Dear all SI participant:
This is Kyung-Hwa, Kim in Korea.
My major roles are simulation and analysis on
signal integrity for high speed digital circuit
and systems.
I want to ask the following questions to you.
**************[Begin of situation]*************
I want to simulate the driver's output signal
which have 10ns's Tr/Tf and 0 to 5V swing with
1000pF capacitive load.
1000pF loads are actually distributed having
equal distance on burn-in board.
(please ref. following config.)
This driver has to drive 24 receivers(64M SD
RAM)
distributed in burn-in board.
We don't use daisy chain system but bus stytle
to distribute driving signals.
(main net length : 50 in., stub length : 2 in.)
Our driver's short circuit current is estimated
about 170mA. This driver locate on seperated driver
board.
Estimated loading conditions for one driver is
496pF(min) or 568pF(max).
Followings are loading condition.
[Receiver's loading for 24 of SDRAM, PKG, socket]
- SDRAM's input capacitance :
3~5pF/ea -> 24 X (3~5pF) = 72/120pF
- PKG & socket :
3~4pF/ea -> 72/96pF
[24 stubs and one main net]
- stub : 4pF/in.(2in. for one stub)
&
nbsp; -> 24 X 8pF = 192pF
- main net :
4pF/in.(40in.) -> 160pF
[receiver] 1 2 3 4 5 6 7 8 9 10 11 12
| | | | | | | | | | | |
| | | | | | | | | | | |
[Driver] --------+--+--+--+--+--+--+--+--+--+--+--+---R to Gnd
Ron=29.4 | | | | | | | | | | | | (30ohm)
Tr/Tr=10ns | | | | | | | | | | | |
Voh/ol = 5/0V 13 14 15 16 17 18 19 20 21 22 23 24
Isc = 170mA
***************[End of situation]**************
***************[Begin of question]*************
Our earlier driver's short circuit current is
40mA when the driver operates at 30ns's Tr/Tf for
same loading condition. The Zo of strip/microstrip
is 50 around. This system operates correctly.
Now we designed upper bus system by 30 ohm
strip/microstrip lines with 10ns's Tr/Tf driver.
But this system doesn't operate at this frequency,
showing over/undershoot and steps in rise/fall
edges.
-----------------------------------------------
[My guess for short circuit current]
(referenced by "CMOS/TTL Digital System Design"
pp.2
02 by James E. Buchanan)
dV/dt = Isc / C
where Isc is the output short circuit current
of the output driver and C is the increase in
load cap. beyond the specified load value.
5V/10ns = Isc / 568pF, 5V/10ns = Isc / 496pF
Isc = 284mA Isc = 248mA
-----------------------------------------------
[Q1]
If our driver doesn't have 284/248mA Isc
characteristics.
We estimated it has 170mA.
What kind of problems could be occured?
It couldn't drive the receiver located far
end of main net?
It could show large delay or steps in rising
edge?
[Q2]
My short circuit current estimation is
acceptable?
I've solved the equat
ion by one lumped
capacitance for net and stub's transmission
line capacitance.
[Q3]
Which effects could be emerged by designing
the strip/microstrip's Zo to 30 from 50? We
reduced the Zo from 50 to 30.
And also I use bus system for 10ns's driver.
[Q4]
We tested CMD(California Micro Devices Co.)'s
PAC DN005Q for our end(load) terminator. This
terminator is consisted of shottkey diode which
has reverse recovery time less than 400ps.
(Please check for this terminator at
www.clamicro.com/prod/spice/s_diode_sub.htm)
But this terminator couln'd clamp any
over/undershoot.
Could anyone let me know why it doesn't
operate correctly?
Does the huge driving current make this
problem?
[Q5]
Our driver has been specially designed
by Analog Device.
We developed driver module containing 8 of
this driver and this module is attached to the
driver board.
One driver's Isc is 170mA. So, the simultaneous
switching current in this module is about
170mA*8= 1.36A. The module size is 3 by 1 in.
Is this switching current too big to this
size of module? If we doesn't have ideal
power/ground plane and short power/ground net,
it could make ground bound or power droop
problem? So, the driving characteristics could
be heavily damaged?
[Q6]
We only checked the driver's Tr/Tf and
voltage swing level.
If the driving current(Isc) has major role
to drive the 24 receivers and transmission load,
how could we measure the driver's Isc correctly?
*****************[End of question]***************
**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****