pic16f88 frequency counter triples result

[truffaldino]

I wrote a frequency counter program in PICBASC PRO for 16F88. This is a pretty standard stuff: timer1 is an asynchronous counter and timer0 is a clock.

When I run this counter in Proteus, everything is OK if I feed B6 timer1 pin with the digital clock signal (it works up to 100 MHZ).
But when the signal is not digital clock, even if it is suare 50%duty cycle with raise/fall time 1 picosecond, it triples the frequency. It shows correst frequency only if rise/fall time is 0.

What could cause this problem? Is it software, configuration or simulator error?

 

[FvM]

it works up to 100 MHZ

Makes clear that the simulation hasn't much to do with a real PIC. Maximum asynchronous input frequency is 16 MHz.

But when the signal is not digital clock, even if it is square 50%duty cycle with raise/fall time 1 picosecond, it triples the frequency. It shows correst frequency only if rise/fall time is 0.

You actually mean picoseconds? And observe a difference between 1 ps and 0 ps rise time? I wonder if Proteus can simulate superluminal phenomena.

 

[truffaldino]

OK, we cannot trust simulator.

I have another question.

I put two microcontroller circuit on the solderless breadboard to test the counter on low frequencies (just 5-10 Hz). First microcontroller is the pulse generator, while the second one is a frequency counter.

It shows zero frequency, while in Proteus everything works fine. Could it be caused by solderless breadboard parasitic effects?

 

[FvM]

Could it be caused by solderless breadboard parasitic effects?

Possible but unlikely. I would rather expect a coding or wiring error. You would check if either the generator or the counter are failing.

 

[truffaldino]

It was indeed a small coding error unnoticed by simulator. I guess all these problems with simulation is caused by bugs in SPICE model of PIC.

 

[truffaldino]

I have another problem with the counter: I have put a jfet buffer and npn self-biased transistor as a preamplifier before counter. Everything works fine up to about 12MHZ when I use ordinary 2N3904 transistor as preamplfier, but when I try to use high-frequency transistors, such as BF199 or PN3563 counter shows zero result.

As for BF199, I run a simulation in LTSPICE and everything is going excellent there. What could be cause of such behavior: stray capactance of solderless breadboard?

 

[Aussie Susan]

At 12MHZ stray capacitance could well be the issue.
However use a scope to see the actual signal traces and in particular voltage and frequency to make sure they are acceptable.
If there is capacitance then the transitions may be slowed.
Susan