Fast-20 active negation current

Tak Asami asami at
Tue Jan 24 08:55:35 PST 1995

Thanks to Bill Ham and John Lohmeyer for answering my question so quickly
regarding Fast-20 active negation current.

The requirement of higher active negation current, however triggers 
another question.
We are supposed to use the "Single-ended test circuit" in Figure 1 as 
a test load for the purpose of slew rate measurement.

In order to keep 520mV/nsec slew rate specificaiton, I cannot turn on
pull-up device at all. I need to slow down the turn-off of pull-down 
device to achieve this.  Otherwise, the voltage source in the load can 
pull up the test node like no tomorrow.
But this means that during the low to high transition, there is zero
current available from active negation component.

Furthermore, the 22mA current requirement can't be applicable to steady
state, since the terminator will not let you go down below 2.0V DC.
So 7mA or less suffice for steady state high.

So, this 22mA reqirement must apply to AC transition phase (as indicated
in the document) below 2.0V, where the devices are in triode region.
>From this, I assume someone decided the Fast-20 loading is much heavier
and slow responding than the one in Figure 1.
If not, I don't know how anyone can keep both Ioh > 22mA and slewte(max)
= 520mV/nsec specs simultaneously.
Can anyone show any electrical model that warranted the high current
to drive Fast-20 bus?  That might give us a clue.

BTW, when we talk about AC transition current of a driver, it has (almost)
nothing to do with DC source current Ioh.  During the transition, more
current is certainly available.  For example, a simple 4mA push-pull driver
can often able to provide 20-30mA of transition current, but only for short
time.  That may be sufficient to cover this case.

So, in summary (in my own prejudiced mind),
1) Ioh > 22mA requirement is NOT a needed DC specification.
   7mA DC driver can provide that much current for AC transition.
   The need for extra transition current should be in the standard
   document in words.
2) To satisfy AC requirement, we need a different AC load model.
   Perhaps a parallel combination of 47ohm and 15pF to ground for
   low to high transition, to VDD for high to low transition?
3) This model should also be used for slew rate calculation.
4) Note the extra transition current was not needed in Fast-10 and
   below. This extra current is likely to violate the slew rate
   specification in that range.

I'd like to hear more comments on this subject.....

Tak Asami ==================================================================
Western Digital Corp.
I/O Product Engineering
asami at
(714) 932-7621 : Voice
(714) 932-6496 : Fax

More information about the T10 mailing list