RE: [SI-LIST] : measuring CMOS Cin with HP4275A

Charles Hill (chuckh@csn.net)
Fri, 10 Apr 1998 16:12:04 -0600

Greg,

Your initial guess is correct. There is a grounding problem, and this
meter was not designed to do this task. Let me explain. To measure
impedance, one must measure both voltage and current in amplitude and
phase. One lead of the HP4275A is a voltage source for excitation, the
other lead is a current to voltage converter that acts to maintain a
virtual ground at that lead. The excitation is known, so the voltage
across the DUT is determined, the current to voltage converter provides the
means to determine the current. Both current and voltage are measured in a
vector fashion, so the impedance is calculated.

Note the DUT should not have any portion grounded since this provides an
alternate path for the current through the DUT and the impedance
measurement suffers. Many LCR meters are designed with this limitation.
This is fine for resistors, capacitors, and inductors.

Another difficulty is the test frequency. If the capacitance you wish to
measure is 10pF, then what is the current at the test frequency? At 1MHz
test frequency, the current is 63uA for 1 volt oscillator level. The noise
floor of the instrument and test setup could be a problem. Another problem
is the phase angle of the DUT will be near either 90 or 0 degrees and the
accuracy of the resistive and reactive parts, respectively, suffers.

There are other instruments that work better. In the auto balancing bridge
category, the HP4192 has a 1K (if I remember correctly) to ground on the
virtual ground side of the load. This should prevent your grounding
problem. I routinely use an HP4396 impedance analyzer to measure Cin of
CMOS devices. Doing a proper measurement does require care, though.
Calibration is a important part. It sounds like you are on the right
track.

Charles Hill, consultant
chuckh@altaeng.com
719-495-3007

----------
From: Greg Edlund[SMTP:Greg.Edlund@digital.com]
Sent: Friday, April 10, 1998 8:38 AM
To: 'si-list@silab.eng.sun.com'
Subject: [SI-LIST] : measuring CMOS Cin with HP4275A

I have a board in the lab with a test structure for measuring the
capacitance of a CMOS input on a powered-up component. I'm using an
HP4275A, which is an LCR meter that utilizes an auto-balancing bridge
circuit. I can make the measurement with the board isolated from any
power source, but as soon as I even connect the board to a powered-off
supply, the meter reads overflow. It sounds like I have some kind of a
ground loop problem. Am I trying to do something that this meter was
not designed to do?

If you could lend your experience, it would be much appreciated. It's
for a good cause, too. When we finish this board, we'll be posting the
Gerbers and schematics to the IBIS web site. The purpose of the board
is to demonstrate how to make measurements that verify an IBIS
datasheet. (It's also useful for verifying SPICE models). Thanks in
advance.

Description of my test set-up:

There are three SMA connectors on the board: one for open cal, one for
short cal, and one that goes to the DUT. They each have an identical
piece of etch connected to the SMA center pin.

The probes are home-built from two dual-conductor, shielded cables. On
the instrument side, one of the cables is connected to Hp and Hc through
a two isolated pieces of copper foil mounted on perf board. The other
cable is connect to Lp and Lc in the same manner. The shields are
connected to chassis ground.

On the DUT side, the Hp and Hc conductors of one cable are connected to
the SMA center pin. The Lp and Lc conductors of the other cable are
connected to the SMA ground. The shields are connected to each other
but floating.

----------
Greg Edlund, Principal Engineer
Server Product Development
Digital Equipment Corp.
129 Parker St. PKO3-1/20C
Maynard, MA 01754
(978) 493-4157 voice
(978) 493-0941 FAX
greg.edlund@digital.com