time, and very wide bandwidth. The downside is heavy
loading and a high probe attenuation factor. For the two
values of R shown, the attenuation is 10X with 450 ohms
and 20X with 950 ohms. Also be aware that it is DC
coupled. Ideally, these connect to a scope’s 50 ohm input
port (which most quality scopes with >150 MHz BW have).
In lieu of that, a 50 ohm termination will have to be added
between the scope’s normal one meg input jack and the
probe’s BNC termination. Unfortunately, the scope’s
approximate 20 pF of input capacity at this port will still be
present under this situation and will present an increasing
VSWR vs. frequency, skewing results above 100 MHz. In
spite of this probe’s shortcomings, it has very low input
capacitance (<1 pF), thereby presenting an almost purely
resistive load to the circuit under test; VHF frequencies and
above will actually have less loading effect than the
aforementioned 10X passive probes. As cautioned, the
probe is DC coupled but is actually an advantage in
standard five volt logic circuitry. In my experience, there was
a 20% loading factor in probing high speed CMOS logic
devices, but the fidelity is preserved and predictable.
The active probe is shown in simplified form in Figure
2B. Typically, the probe initially couples the signal through a
very small coupling capacitor (although some are DC
coupled through an RC network) into a high input
impedance amplifier. The input network is adjusted to give a
10:1 attenuation ratio with a very small Cin (<1 pF). The
amplifier will have a gain of X1 to X10. This gives the probe
an attenuation ratio of anywhere from 1X to 10X. It is rare
to find these probes with a 1X configuration (no
attenuation).
More commonly, they will be in a 5X or 10X
configuration. In all cases, the output is converted to 50
ohms Z and sent down a 50 ohm transmission line to the
scope’s vertical input. The down side of these probes
include the Bulky probe tip (in some models), external DC
power required, limited dynamic range, and a shocking
sticker price averaging $2,000 to $3,000. The highest priced
one I ever saw came in just under a whopping $16,000.
Their upsides are superior bandwidth (some > 10 GHz),
excellent rise time, and high R/ low C probe input specs.
However, these phenomenal specs do not come without
some strings attached. When dealing in precise and faithful
probing in the realm of microwave frequencies that these
probes are capable of, the probe’s accessories and
techniques become almost as important as the probe itself
to insure quality results. Depending on the probing
situation, special grounding adapters have to be used and
special resistive probe tips have to be interchanged (as
many as a dozen or more). In some cases, special sockets
have to be installed and soldered directly to the circuit
board which will mate up to an accessory probe tip. Include
a thorough understanding of the circuit being probed so
that the correct accessories will be installed and it’s easy to
see that these high-end probes are not for the amateur.
Constructing one of these probes at the hobbyist level
would be very difficult to say the least. However, in spite of
the shortcomings of the common 10X passive probes, they
are still the workhorse of the industry and are very
worthwhile for general-purpose probing. Their dynamic
range far outpaces any other type probe made and often
approaches upper limits of 500-600 volts. I would always
include a couple in my arsenal of probes.
What I Did About the Problem
I have always been intrigued by the concepts of the low
Z passive probes and active probes. After years of
frustration using the common 10X passive probes, I decided
it was time to switch to one of these superior probes. Since
I could not justify the expense involved in a commercial
product, I thought I would design and construct one myself.
Each probe has its benefits and drawbacks. The low Z has
excellent fidelity, rise time, and bandwidth but gives way
too high loading and attenuation. The active probe also has
these great features but with the added benefit of low
loading and low attenuation. However, it comes with the
added burden of a bulky probe size (in my design), limited
dynamic range, necessity of external power, and difficulty
constructing and calibrating the input network. In mulling
over the various pluses and minuses of these probes, I
decided to construct a hybrid of the two, which I
humorously call the “Passive-Active” probe. Figure 3 shows
the completed unit.
The probe starts out in the traditional low Z style with
a resistive divider feeding a 50 ohm transmission line, but
with one exception: a much higher input impedance (i.e.,
3,400 ohms as opposed to 1,000 ohms). This is fed to an
active probe’s traditional amplifier circuit before being
applied to the scope’s input. The amplifier makes up for
most of the divider losses. This probe still suffers some of
the unavoidable consequences of its commercial
counterparts, such as a limited dynamic range, moderate
circuit loading, and necessity of external power.
Due to design constraints, the optimum dynamic range
was determined to be 8V p-p input at its upper end with the
low end of that range being 15 mV p-p input to obtain two
divisions of vertical deflection (typical sensitivity for wide
FIGURE 3
November 2010 47
