cable after the last SCSI device needs to be terminated with active terminating resistors; its insufficient to terminate the buss at the last device on the cable.

If one (such as I) has implemented U320 Twist'n'Flat ribbon cable, then the SCA adapter can NOT provide termination to the SCSI buss. The cable itself must be terminated with appropriate active termination for the desire speed to be negotiated.

What I've found is that with the AHA-2940UW, I can run either U320 device at UW, i.e., synch 40, either terminated or unterminated w/out issue. Obviously I'll run the last device on the SCSI cable implementing the SCA w/active termination just to be cool.

However, the 39160 is a different animal. He wants TERM PWR supplied onto the buss by ONE of the devices on the buss, else he won't play. If neither device on the bus is terminated, he'll play nicely w/out TERM PWR on the bus (albeit only at synch 40).

I've got an Amphenol U320 3,4 position Twist'N'Flat w/active terminator cable on the way (new - out of the box - for $7.99 w/expedited shipping on the way).

I'm a little kid waiting for Christmas....

I want my HDD's to work at least U160 and not have to start entertaining RAID ideas...

Long and short of it: SCA w/active termination will clamp one's SCSI buss synch to SE at UW speed; the U320 cable needs to be properly terminated to the synch speed one desires.

Just as an aside: one MIGHT possibly get U320 synch out of a flat 68 provided it is of the highest quality and not too many loads are thereon, and the cable isn't too long: but one will still have to terminate the end of the cable (not the buss).



I'm not positive of that yet, but given HDTach's report of a flat 40MB/s from outter to inner tracks - for both HDD's - I'm very optimistic the issue does not lie with my HDD's; transfer rates are SCSI host negotiated. Moreover, the 39160 provides SCSI TERM PWR. So the issue that the penultimate device on the SCSI cable required an SCA having TERM PWR jumper was illuminative, that and the lab test results.

I can't recommend DiscounTechnology heartily enough; they'll work with their customers.

Tyrsonswood

Hope that works out for you. I know that the proper terminator and cable/cables will make or break an installation. In my case though that's not the issue as the cables and terminators work just fine with other drives, just not the 73gig and 148gig...... all my Seagate 36gig U320s run fine.

4 of those Seagates running exposed on a desktop sounds wicked, BTW.


Never did get around to trying a low level format but I doubt it would make any difference. SCSI project is kinda just sitting there at the moment.

__________________

WxMan1

I am SO bummed out.





I just got 'nother Amphenol U320 cable (this time with the terminator) today. AHS-39160 hangs at SCSI device detect.



The only conclusion I come to is that the alleged Fujitsu SCSI HDD's I bought on eBay from GreatDeals00 Aug 2007 are SE.



Its not all a total lost cause: I've been learned some things; LVD cable needs to be terminated with a speed rated terminator (not the SCSI buss - and YES there IS a difference).

I can get my hands onto a coupla 20GB IDE HDD's for nuthin'. Think I'm going to fire up that FastTrak Tx2000 (v2.00.0.28) and put my apps on there. Sucks cause the system drive can't be RAID'd



Dunno if running BOINC off the RAID array will make all that much diff.



Did I mention that I'm bummed out?



Its not all that big of a deal ya know?



What business do I have running U160 anyways? I am, after all, running on a GF 2 GTS (64MB 5.5 ns @ 200/333). Yeah, I know its downclocked.

I crank it up when I'm gaming to 226/394

And besides, I've gots sidebanding AND fast write enabled IN THE BIOS (just pointing that out to make it clear I'm not a total sad sack).



But gaming is like NEVER lately. Need to replace my Saitek x36f flight stick before hopping into the cockpit anymore these days. O.k., so I can play WinSPMBT (LZ X-Ray scenario). O.k., sure, but its not the same as tooling around town at Mach speed w/serious ordinance strapped on, or pulling 90 deg. turns at 80 MPH w/800 ponies kickin' my 500kg ass.



See, I gots this Samsung 22" flatpanel monitor that was way too small for Dad. Native resolution for that is 1600x1050 and the ol' LeadTek just don't have the oomph for any of that any more these days (not with any decent frame rates).

I'll pick up a nVidia 6800 for dirt one of these days.



Who knows? Maybe Santa be bringin' me a fine and dashing Fujitsu U320 SCSI HDD one of these days.



I can't believe all this trouble started with that $20 DEAL on that brand new in the unopened anti-static-bag AHS-39160.

WxMan1

A used 15K Fujitsu MAU3036NC 36GB on the way...

Its co$ting me $7.95 + that amount for std shipping.

I went with that instead of the refurb'd 2xMAS3367 @ $30 + $10 shipping.

FWIW: eBay they still has two more of those used 36GB MAU's for $7.95 ea (25% discount deal expires by 1000 EST 8 Nov 2011)

FWIW: shipping for second HDD costs only an additional $3.95.

FWIW: eBayer selling these HDD's has 100% @ 4000+ auctions.

Tyrsonswood

Good price..... I'm loaded with SCSI drives though. I'm also broke, so won't be buying much of anything at the moment.

__________________

WxMan1

MAU HDD came: same ol' same ol'.

The only weak link in the chain now are the SCA cards themselves; NONE of them will do LVD. I can get a pair to work in tandem using TERM PWR jumper of another SCA (it doesn't matter if the SCA is powered to molex or not).

Of the two non-terminated SCA I have: one will hang the bus, the other skips the SCSI ID of the HDD.

Of my recently purchased SCSI U320 LVD cables: one is single-mode LVD crimped terminator only. The other cable needs a terminator. The terminator I purchased for it is of the wrong gender.

So I have a $5 non-terminated LVD/SE SCA card on the way...

With a single LVD compatible SCA card I should be able to get my Fujtitsu MAU U320 LVD working at U160. Then I can check the other two HDD's - alleged Fujitu albeit merely ModusLnk - and ascertain their LVD - or mere SE - compatiblity.

Hopefully both original HDD's will prove U160 compatible; then I only need two LVD compatible SCA cards and a SCSI gender changer...

Matth

You are NOT going to like this...
Fujitsu Refurbished Drives

The drives are refurbished & rebranded in an unnofficial channel, and identification of drive and capabilities seems to be a serious problem.

The failure rate on these refurbs also seems to be astronomical - did they just map out some bad sectors and call the drive good?

__________________
Mary had a little lamb,
Her father shot it dead
Now Mary takes her lamb to school,
Between two crusts of bread

WxMan1

Thanx for the URL Matth! I am aware of that though.

My Folks received the newly ordered SCA card coupla days 'go; will be pickin' it up tomorrow. Its spoda do dual-mode LVD / SE, so as long as the Twist 'n' Flat U320 (w/crimped U320 terminator) is rock-solid, I'm exepcting U160 by this time tomorrow.



The used MAU I recently purchased is awesome! I can't believe how quiet it is; given its 15k rotational speed I seriously expected some whine. Nevertheless, that notwithstanding, it has less whine than the 36GB ModusLnk (10K) that replaced it.

Its a bit noisy on random seeks, but when HDD is de-fragged using UD3, even boot-time disk access is fairly silent. HDD accesses are akin to loud ticks of a wrist-watch. I'm very pleased with this MAU (wish I could afford a MAX).

The expectation - - is that the dual-mode SCA card should allow my AHS-39160 to recognize my used MAU 36GB at U160. If so, then I can use that same SCA to diagnose my existing allegedly Fujistsu 36GB & 74 GB, albeit ModusLnk, synch speeds. IF - - they negotiate U160, then an additional $15 (including shipping) will get ALL my HDD to run U160.

I'll be Happy Camper for a little while!

The most interesting discovery: the AHS-39160 is finicky w/respect to the SCA card, i.e., a SCA w/out TERM PWR jumper will hang the SCSI buss, and a SCA w/active termination will ignore the SCSI ID of the HDD w/out a SCA card on the buss having a TERM PWR jumper (whether a molex power connector is plugged into the SCA or not). Given the aforementioned conditions it'll run SE either terminated or not. However, as far as the AHA-2940UW is concerned: it don't care one iota and will run fine with any one (or combination) of the three SCA cards I've have (bus / cable terminated or not).



Just throwing that out there as fuel for thought (in case somebody has similiar issue down the road)...

'Cause all this stuff is goin' down the memory hole since everybody is jumpin' on the SATA bandwagon.

WxMan1

latest SCA card allows both ModusLnk drives to be recognized as U160 including - obviously - the most recent MAU U320 HDD.

Testing the MAU w/HD Tach reveals 89 MB/s burst, 78 MB/s avg sequential read and 5.8 ms latency. The sequential read graph shows about 82MB/s across the outter 3/4's of the HDD tracks where it drops off to almost 70 MB/s. I thought I'd get better than that: specs for the MAU allege 147MB/s internal x-fer rates.

That notwithstanding: the x-fer rates at U160 are more than TWICE what the HDD's dish out at UW.

CPU utilization is between 8-16%, which seems a bit high. Probably due to my CPU being a mere PIII 1400-S?

I don't remember what those Fujitsu refurbed to ModusLnk were alleged to be, i.e., the 36 GB, either MAN, MAP or MAS, and the 73GB plausibly a MAT, I'm eager to bench those ModusLnks. The specs for MAN (52-84 MB/s); for MAP (61.8-106.9); for MAS (93-118); for MAT (73-132). I won't be able to benchmark them until after reeipt of two more SCA cards on order (hopefully by Monday next week).

WxMan1

http://farm8.staticflickr.com/7148/6...26b590b9_b.jpg

http://farm8.staticflickr.com/7144/6...999e8fee_b.jpg

http://farm8.staticflickr.com/7031/6...462bde31_b.jpg

Tyrsonswood

Other than CPU utilization that looks about right..... Your not going to see 160 speeds in reality.

__________________