Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
I have several SCSI drives from old servers that I need to pull data
from onto my laptop.
I have been looking for USB or FireWire adapters with no luck.
Adaptec seems to have discontinued support for most of these and they
are all 50-pin devices anyway, which doesn't help me.
Has anyone ever seen a USB, FireWire or PC Card(PCMCIA) to SCSI-3
adapter?
I have several SCSI cable adapters (internal/external, hot-swap/
internal, etc.) so I can deal with that, but the SCSI end must be 68-
pin.
I would appreciate any suggestions. Thanks.
<los.varas@gmail.com> wrote in message news:1185555234.656545.99340@x35g2000prf.googlegro ups.com...
>I have several SCSI drives from old servers that I need to pull data
> from onto my laptop.
> I have been looking for USB or FireWire adapters with no luck.
> Adaptec seems to have discontinued support for most of these and they
> are all 50-pin devices anyway, which doesn't help me.
> Has anyone ever seen a USB, FireWire or PC Card(PCMCIA) to SCSI-3
> adapter?
> I have several SCSI cable adapters (internal/external, hot-swap/
> internal, etc.) so I can deal with that, but the SCSI end must be 68-
> pin.
>
Re: Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
> I have several SCSI drives from old servers
> that I need to pull data from onto my laptop.
> I have been looking for USB or FireWire adapters with no luck.
Yea, it's slow (USB 1.1, not 2.0) but it's a one time use project.
If the device is wide, just use a narrow to wide adapter or cable.
Speed is not of the essence, just quick connectivity.
Or: connect the SCSI drives to a spare PC with SCSI interface
(if you have one or can borrow one)
and share the results via the network to the laptop.
A bootable "Live Linux" CD could be part of the solution
if the system doesn't even boot on its own
or doesn't have the SCSI drivers.
--
Re: Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
Jeff Jonas wrote:
>
> How about
> http://www.geeks.com/details.asp?inv...B2SCSI&cpc=SCH
> $14.99
> USB 1.1 To SCSI Adapter
>
> Yea, it's slow (USB 1.1, not 2.0) but it's a one time use project.
> If the device is wide, just use a narrow to wide adapter or cable.
Two issues to remember:
1) Wide termination
Some Wide-SCSI disks don't like it if their high-bus is unterminated. So
use an 50/68pin-Adapter with terminator for this job onboard.
2) Termination power
USB cannot provide enough energy for the terminators. Calculated with
two active SE terminators [1] for the bus and a passive SE
high-terminator on the adapter the worst case current on TERMPOWER is:
2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 220Ohm) = 1.14A
With all terminators passive:
2 * 18 * (5V / 220Ohm) + 9 * (5V / 220Ohm) = 1.02A
With floating high bus on the disk (not recommended):
2 * 18 * (5V / 220Ohm) = 818mA
In all cases this will overload an USB port. The link say nothing about
the hostadapter is shipped with external power supply, so at least one
of the disks that are connect must supply TERMPOWER to the SCSI bus.
Re: Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
"Michael Baeuerle" <michael.baeuerle@gmx.net> wrote in message news:modsn4-oq1.ln1@micha.freeshell.org
> Jeff Jonas wrote:
> >
> > How about
> > http://www.geeks.com/details.asp?inv...B2SCSI&cpc=SCH
> > $14.99
> > USB 1.1 To SCSI Adapter
Not really suited for bare drives.
> >
> > Yea, it's slow (USB 1.1, not 2.0) but it's a one time use project.
> > If the device is wide, just use a narrow to wide adapter or cable.
>
> Two issues to remember:
Only 2?
>
> 1) Wide termination
> Some Wide-SCSI disks don't like it if their high-bus is unterminated. So
> use an 50/68pin-Adapter with terminator for this job onboard.
You still need conversion from that USB 1.1 To SCSI Adapter male 50p HD connector.
>
> 2) Termination power
> USB cannot provide enough energy for the terminators.
That's debatable.
> Calculated with two active SE terminators [1] for the bus and a passive SE
> high-terminator on the adapter the worst case current on TERMPOWER is:
> 2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 220Ohm) = 1.14A
Huh?
Active terminators only draw current when data and/or commands are trans-
ferred. To draw 400mA, all the data and control lines need to be continu-
ously active *all the time*. Second, the high-byte terminator does not draw
any power as it doesn't carry any signals and isn't a real terminator. It's only pull-up. Even *if* a real terminator, the
resistance is 550Ohms/line.
Re: Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
Folkert Rienstra wrote:
>
> Michael Baeuerle wrote:
> >
> > Two issues to remember:
>
> Only 2?
>
> > 1) Wide termination
> > Some Wide-SCSI disks don't like it if their high-bus is unterminated. So
> > use an 50/68pin-Adapter with terminator for this job onboard.
>
> You still need conversion from that USB 1.1 To SCSI Adapter male 50p HD connector.
An old external cabinet with matching connectors (as used for CDROMs)
will do.
There are also PC case slot covers with HD female extern and 50Pin male
intern that can be used for this task.
> > 2) Termination power
> > USB cannot provide enough energy for the terminators.
>
> That's debatable.
>
> > Calculated with two active SE terminators [1] for the bus and a passive SE
> > high-terminator on the adapter the worst case current on TERMPOWER is:
> > 2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 220Ohm) = 1.14A
>
> Huh?
> Active terminators only draw current when data and/or commands are trans-
> ferred. To draw 400mA, all the data and control lines need to be
> continuously active *all the time*.
Right, but this was intended for my "worst case" calculation.
> Second, the high-byte terminator does not draw
> any power as it doesn't carry any signals and isn't a real terminator.
> It's only pull-up. Even *if* a real terminator, the
> resistance is 550Ohms/line.
My fault, thanks for correction. The high bus never should have asserted
signals in narrow mode.
The correct worst case current is then:
2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 550Ohm) = 1.01A
with real high terminator and:
2 * 18 * (2.85V / 110Ohm) = 933mA
with pull-ups.
> > [...]
> > In all cases this will overload an USB port.
>
> Hardly. And since there is no SCSI bus *to speak of* it's likely
> there isn't even a terminator on the USB-SCSI converter.
If you plug the USB-SCSI converter directly to the device a single
terminator should be sufficient. Then the maximum current is only
approx. 500mA and it should work.
But with cable you should connect the second terminator if it is not
integrated.
> > The link say nothing about the hostadapter
>
> It's not really a "hostadapter".
Why not? It communicates via USB instead of PCI, but IMHO that's no
difference in principle.
> > is shipped with external power supply, so at least one
> > of the disks that are connected
>
> Supposedly one (in a cabinet). Either that or a multidrive cabinet.
If all fails, an external power supply can do it with a fuse and a
diode. I have done it like this for an Iomega IEEE-1284 hostadapter with
a CDROM connected that cannot provide TERMPOWER: http://micha.freeshell.org/traveller/index.php
Re: Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
"Michael Baeuerle" <michael.baeuerle@gmx.net> wrote in message news:vss1o4-k61.ln1@micha.freeshell.org
> Folkert Rienstra wrote:
> >
> > Michael Baeuerle wrote:
> > >
> > > Two issues to remember:
> >
> > Only 2?
> >
> > > 1) Wide termination
> > > Some Wide-SCSI disks don't like it if their high-bus is unterminated. So
> > > use an 50/68pin-Adapter with terminator for this job onboard.
> >
> > You still need conversion from that USB 1.1 To SCSI Adapter male 50p HD connector.
>
> An old external cabinet with matching connectors (as used for CDROMs) will do.
> There are also PC case slot covers with HD female extern and 50Pin male internal
> that can be used for this task.
>
> > > 2) Termination power
> > > USB cannot provide enough energy for the terminators.
> >
> > That's debatable.
> >
> > > Calculated with two active SE terminators [1] for the bus and a passive SE
> > > high-terminator on the adapter the worst case current on TERMPOWER is:
> > > 2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 220Ohm) = 1.14A
> >
> > Huh?
> > Active terminators only draw current when data and/or commands are trans-
> > ferred. To draw 400mA, all the data and control lines need to be
> > continuously active *all the time*.
>
> Right, but this was intended for my "worst case" calculation.
Which you then used as a permanent figure that USB would not support.
What matters is what the average current draw is in a certain timeframe
and whether that will trip any fuse.
>
> > Second, the high-byte terminator does not draw
> > any power as it doesn't carry any signals and isn't a real terminator.
> > It's only pull-up. Even *if* a real terminator, the resistance is 550Ohms/line.
> My fault, thanks for correction.
> The high bus never should have asserted signals in narrow mode.
> The correct worst case current is then:
Not worst case, the max. peak current.
And it is highly debatable whether this worst case scenario even exists.
I think that only a few control lines (if any) can be active at the
same time, reducing that 18 number.
> 2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 550Ohm) = 1.01A
> with real high terminator
> and:
> 2 * 18 * (2.85V / 110Ohm) = 933mA with pull-ups.
>
> > > [...]
> > > In all cases this will overload an USB port.
> >
> > Hardly. And since there is no SCSI bus *to speak of* it's likely
> > there isn't even a terminator on the USB-SCSI converter.
>
> If you plug the USB-SCSI converter directly to the device a single
> terminator should be sufficient. Then the maximum current is only
> approx. 500mA and it should work.
>
> But with cable you should connect the second terminator if it is not
> integrated.
>
> > > The link say nothing about the hostadapter
> >
> > It's not really a "hostadapter".
> Why not? It communicates via USB instead of PCI, but IMHO that's no
> difference in principle.
It's a converter. The host adapter is the USB controller.
>
> > > is shipped with external power supply, so at least one
> > > of the disks that are connected
> >
> > Supposedly one (in a cabinet). Either that or a multidrive cabinet.
> If all fails, an external power supply
The one powering the drive.
> can do it with a fuse and a diode.
The fuse is a safety precaution preventing the terminator to start a fire
in case it developes a fault. Since this is a temporary solution in a managed
environment it isn't necessary.
The diode isn't necessary either since this will be the only power source.
Re: Looking for USB, FireWire or PC Card to 68-pin SCSI adapter
Folkert Rienstra wrote:
>
> Michael Baeuerle wrote:
> >
> > Folkert Rienstra wrote:
> > >
> > > Michael Baeuerle wrote:
> > > >
> > > > [...]
> > > > 2) Termination power
> > > > USB cannot provide enough energy for the terminators.
> > >
> > > That's debatable.
> > >
> > > > Calculated with two active SE terminators [1] for the bus and a passive SE
> > > > high-terminator on the adapter the worst case current on TERMPOWER is:
> > > > 2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 220Ohm) = 1.14A
> > >
> > > Huh?
> > > Active terminators only draw current when data and/or commands are trans-
> > > ferred. To draw 400mA, all the data and control lines need to be
> > > continuously active *all the time*.
> >
> > Right, but this was intended for my "worst case" calculation.
>
> Which you then used as a permanent figure that USB would not support.
> What matters is what the average current draw is in a certain timeframe
> and whether that will trip any fuse.
If you draw more current as allowed, the supply voltage can drop below
the Limit immideately. So malfunction can occur long before the fuse
will blow. The fuse is slow and only prevents damage but does not ensure
proper operation in any way.
> > > Second, the high-byte terminator does not draw
> > > any power as it doesn't carry any signals and isn't a real terminator.
> > > It's only pull-up. Even *if* a real terminator, the resistance is 550Ohms/line.
> >
> > My fault, thanks for correction.
> > The high bus never should have asserted signals in narrow mode.
> >
> > The correct worst case current is then:
> > 2 * 18 * (2.85V / 110Ohm) + 9 * (5V / 550Ohm) = 1.01A
>
> Not worst case, the max. peak current.
Name it as you like.
> And it is highly debatable whether this worst case scenario even exists.
> I think that only a few control lines (if any) can be active at the
> same time, reducing that 18 number.
Ack. All lines active should not happen in real world.
Active lines for a MESSAGE IN phase with 0xFF data as example:
----------------------------------------
Data (inverted): 8
Odd parity (inverted): 1
BSY (Bus busy): 1
MSG + C/D + I/O (MESSAGE IN phase): 3
REQ + ACK (Interlock overlap): 2
----------------------------------------
In addition, the initiator can set ATN at any time what leaves only RST
and SEL remaining inactive.
So 16 of 18 lines active can happen under normal operation.
To get the average current let's take the very common DATA IN phase with
0xF0 data (4 data lines active):
----------------------------------------
Data (inverted): 4
Odd parity (inverted): 1
BSY (Bus busy): 1
I/O (DATA IN phase): 1
REQ or ACK (no overlap): 1
----------------------------------------
2 * 8 * (2.85V / 110Ohm) + 9 * (5V / 550Ohm) = 496A
Even the average is near the USB limit, so the real current pattern
surely will not be allowed by the USB spec. It may nevertheless work in
the illegal area as most bus-powered 2.5" disks do ...
> > > [...]
> > > It's not really a "hostadapter".
> >
> > Why not? It communicates via USB instead of PCI, but IMHO that's no
> > difference in principle.
>
> It's a converter. The host adapter is the USB controller.
Following your scheme an AHA-2940 is also a "converter" because the
hostadapter is the PCI controller. But even an AHA-1542 in a modern PC
is mostly called "SCSI hostadapter" and is connected double indirect via
PCI and ISA controllers to the FSB of the CPU.
This is a question of definition. In principle all I/O busses like ISA,
PCI, USB and SCSI simply transports data. So if you name the USB
function "controller" and be really consequent, the PCI function is a
controller and the SCSI function is a controller too.
> > > > is shipped with external power supply, so at least one
> > > > of the disks that are connected
> > >
> > > Supposedly one (in a cabinet). Either that or a multidrive cabinet.
> >
> > If all fails, an external power supply
>
> The one powering the drive.
For example.
> > can do it with a fuse and a diode.
>
> The fuse is a safety precaution preventing the terminator to start a fire
> in case it developes a fault. Since this is a temporary solution in a managed
> environment it isn't necessary.
> The diode isn't necessary either since this will be the only power source.
Ack, for such a Q&D solution TERMPOWER can be connected directly to 5V.
But the example show that even the correct solution is not really
complicated.
Looking for USB, FireWire or PC Card to 68-pin SCSI adapter 
Linear Mode
