jdp@caleb.UUCP (Jim Pritchett) (01/02/91)
[]
What is the maximum cable length for the SCSI bus? What is the recommended
length?
Thanks,
--
Jim Pritchett
UUCP: texsun.central.sun.com!letni!rwsys!caleb!jdp
or spudge.lonestar.org!caleb!jdp
or letni.lonestar.org!dms3b1!caleb!jdppete@fidata.fi (Petri Helenius) (01/04/91)
In article <2167.AA2167@caleb> jdp@caleb.UUCP (Jim Pritchett) writes:
What is the maximum cable length for the SCSI bus? What is the recommended
length?
Thanks,
We've been using little over 3meter cable and it works fine. I've heard
of up to six meter cables being used and I thing the official spec for
SCSI-1 is 3 meters or 10 feet...
--
--------------------------------------------------------------
Petri Helenius, Fimeko-Data Oy Phone +358-0-458 2421
----------------------------------! Telefax +358-0-458 2425
Looking for Unix(r) FAX-system ? ! Internet pete@fidata.fi
Mail queries to unifax@fidata.fi ! FidoNet 2:504/23.0
--------------------------------------------------------------ts@cup.portal.com (Tim W Smith) (01/04/91)
By being very careful to get good cables, and good termination, and other magic things that I don't understand, you can get something like *100* feet. The person I have heard about who was doing this, however, is one of the major experts on cabling systems and noise and stuff like that, so this is probably one of those things that is in the "kids, don't try this at home" category. I don't have a reference to this, since I got it from my boss, who was a speaker at the recent SCSI seminar in New England, and he got it from one of the other speakers, who was the cable guru in question. I'm also not sure that it was 100 feet. It might have been 100 *meters*! Tim Smith
ritchie@hpdmd48.boi.hp.com (David Ritchie) (01/04/91)
> >We've been using little over 3meter cable and it works fine. I've heard >of up to six meter cables being used and I thing the official spec for >SCSI-1 is 3 meters or 10 feet... >-- >-------------------------------------------------------------- > Petri Helenius, Fimeko-Data Oy Phone +358-0-458 2421 6 meters in length, with no longer than 0.1 meter stubs. -- Dave Ritchie ritchie@hpdmd48.boi.hp.com
thomas@neon.in-berlin.de (Thomas Ziemer) (01/04/91)
jdp@caleb.UUCP (Jim Pritchett) writes: > What is the maximum cable length for the SCSI bus? What is the recommended > length? About 6 meters. Can that be true? Bye ... thomas -- __ _______________________________________________________________________ (00) N E O N Research Caboose, Berlin Computer Science Research \`\/ Thomas Ziemer, 1000 Berlin 20, (West) Germany, Phone#: +49-30-361 89 95 "" Smart Addresses: thomas@neon.in-berlin.de, thomas@neon.mbx.sub.org
src@scuzzy.in-berlin.de (Heiko Blume) (01/05/91)
ts@cup.portal.com (Tim W Smith) writes: >By being very careful to get good cables, and good termination, >and other magic things that I don't understand, you can get >something like *100* feet. in fact the maximum length depends on the kind of SCSI bus you use. "Single Ended" SCSI allows up to six meters, "Differential" SCSI allows up to 25 meters. you can use Single Ended to Differential adapters to achieve greater distance between your host adapter and the devices, if these are Single Ended. i.e. Host <-SingleEnded-> Adapter <-Differential-> Adapter <-Single-Ended-> Device(s) tho, i'm not sure about the max length with this, if the adapters don't have their own power supply. the optimum cables are: - 132 ohms imp. 50 cond. flat or 25 signal twisted pair cable for Single Ended - 122 ohms imp. 50 cond. or 25 signal twisted pair cable for Differential both should be shielded. the mileage with your Devices' SCSI implementation may vary, of course. -- Heiko Blume <-+-> src@scuzzy.in-berlin.de <-+-> (+49 30) 691 88 93 public source archive [HST V.42bis]: scuzzy Any ACU,f 38400 6919520 gin:--gin: nuucp sword: nuucp uucp scuzzy!/src/README /your/home
urlichs@smurf.sub.org (Matthias Urlichs) (01/05/91)
In comp.periphs.scsi, article <1991Jan04.191812.16844@scuzzy.in-berlin.de>,
src@scuzzy.in-berlin.de (Heiko Blume) writes:
<
< Host <-SingleEnded-> Adapter <-Differential-> Adapter <-Single-Ended-> Device(s)
I am running such a setup right now (the hard disks were too loud and got
moved to the kitchen ;-) : the single-ended cables are each ~ 1.5
meters, the differential is 20 meters long. No problems.
NB: The adapters should not have to be told which IDs are on what side of the
bus. They can find out for themselves by watching the signals at runtime.
<
< tho, i'm not sure about the max length with this, if the adapters
< don't have their own power supply.
<
If not, the adapter should have a standard power connector, just like the
SCSI drives you're connecting to it.
Powering anything bigger than a terminator pack off the SCSI bus probably is
not a good idea.
< the optimum cables are:
< - 132 ohms imp. 50 cond. flat or 25 signal twisted pair cable for Single Ended
< - 122 ohms imp. 50 cond. or 25 signal twisted pair cable for Differential
The ohms may be important depending on how sensitive/fast your devices are.
If there is a mismatch there can be echoes from the terminators, which can
disrupt bus communications.
< both should be shielded.
<
On the other hand, if radio&TV antenna and any heavy-duty applicances are
sufficiently far away, you may be able to forget about the shielding.
Advantages: it's cheaper if you already have flat cables lying around, and
routing ribbon cables around the house may be easier than a thumb-thick
25-twister cable.
Also, you can press connectors onto flat cable quick&easy (twisted cable has
nontwisted sections every half meter or so for that purpose. Don't ever use
nontwisted cable for the differential bus); shielded cables have to be
soldered and checked, which is difficult with a long cable which has been
routed through walls and whose ends you can't place next to each other any
more...
--
Matthias Urlichs -- urlichs@smurf.sub.org -- urlichs@smurf.ira.uka.de /(o\
Humboldtstrasse 7 - 7500 Karlsruhe 1 - FRG -- +49+721+621127(0700-2330) \o)/src@scuzzy.in-berlin.de (Heiko Blume) (01/06/91)
urlichs@smurf.sub.org (Matthias Urlichs) writes: >In comp.periphs.scsi, article <1991Jan04.191812.16844@scuzzy.in-berlin.de>, >< tho, i'm not sure about the max length with this, if the adapters >< don't have their own power supply. >< >If not, the adapter should have a standard power connector, just like the >SCSI drives you're connecting to it. >Powering anything bigger than a terminator pack off the SCSI bus probably is >not a good idea. that's what i meant by 'own power supply', that they do *not* get it from the cable.. >< the optimum cables are: >< - 132 ohms imp. 50 cond. flat or 25 signal twisted pair cable for Single Ended >< - 122 ohms imp. 50 cond. or 25 signal twisted pair cable for Differential >The ohms may be important depending on how sensitive/fast your devices are. >If there is a mismatch there can be echoes from the terminators, which can >disrupt bus communications. yep, tho the above numbers are what the standard (17b) recommends. however since it might be difficult to obtain such cables, it says lower impedance might work too. if i were to use a 20m cable i'd really try hard to get the optimum material, tho :-) what kind of adapters do you use anyway?? -- Heiko Blume <-+-> src@scuzzy.in-berlin.de <-+-> (+49 30) 691 88 93 public source archive [HST V.42bis]: scuzzy Any ACU,f 38400 6919520 gin:--gin: nuucp sword: nuucp uucp scuzzy!/src/README /your/home
wcarroll@encore.com (Mr. New Dad) (01/08/91)
From urlichs@smurf.sub.org (Matthias Urlichs): > src@scuzzy.in-berlin.de (Heiko Blume) writes: > < > < Host<-SingleEnded->Adapter<-Differential->Adapter<-Single-Ended->Device(s) > > NB: The adapters should not have to be told which IDs are on what side of the > bus. They can find out for themselves by watching the signals at runtime. The adapters shouldn't need to know which IDs are on what side of the bus. They don't need to "know" anything, since they don't need onboard intelligence. They just convert the physical interface. > < the optimum cables are: > < 132 ohms imp. 50 cond. flat or 25 signal twisted pair cable for Single Ended > < 122 ohms imp. 50 cond. or 25 signal twisted pair cable for Differential > Don't bother calling your local cable vendor and asking for 132 ohm 50 cond cable. They won't have it. And even if they did, you'd never find connectors with a 132 ohm impedance. Just use standard 100 ohm cable. > < both should be shielded. > < > On the other hand, if radio&TV antenna and any heavy-duty applicances are > sufficiently far away, you may be able to forget about the shielding. As long as you don't want to watch TV while computing and you don't have technically-astute neighbors, it shouldn't be too hard to get away with it. If it screws up their reception and they call the station to complain, the station will likely just tell them to buy a bigger antenna or a better receiver. I only know one person who has ever proved to the FCC that a neighbor was above legal RFI emmision levels, and it took MONTHS of that person's time to do it. Please note that I am not advocating the breaking of laws, or encouraging you to break any laws. I am just stating personal opinion and experience. And to add a note to the descriptions of 100 foot buses: There are timings written into the SCSI spec for things like bus delays, skew delays, etc. Many SCSI interface chips hardware time these delays. If the bus is sufficiently long to cause these times to exceed the spec values, those interface chips may run into problems. I have never experienced this personally (having never run a bus longer than ~50 ft), but I accept it as a fact. -- William R. Carroll (Encore Computer Corp., Ft. Lauderdale FL) wcarroll@encore.com uunet!gould!wcarroll "This comment without commentary is not the view of the staff or management of WKPX, the Broward County School Board, ... or even me." - KZ