Frequency counter modification for fewer chips

[neazoi]

I have found this vintage frequency counter (I like it because it does not require an mcu) and I want to use it for HF with at least the 100Hz option that is mentioned in the article.

1. Can this be modified somehow to display to 10Hz (cutting off the MHz symbols)

2. Can the number of chips be reduced (at least in the 100Hz option)?
for example the master oscillator could be made discrete.

3. Are you aware of any simpler MCU-free counter for HF (up to 30MHz) that is able to count to 100Hz?

Thank you.

- - - Updated - - -

4. The refresh rate of 2 seconds is quite annoying. can I do something to reduce it to say 1 sec, or even 1/10th of a second?

 

[BradtheRad]

I built my homebrew 6-digit frequency counter around two 14553 IC's (similar to the 4553's in the article). It updates each second. Readout 1 to 999,999 Hz. I've had it 30 years.

Timebase is from a 555 IC (10 Hz divided by 10 via 4017 IC). In hindsight I should have made a stable timebase by using a crystal, 32 KHz divided down to 1 Hz via a 4020 counter IC.

You might get away with making updates every 1/10 sec, thus having a range 10 Hz to 10 MHz (by adding a zero to the right of 1-999999).

Or else if the IC's cannot keep pace, then you divide the input by 10 to obtain range 10 Hz to 10 MHz. (So you add a zero to the right of 1-999999.) Updates every sec.

Then dividing by 10 twice is the same as dividing by 100. It gives you 100 Hz to 100 MHz. ( (So you add two zeroes to the right of 1-999999.) Updates every sec.

- - - Updated - - -

I followed schematics in Forrest Mims' Engineer's Notebook. Can be viewed on Google-Books.

https://books.google.com/books?id=a_B4dCFFL2oC&pg=PA32&lpg=PA32

 

[KlausST]

Hi,

I see you don't want to use an MCU. This is how I'd do it because:
* it is most flexible
* it can handle disply mutiplexing
* it can control frequency dividers while adding the leading zeroes accordingly
* it can handle various update rates
* it can be controlled by a PC
* it can send the measurement value to a PC
* it can use "counter mode" for higher frequencies
* or "time measurement" mode for high resolution (milli-Hz) at lower frequencies.
* automatic range select
* low part count
* can drive low power LCDs for low power battery operation.
* very accurate because it depends on Xtal frequency. May be fine tuned by software.
...
The drawback is the software effort. But this can be made step by step. I assume there is plenty existing code in the internet.

********
All other solutions will have reduced functionality.
I recommend to choose your requirements first.

Other solutions I see:
* FPGA
* PLD
* ready to buy frequency counter ICs like ICM7216
* ready to buy frequency counter modules
* ready to buy frequency counters (complete measurement device), new or used.
* multiple ICs solutions
...
More and more I use the interface to a PC. It enables to automatically generate reports. Data input to excel or other applications for easy chart generation. Long time frequency tracking. Archivating measurement data. Remote access via internet.

Klaus

 

[neazoi]

I built my homebrew 6-digit frequency counter around two 14553 IC's (similar to the 4553's in the article). It updates each second. Readout 1 to 999,999 Hz. I've had it 30 years.

Timebase is from a 555 IC (10 Hz divided by 10 via 4017 IC). In hindsight I should have made a stable timebase by using a crystal, 32 KHz divided down to 1 Hz via a 4020 counter IC.

You might get away with making updates every 1/10 sec, thus having a range 10 Hz to 10 MHz (by adding a zero to the right of 1-999999).

Or else if the IC's cannot keep pace, then you divide the input by 10 to obtain range 10 Hz to 10 MHz. (So you add a zero to the right of 1-999999.) Updates every sec.

Then dividing by 10 twice is the same as dividing by 100. It gives you 100 Hz to 100 MHz. ( (So you add two zeroes to the right of 1-999999.) Updates every sec.

- - - Updated - - -

I followed schematics in Forrest Mims' Engineer's Notebook. Can be viewed on Google-Books.

https://books.google.com/books?id=a_B4dCFFL2oC&pg=PA32&lpg=PA32

I have just realized that one of the 74hc390 could be removed. A half of it, is not used anyway. But the second half could also be removed according to the article. See the page 5 in the possible modifications section. Since I need the 100Hz option the second half of the 74hc390 could be removed. That is one chip less. So now it is seven chips, not bad.

I was thinking of this 4553 and maybe you could help on this. Maybe there is another similar chip that has the functionality of the 4511 embedded into it (or maybe not)?

Also I see the functionality of the 74hc132 is important. but this is just gates, so maybe I could replace this with a few discrete diode-transistor logic?

- - - Updated - - -

Hi,

I see you don't want to use an MCU. This is how I'd do it because:
* it is most flexible
* it can handle disply mutiplexing
* it can control frequency dividers while adding the leading zeroes accordingly
* it can handle various update rates
* it can be controlled by a PC
* it can send the measurement value to a PC
* it can use "counter mode" for higher frequencies
* or "time measurement" mode for high resolution (milli-Hz) at lower frequencies.
* automatic range select
* low part count
* can drive low power LCDs for low power battery operation.
* very accurate because it depends on Xtal frequency. May be fine tuned by software.
...
The drawback is the software effort. But this can be made step by step. I assume there is plenty existing code in the internet.

********
All other solutions will have reduced functionality.
I recommend to choose your requirements first.

Other solutions I see:
* FPGA
* PLD
* ready to buy frequency counter ICs like ICM7216
* ready to buy frequency counter modules
* ready to buy frequency counters (complete measurement device), new or used.
* multiple ICs solutions
...
More and more I use the interface to a PC. It enables to automatically generate reports. Data input to excel or other applications for easy chart generation. Long time frequency tracking. Archivating measurement data. Remote access via internet.

Klaus

I have already built counters using PIC micros (others projects) and yes they provide great features in very few components.

However, the microcontroller requires a programmer hardware and software, in order to be programmed. A PC is also required for the programming operations. This might look common practice to you but many of the homebrewers do not know how or don't have the skills to write a program for the MCU neither the time to learn MCU programming. Thus, they rely on programs others have built and they cannot alter their operation to their needs. There are several other disadvantages one may think. For example, I care my projects could be built even 40 years later, by someone reading a "forgotten" article published in the past on a magazine. This is something I also do, looking at old magazines to find simple circuits or block diagrams to get ideas. Even if the transistors do not exist anymore, most of the times they have equivalents and sometimes they can be replaced by general purpose modern transistors. Hopefully the TTL/CMOS series chips will exist after many years, but specific micros won't. Another reason may be the availability, for example in my local town, electronics shops do not have micros and you have to order them, but they mostly have TTL/CMOS series ICs.
You get the feeling.

 

[betwixt]

The device with functionality of the 4553 and 4511 and 74HC132 is the Intersil IL7216 as Klaus suggested. All it needs is a crystal and a few capacitors, it drives 8 digits of LED directly. However, it has been obsolete for a number of years. I have one in a frequency counter here that covers up to 1.5GHz by using a prescaler.

Regarding the update rate, you have to consider how a frequency counter works: The count is cleared, then for a fixed time "The gating period", it counts incoming pulses. At the end of that time it transfers the result to the display then repeats. To get faster display updates you have to reduce the gate time so the count finishes sooner and can pass the result to the display quicker. For higher resolution, you have to count every pulse which might involve a longer gate period and hence slower display update.

There are some tricks to measure low frequencies, typically reversing the gate clock and input signals so you count clocks per signal cycle instead of the other way around. For very low frequencies, (<1Hz) some counters use PLL circuits to multiply the frequency to make it quicker to read but there are inherent difficulties in doing that so it is seldom used. PLL multiplication can increase resolution but it doesn't improve update rate.

Please bear in mind that all of the requirements you have can be met with a single inexpensive MCU. Search for PIC16F628 frequency counters and you will see many single chip designs that count accurately up to 50MHz or higher. Total build cost could be less than about 20 Euros including the LCD display. That is much cheaper than buying the individual ICs alone.

Brian.

 

[neazoi]

Regarding the update rate, you have to consider how a frequency counter works: The count is cleared, then for a fixed time "The gating period", it counts incoming pulses. At the end of that time it transfers the result to the display then repeats. To get faster display updates you have to reduce the gate time so the count finishes sooner and can pass the result to the display quicker. For higher resolution, you have to count every pulse which might involve a longer gate period and hence slower display update.

Brian.

What if I increase the clock rate from 0.5Hz to... ?
Will it give faster refresh rate?

In my PIC based counter the gate time is 0.1 second which gives resolution of 10Hz.

 

[KlausST]

Hi,

I have already built counters using PIC micros (others projects) and yes they provide great features in very few components.

However, the microcontroller requires a programmer hardware and software, in order to be programmed. A PC is also required for the programming operations. This might look common practice to you but many of the homebrewers do not know how or don't have the skills to write a program for the MCU neither the time to learn MCU programming. Thus, they rely on programs others have built and they cannot alter their operation to their needs. There are several other disadvantages one may think. For example, I care my projects could be built even 40 years later, by someone reading a "forgotten" article published in the past on a magazine. This is something I also do, looking at old magazines to find simple circuits or block diagrams to get ideas. Even if the transistors do not exist anymore, most of the times they have equivalents and sometimes they can be replaced by general purpose modern transistors. Hopefully the TTL/CMOS series chips will exist after many years, but specific micros won't. Another reason may be the availability, for example in my local town, electronics shops do not have micros and you have to order them, but they mostly have TTL/CMOS series ICs.
You get the feeling.

Ok. Makes sense.
Didn´t know that´s for public and homebrewers.
Then PLD and FPGA is no option, too.

Maybe useful, maybe not:
In times of PCs: A bare counter which sends it´s (binary) value in UART style. Via RS232, via UART-USB bridge, via Bluetooth, via WiFi...
Benefit: needs no display (which could be a drawback, too)...

Klaus

 

[betwixt]

What if I increase the clock rate from 0.5Hz to... ?
Will it give faster refresh rate?

In my PIC based counter the gate time is 0.1 second which gives resolution of 10Hz.

I'm not sure which clock you mean but if you deviate from 1 or a multiple of 10, the resulting digits will be multiplied or divided accordingly. For example, if you change from 0.5Hz to 1Hz, you half the gate period and the display shows half the frequency. This is sometimes useful if you use a prescaler (divide by 64 for example) but not useful otherwise. The frequency count cannot be updated until the counting period has ended, otherwise you only see a partial count.

The resolution you have is correct, if you gated for one second, you get the exact cycle count in one second so the resolution is to 1Hz. Making the gate period ten times shorter means you lose the last digit so the resolution becomes 10Hz and so on.

I'm not sure the argument about MCUs going obsolete is valid, try getting hold of some older TTL or CMOS devices these days, it is just as difficult. At least with MCUs you can generally replace them with a 'better' one but retain the original functionality. I've been using PICs for more than 25 years and never had problems with getting originals or pin-compatible substitutes. I agree the process of getting the program into the silicon needs extra equipment but I can even write, simulate and debug for PICs on my mobile phone while on the move so the technology is easily available and easy to use.

Brian.