RE 3: Re 2: Revised proposal on MSE high-impedance spec

Walter Bridgewater wally at
Tue Jun 22 11:32:52 PDT 1999

* From the T10 Reflector (t10 at, posted by:
* wally at (Walter Bridgewater)

Wouldn't a SCSI device that drove to 5.5 volts with a P-channel
active negation, already be expecting that the termination would
be clamping the high going current?

So the termination could be sinking current to a 2.5 volt supply
through a 44 ohm resistor.  This is for both ends of the termination.
Then, in parallel with that, you could have a clamp to 3.0 volts,
through 100/15, or 6.66 ohm.

Since even for SCSI-2 devices, the spec said that they shouldn't 
source current above 3.7 volts.

At 3.7 volts, termination would sink 27.27mA  (3.7-2.5)/44,
& @ 3.7 volts, clamping would sink 105mA,  (3.7-3.0)/6.66

But, the driver is supposed to sink 0mA at 3.7 volts, so,
it could never really make it to 132mA.

I agree with John, we can't keep from going forward because of
somebody mis-designs.

But, on the other hand, one way out of this entirely, would be to
abandoned Single-Ended SCSI completely.



> *
> At 6/22/99 09:43 AM , Gene_Milligan at wrote:
> ><<With multiple high-Z MSE receivers I believe this intersection could be
> >on the 20 mA segment of Figure 26. However, I don't think the 160 mA in
> >Gene's original comment is possible due to the limit in Figure 26.>>
> >
> >     I guestimated the upper limit which did not quite make it into the
> >figure was at about 10 ma per receiver. I presumed we did not need to worry
> >about 32 bit wide busses which could have 32 SCSI devices and judged the
> >limit to be 16 SCSI devices. Thus 160 mA.
> Gene and Richard,
> Richard is much closer to reality with his 7 mA number for SPI-3 buses.  20
> mA might be possible theoretically with a little ground shift, but I think
> it is unlikely that we will even  see 7 mA with modern active negation
> drivers.  
> Gene's 160 mA is an extreme worst-case number that could only happen with
> 15 (not 16) 3-volt devices and one 5-volt active negation device that has a
> P-channel negation transistor which clamps the signal to 5 volts.  I think
> there were a few such devices designed in the early days of active negation
> (circa late SCSI-2 and early Fast-20).  People quickly learned that such
> strong active negation transistors caused more problems than they helped.
> In fact, these devices worked better with active negation disabled -- so if
> we can draw enough current to blow out these deviant transistors, we'll be
> doing them a favor!  ;-)
> All seriousness aside, should we handicap our ability to move forward with
> new processes just to avoid a potential problem with very old devices?  How
> likely is it that we'll encounter any of these old devices?
> John
> --
> John Lohmeyer                  Email: lohmeyer at
> LSI Logic Corp.                Voice: +1-719-533-7560
> 4420 ArrowsWest Dr.              Fax: +1-719-533-7036
> Colo Spgs, CO 80907              BBS: +1-719-533-7950

> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at

* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at

More information about the T10 mailing list