(Above) Enclosed in a 16 cm x
16.5 cm plastic box, the preamp has a 60 cm cable to supply power and
take the signal to the frequency meter/counter. The cable was
originally part of a USB mouse. The labels were printed on a laser
printer. Yes, that's clear plastic tape over the labels so my fingers
will not rub the toner off the paper.
Overview
After finishing
Little More Serious Frequency
Meter I had planned to make a suitable preamp and prescaler
for it, and set about to gather ideas and parts. I was inspired by one
fellow who had built the meter and the
2
line X16 character LCD display to show the output, succeeded in
designing a preamp based on the BRF96 and modifying the circuit to get
it to work at 99.999999 MHz. My intentions are to use the frequency
meter between 100 and 200 MHz, so I found a prescaler, the Motorola,
now On Semiconductor, MCT10280. For the preamp, I was able to buy some
2N3663 transistors. I would have like to have tried the BRF96, but
could not find them stocked at any of my favorite distributors.
What was probably the most difficult part of the design was deciding
what I really wanted the circuit to do and how to go about doing it,
given the limitations of the available components. The resulting preamp
can drive the counter from as little as 20
millivolts P-P input at a few KHz, but needs over 300 millivolts at 20
MHz, and can switch in a divide by 10 prescaler to extend the range
to 300 MHz. I have used this at 338 Mhz to date. The data sheet
limit for the 74HC4060 is 30 MHz, so performance over 300 MHz is
expected by beyond specification.
A
Counter mode allows a
direct coupled signal to drive the frequency meter in the counter mode.
The resulting preamp/prescaler design intended to operate within these
frequency limitations:
MODE
DESIGN SPECIFICATION
Count
DC to several hundred kHZ (in practice, several MHz)
F/1 Frequency, no
prescaler 10 Hz to 30 MHz
F/10 Frequency, 10X prescaler 10 MHz to 300 MHz
(Mine worked at 338 MHz.)
These parameters are expected with an approximately 50% square wave up
to frequencies of seveal MHz, and symmetric sine waves at higher
frequencies. The primary limitation is based on the maximum clocking
rate specification for the MM74HC6040 ripple counter in use in the
Slightly More Serious Frequency Meter project. Selection of faster
parts and careful circuit layout can extend the upper limit of the
useful frequency range. The lower frequency imit is the lowest sine
wave input frequency for the MCT12080 at which the input of the
MCT10290 does not osccilate.
Circuit
(
(Above) The cirucit blocks
are,from left to right, input protection circuit, prescaler, FET buffer
and bipolar limiting amplifier, and Schmitt trigger buffers.
Signals applied to the input connector
can be switched either through the AC path which includes the preamp
and the prescaler, if switched into the circuit, or the DC path, which
routes the signal to a Schmitt trigger buffer that then sends the
signal on to the counter.
Regardless of how the signal is routed, it must first pass through an
input protection network, which includes two schottky diodes and a
zener clamp. The 1N5711 schottky diodes prevent the input signal from
going more than a schottky diode drop below ground or above the power
supply. I used Schottky diodes because they have a lower voltage drop
than the PN protection diodes on the CMOS integrated circuit they are
intended to protect, and as such, they will draw much more of the
current from excessive input voltages than the input protection diodes
in the integrated circuits.
The two 1N5226 zener diodes in series prevents the power supply from
rising above 6.6 volts in case the input is accidentally connected to a
low impedance source that is higher than 5 volts. The 47 Ohm resistor
limits the input current in case of excessive voltage being applied to
the inputs.
The input of the frequency meter requires a full 5 volt CMOS logic
swing, and the prescaler's output is less than 1 volt peak-to-peak, so
the prescaler, when switched into the circuit, the signal goes through
the prescaler, then the preamp, and the preamp drives the frequency
meter through the Schmidt trigger buffers.
The MCT10280 prescaler can be set to divide by 80, 40, 20, or 10, as a
function of which pins are tied to the power supply. I set this one to
divide by 10 since it is adequate for my needs, and the mental
calculation of multiplying the meter reading by 10 is not taxing.
One problem with the MCT10280 is that if it doesn't have an adequate
input, the output is very noisy, which shows up as counts in the couple
MHz range on the frequency meter. This noise shows up if the
signal amplitude the signal frequency is too low. For this reason, I
only intend to use the prescaler with inputs between 10 MHz and 300 MHz.
Whether the signal is divided by ten or not, it must pass through the
preamp. In the Preamp, a 2N5485 N-channel FET is connected as a source
follower. This provides a high input impedance to the input signal and
drives the next stage, the 2N3663 limiting amplifier with a nice low
impedance signal. This results in high AC gain.
DC buffer
Signals from the 2N3663 are nice, clean square waves with fast rise and
fall times, and when driving the 74HCT02 input on the frequency meter,
could result in the input of the 74HCT02 oscillating. To prevent this,
the signal is passed through a 74HCT14 Schmitt trigger. The signal
conductor in the cable has an impedance of about 150 ohms, so it is
driven through two 300 ohm resistors in parallel to keep the ringing on
the frequency meter end of the cable to a tolerable level. There is no
termination on the frequency meter end of the cable because terminating
it would reduce the amplitude of the signal below the CMOS thresholds.
Power for all the circuit comes through the cable and is regulated by
the 78L05 regulator.
Assembly
and Test
The great bulk of the work in assembling
this was in cutting the circuit board to shape and making the holes for
the input connector, switches, and the cable in the plastic case. After
that, it was mostly a matter of placing the components for minimum lead
length and hooking them up point-to-point.
The switches could not be mounted on the circuit board because the
standoff height was not sufficient and I hadn't anything to extend the
standoffs with, so flying leads were used to connect the slide switches
to the circuit board. It would have been nice to no have the flying
leads, but they appear to not be causing any problems with
performance. I have not investigated the circuit's upper
frequency limit,
where small differences in layout and wiring could have large effects
on performance, except to note that with the prescaler switched into
the circuit, the frequency meter is able to measure the frequency of a
100 MHz RF source.
I built up a
subassembly with the prescaler and tested it before mounting the
subassembly on the main board. I would have almost been worth
making a
printed circuit board for this. See the photograph of the surface mount
subassembly below.
(Above) The MCT10280 prescaler is
mounted along with chip resistors and chip capacitors and the leaded
78L05 on a separate circuit board that was in turn mounted on the
larger assembly. The capacitors were taken from discarded cell phones.
(Above)The threshold for square
waves is lower than for sine waves at frequencies below 100 Hz because
the edges of the square waves make it through the AC coupling. These
measurements were taken without the prescaler connected.
Once assembled, I measured the performance and found that I needed
nearly a volt of input signal to get the frequency meter to read
properly. I added an extra ground wire between the 2N3663 emitter and
the ground pin on the input connector, and that helped a little bit.
The big improvement came when I shorted the 2.2k resistor on the input
of the frequency meter. It is not entirely clear to me why this made
such a dramatic difference, but it lowered the threshold to a little
over 300 millivolts at 20 MHz.
I used the preamp in the prescaler switch position when I connected it
to a buffered Colpitts oscillator I am considering for a future
project. As this oscillator was operated from 76 MHz to 306 MHz, the
frequency meter read out faithfully in agreement with the oscilloscope,
which until I made this probe, was my only way of seeing signals in
this frequency range. I have not yet had an opportunity to use the
prescaler at higher frequencies.
One note on using it in the prescaler (F/10) postion: The MCT10280
prescaler was designed to be used on a printed circuit board, with ECL
amplitude signal sources permanently connected to it. There is ESD
protection on the input pins but the details of the EDS protection are
not given on the datasheet, so there is no wayto know to what extent
the input prtoection network shown in this schematic can protect the
MCT10280's input, other than to test several chips to failure. In leiu
of any volunteers to perform this testing with their own parts, I can
only recomend treating the input carefully when the F/F10 is in the
F/10 position. Make sure the frequency meter's ground is connected to
the signal source's ground. Be careful not to apply large voltage
transients to the input. Know what you are connecting the input to each
time.
Clearly, the cable between the preamp and the counter is
less than optimum. If the upper frequency limter were to be higher than
20 MHz or the cable longer, a technically better arrangement would be
to place a
buffer with a driving end termination in the probe body, drive a
shielded cable, and add preamp with a receiving end termination to the
frequency meter circuit board.
You might also
want to see these
related projects on www.projects.cappels.org:
A
Little More Serious Frequency
Meter
Serial Interface
for Truly MTC-C162DPLY-2N, 2 line X 16 char LCD display
Frequency
Meter and Pulse Generator
100 MHz
Modulated RF source
First posted in September, 2005. Update
20 Setp. 2005.
