Inspired by the work of Onno PA2OHH and Hans G0UPL
this is my variation of the 2 chip LED display frequency counter.

The circuitry and operation is very simple, the only trade-off being that
the display must be interpreted from the state of 6 adjacent LEDs:
the frequency count being the aritmetic "sum" of the illuminated LED's.
This you must do mentally or with pen & paper, but your eye/brain
adapts suprisingly quickly to this presentation.

A 4060 xtal osc/14 stage ripple counter, which gates on/off a 4040
12 stage ripple counter which is used to count/store the incoming signal frequency.
Half of each cycle is used for counting the incoming signal.
Other half displays the frequency "count" on the LEDs
With a 4MHz xtal reference (4060 on-chip oscillator):
Mhz position: 15625 Hz sample 64uS period - 32uS count, 32uS display
Khz position: 244.1 Hz sample 4.096 mS period - 2.048 mS count, 2.048 mS display
No display "flicker" is seen.
During the sampling interval; input transistor-preamp/buffer is switched/gated on
and 4040 ripple counts incoming signal-frequency.
4060 rising edge thru 100pF capacitor spike resets the 4040 to zero, just prior to counting.
All LEDs will be dark because their cathodes are at +5v
During display interval; input transistor is gated off,
and LEDs can illuminate (cathodes now grounded thru 4060) to display count that is stored in the 4040
Cycle then repeats
The 6 LED's indicate a frequency value of 8, 4, 2, 1, 0.5, 0.25 (MHz switch pos'n)
and 125, 62, 31, 16, 8, 4 (KHz switch pos'n) See diagram.
I used small (SMD) hi-intensity blue LEDs; unsoldered from LED ribbon strip light.
These are well under-run but still bright whilst drawing only 0.5mA each (4K7 limit resistor)
I have used the 74HC 4xxx series (high speed CMOS) family logic IC's
The useable maximum frequency limit (at 5V rail) is 40 MHz.
The LED display upper limit capability is just< 16MHz (8+4+2+1+0.5 ....etc)
In the MHz switch pos'n the counter displays the most significant 6 highest 'bits' of the count.
Accuracy here would then be 1 part in 64 (about 1.5%)
In the KHz position the counter is "over-clocked" 64 times (15625 ÷ 244)
however provided the input signal is stable we read the truncated frequency count accurate to 4KHz.
Arithmetic addition of the two readings will give a figure of accuracy of 1 bit in 212 or 1 part in 4096 ( 0.02%)
So you can see; that this very simple idea is indeed quite usefull!

See also: "VK6FH-freq' counter"

Test mode: counting action of 4040 ripple counter: