Will a scanner with a SCSI II interface work with a fast SCSI interface card seen here http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=2075008922

Someone has asked me and Idon't have a clue!

Cheers guys

Yes, in fact I use the variation on this card (2906) for a scanner and tape drive.

Thanks!

SCSI II == fast SCSI, BTW.

"Fast" is and extension to the established SCSI II spec, giving fast rise times and short cable lengths.. You can have SCSI II devices that are not fast.

The SCSI Trade Association would seem to disagree with you, since they seem to list Fast SCSI as the canonical name of the second generation SCSI protocol.

Well, they're a bunch of modifier obsessed weenies, anyway. I mean, geez, we have:

SCSI
SCSI-2
Wide SCSI
Fast SCSI
Fast/Wide SCSI
Ultra SCSI
Ultra Wide SCSI
SCSI-3
U160 SCSI
U320 SCSI

...each with a subtle variation in cabling requirements. Moreover, some of them come in 3 cabling variants: LVD, HVD, SE.

I don't think that any mere mortal can keep track of it.

What's next? Hyper SCSI Ultra Mega 640: Washes whiter than white! Kills all known germs! Decrypts WW2 German naval transmissions!

Except that we've probably already seen that announced already.

I agree, I would have tried to go SCSI on my new computer if it weren't for the terrible confusing terminology they've got for the stuff.

Ha, Rob, I use that same card I think, and also use it for a tape backup and a scanner, mostly through your help. Thanks for sending me that software!

This is why I get confused! I bought the card 2 years ago for a cd-writer and never looked at scsi since! The guy bought it, so he must know what hes buying (i hope).

Cheers

I would have tried to go SCSI on my new computer

I did -- I got an Adaptec 19160 card and a 9.1Gb IBM SCSI 10Krpm disk. I've not regretted it (touch wood). This was a while ago. If I had to do it again, I'd buy a 36Gb 15Krpm disk.

That said, my scanner's USB (bus powered). It's a Canon N650U.

Really, there are only a few different modifiers.

First, Wide/Narrow doesn't affect anything else. Narrow was the ``norm''; everything past Ultra2 is wide-only. Wide doubles the bandwidth as well as doubling the number of available IDs from 8 to 16 (really 7 to 15, since one is always used by the controller itself).

The other portion of the name determines the bandwidth. It just goes SCSI (5MB/s)/Fast SCSI (10MB/s)/Ultra SCSI (20MB/s)/Ultra 2 (40MB/s)/Ultra 160/Ultra 320/Ultra 640. You'll note that the bandwidth between Ultra 2 and Ultra 160 quadruples. That's because of the transition to no narrow buses; the prior ones are based on the narrow bandwidths. Obviously, they realized that they couldn't keep coming up with more superlatives, so they switched to numbering, then they realized that the numbers might as well have some meaning. They'll probably switch to something else the next time. Honestly, I don't think it's any harder than remembering PIO/DMA/UDMA/ATA100/etc. for IDE buses, not to mention not needing an add-on protocol like ATAPI.

Then there's the cabling. HVD (high-voltage differential) is very uncommon. About five years ago it was simply uncommon, but I've only ever seen it once. That just leaves LVD (low-voltage differential) and single-ended (SE). These have to do with the electrical level of the protocol. Most LVD devices will work on an SE bus, but revert to SE mode. You saw some of both in the Ultra days, but before that it was basically all SE and after it's all LVD.

I've got a few old 9GB differential drives but as you've pointed out, nobody uses it anymore apart from legacy systems.

It gets even more complicated once you factor in the number of different connectors as well! There's 50 pin Centronics, 50 pin high density, 25 pin wierd Apple thing...

- Trevor

I don't think it's any harder than remembering PIO/DMA/UDMA/ATA100/etc. for IDE buses,

Remember the old joke about PCMCIA? That stands for "People Can't Memorize Computer Industry Acronyms"

Funny thing is -- that's about the only one I can remember.

tanstaafl.

To be honest Bitt - and no matter how I say it here, it is going to sound as if I am being rude. You are wrong. There is a distinct difference between the realities of the engineering standards, and what (for example) the marketing departments of various different SCSI vendors decide is a cool name to sell the next performance iteration of a standard. You have unfortunately been taken by the marketing side ("Bitt - use the Force!") rather than the engineering side. Sorry Bitt, in this case I must try and take the high ground and say you're off base here. I can recommend "The Book of SCSI" by Gary Fields which explains the standards and also the progression of technologies, and the historical context of competition with the IDE based world. A lot of the material in the book is also available in the SCSI FAQ. Please don't just whack what I said without going to look it up: the SCSI standards have progressed logically from SCSI 1 to 2 to 3 (standards sets) and in each family standard, there have been improvements in physical transmission technologies, transmission path widths, state machines, device families, command sets, and many other physical and logical aspects. It is one of the best managed and logical engineering progressions I have had to deal with in some 20 years of electronics and software design, and to simply look at it from the point of view of "MarketSpeak" acronyms is pretty insulting to the people who have worked on it's development over the years.

As regards the transmission technology, well HVD was a mistake, since it required extremely expensive driver chipsets manufactured in small quantities by one maunfacturer with no second source (= expensive disks that people wouldn't want to buy). It didn't sell very well (for some reason, along with several other variations of the standards that you have proably not seen in the wild), especially since other manufacturers were offering far cheaper IDE technology drives. LVD was the saviour of SCSI and is the only reason it remained competitive - would you have been happy to cable your SCSI drives with a 6 cm gap between connector node points to guarantee performance? That's what HVD would have meant with standard cables, and you would have had to have paid twice as much per disk for the priviledge of being able to buy SCSI disks with performance to match what was available on an IDE drive at less than half the price?

Well, technically, you're right.

However, I don't believe that I have ever heard anyone in the real world use the term SCSI-II to mean anything other than Fast SCSI. If they'd meant the Wide portion of the SCSI-II spec, then they would have explicitly said Wide, since it was a new concept to SCSI when it came out. (Actually, speeding up the bus was a new concept for SCSI, too, but that's certainly obvious in its implications, whereas Wide was not.)

I actually own a copy of The Book of SCSI, I believe. Thin blue and green book -- kinda looks like a Dixie cup? Looks like no. Oh, well.

I'll admit that I'm weak on the physical transmission specs, as I really have no need to be aware of them, but I don't think that my summary of the ``what the user sees'' aspect is far off base. If so, please tell me where. And I certainly didn't intend to dismiss the technical achievements of the SCSI engineering community. I realize that it's not as if silicon suddenly got faster and they plopped a faster clock into SCSI.

I certainly didn't mean to offend, and, honestly, I'm still not quite sure how I did. It's not as if I was implying that those terms were the end-all and be-all of SCSI. It is sometimes nice to be able to know what can hook up to what without knowing how the electrical current is rising and falling.

However, I will see if I can check out that book somewhere. It sounds like it could be interesting.

Now I feel pretty awful. I really didn't mean to come over as the "SCSI Heavy" - I've had too many arguments over the years about SCSI vs. IDE to not be aware of the levels of fanaticism that exist in both camps. Sorry Bitt, I think I was a bit heavy handed there and it was not neccessary. I hope you can let me off the hook for allowing myself to pump up my blood pressure .

I'll be honest: there are few differences between IDE and SCSI that are worth arguing about. Occasionaly, one camp (or the other) gets a bit ahead on some aspect of performance, but one thing that a lot of people seem to forget is that IDE and SCSI actually came from the same disk technology! IBM is in fact responsible for both, and they are in the end related at the root of a technology tree spanning nearly 20 years. The earliest forms of SCSI, called SASI (Shugart Associates Small disk Interface) and the ST506/ESDI (Enhanced Small Disk Interface) interface were pretty much the same thing. ST506 has finally evolved into a Disk-only high-performance single threaded disk interface standard, whereas SCSI has turned into a lower performance, multi-device, multi-threading oriented standard for all sorts of hardware. It could be argued that USB has come from the experience of SCSI: there are all sorts of technical connections that could be made to what IDE has become.

I am always suspicious of MarketSpeak, as I have had to work with technical marketing people all my life. What I have always (without fail) found puzzling is that not one single one of them has had an engineering background, yet they are selling engineering products.

Speaking of which, does IDE implement tagged queueing yet? That seems to be the one area where it has always lagged behind SCSI, at least as far as disk access goes.

I am always suspicious of MarketSpeak, as I have had to work with technical marketing people all my life. What I have always (without fail) found puzzling is that not one single one of them has had an engineering background, yet they are selling engineering products.


That's easy to answer - I'm convinced that marketing is basically the art of white lies, putting the positive spin on the product, and downplaying any flaws, which is far easier to do when you don't really understand it's true capabilities and deficiencies. I'm sure that someone with an engineering background could do marketing, but they wouldn't last very long unless they had absolutely no soul - the guilt complex would be too depressing, and would eventually creep into their work...
Mind you, once upon a time I temped at the Institute of Marketing - and everyone I met there seemed to be shallow and soul-less anyway

does IDE implement tagged queueing yet?

Just recently:

About IBM Deskstar 180 GXP
Deskstar 180GXP is the world’s fastest (IDE interface) desktop hard disk drive with IBM-exclusive technology called “tag ‘n seek” for its simplicity. Known technically as tagged command queuing, “tag ‘n seek” technology enables the drives to perform nearly 25 percent faster than its predecessor and widen the Deskstar performance gap over competitors by up to 20 percent. Formally announced in September 2002, Deskstar 180GXP is a 7200RPM drive with maximum cache size of 8 MB and highest capacity offering at 180GB.

I can recommend "The Book of SCSI" by Gary Fields

As can I. Working for a company that has roots in SCSI, that book is a must read to move into a storage related position.

Of course SCSI is going to get even more complex with the move to SCSI over fibre, SCSI over IP, and other such advancements.