Frequency counter for PC, reviews comments etc

[neazoi]

Hi, I have found this circuit https://www.qsl.net/pa2ohh/pccnt.htm
It uses a very simple schematic and DOS software (not a problem for me) to count.
I do not know which of the two methods it uses for frequency counting but it has a good resolution and variable gate time on the software side, plus it can count to 80MHz whereas most microcontroller counters can count to 50MHz without a prescaler.
It also accepts an external reference frequency which is good because an OCXO or a GPSDO can be used.
I would like your comments on it, the pros and cons you can see.
What if I would like to make it sinewave compatible, shall I include a broadband amplifier at the input and just led the clipping diodes do the swuaring job, or better include a schmitt trigger?

 

[tanveerriaz]

chk this
https://electronicayciencia.blogspot.com/2011/07/frecuencimetro-para-el-pc.html

 

[neazoi]

chk this
https://electronicayciencia.blogspot.com/2011/07/frecuencimetro-para-el-pc.html

I have found many (better) frequency counters than the link you posted. My question was about the specific counter poster on #1

 

[Dan Mills]

It is a reciprocal counter using the PCs 8254 counter /timer hardware as the timebase for the count.

Basically it is just a prescaler chain tied to some pins on the parallel port, with a second reference chain that the software can measure to calibrate out the fact that the clock used for the 8254 on a classic PC was not exactly a paragon of accuracy.
It very much does have a prescaler, in fact that is really what it is in essence....

The variable 'gate' time is really I suspect a variable number of periods to average.

This is not going to work well in the presence of any sort of multitasking, and probably even a modern video card would give it fits (Never mind stuff like SMM mode that you cannot get away from on a modern PC).

Today , there are easier ways to skin this.

Regards, Dan.

 

[D.A.(Tony)Stewart]

The decision to make or buy a counter comes with knowing all the signal integrity issues you have expect for an input signal.

If there is any resonance or spikes during the transition interval, a good counter will have an adjustable gain, AC or DC coupling with a fixed hysteresis and high bandwidth, low noise.
There will be the ease of use for averaging and getting more resolution and less jitter.

This primitive design uses a binary counter to scale the signal down a few decades, which limits the useful range.

But if you have a PC to dedicate for this, it might work, after many hours of frustration for a primitive design.

just nothing comparable to a used $100 counter on EBAY.
**broken link removed**

 

[neazoi]

The decision to make or buy a counter comes with knowing all the signal integrity issues you have expect for an input signal.

If there is any resonance or spikes during the transition interval, a good counter will have an adjustable gain, AC or DC coupling with a fixed hysteresis and high bandwidth, low noise.
There will be the ease of use for averaging and getting more resolution and less jitter.

This primitive design uses a binary counter to scale the signal down a few decades, which limits the useful range.

But if you have a PC to dedicate for this, it might work, after many hours of frustration for a primitive design.

just nothing comparable to a used $100 counter on EBAY.
**broken link removed**

thanks, I wonder how accurate is the hp 54520a scope counter (or other dso), have you got any guess?

 

[betwixt]

I too have doubt about that design, at least the claims of accuracy as the only method to read the signals rely on polling the parallel port. It means the accuracy of the PC clock and processing speed are critical to it's accuracy. Note that when it says it must be run in DOS mode it does not mean DOS mode under Windows, it has to be the native DOS operating system and with no other time consuming tasks running. Virtualizing the parallel port will not work.

Most of the commercial counters are only accurate because they are calibrated to be so. The designs are optimized for stability but the underlying reference signal still has to be set against a better reference. I have a HP signal generator here which claims it can be set within 10Hz at 3.2GHz, I have never checked that claim but I'm pretty sure that even the slightest knock of a timmer capacitor somewhere inside would send it's accuracy into orbit.

Brian.