Frequency Measurements

MOR_AL · Posted: Fri Dec 04, 2020 3:31 pm Post subject: Frequency Measurements

"Sorry about my English"
I designed a frequency meter with PIC microcontroller in Assembly language. The accuracy of this frequency meter equaled that of the microcontroller crystal.
Now I am studying the Basic language of Bascom-AVR and I have a question.
Is there a way to obtain frequency measurements with the precision of the crystal used in the microcontroller with Bascom's Basic language?
MOR_AL

(BASCOM-AVR version : 2.0.8.2 , Latest : 2.0.8.3 )

albertsm · Joined: 09 Apr 2004 Posts: 5917 Location: Holland

i think your English is ok. But it is the lack of info.
it is like asking if you can do pwm.
sure you can. but the important info is missing.
which micro do you plan to use (there are variants in timers). what is the range of the frequency you want to measure.
which xtal you use, and what is the tolerance.
then you can use the search function of the forum.
i have seen shared code too.
as you used the pic you should know that freq. measurement is done using the timers.
it has little to do with bascom language.

MOR_AL · Posted: Fri Dec 04, 2020 11:42 pm Post subject:

Duval JP · Joined: 22 Jun 2004 Posts: 1161 Location: France

have a look at
https://www.mcselec.com/index2.php?option=com_forum&Itemid=59&page=viewtopic&t=14403&highlight=frequency+counter

EDC does very good sample
JP Wink

JC · Joined: 15 Dec 2007 Posts: 586 Location: Cleveland, OH

As stated, need more information.

How wide are the pulses you wish to measure, (capture)?

How close can they be, (what is the maximum frequency of the signal)?

How fast does the signal change frequency?

How fast do you need to update the measurement?

What is the output of the project, an LCD / GLCD for a human to read, or is it passing the information on to another micro/PC for real time data processing?

Many of the AVRs can measure a pulse width, or count pulses on a Timer/Counter, module directly within the hardware, without the micro "processing" the data.

The XmegaE5 and, IIRC, some of the newer generation XTiny micros also have a programmable logic unit within the micro.
With a wide range of input frequencies, a frequency counter likely will use a different measuring technique for low frequencies and for high frequencies.
The programmable logic unit can be used for a Divide By 2 (or perhaps more) without using an external chip.

Lastly, one of the great features of Bascom is that you can easily mix assembly language with the Basic language, if you found a reason to do so.
(It is rarely needed).
Likewise, although you can generally configure the various modules within the micro through Basic, you can also simply set up the registers yourself, directly, from within your Basic program.
So as mentioned above, the limitations of a project using Bascom and an AVR are really the limitations of the micro, not the language.

The "limitations" of an 8-bit micro running at 20 MHz are pretty similar regardless of the micro, (not trying to start a micro war here!).
Know, however, that the newer AVRs can run at 24 MHz, and the Xmegas can run at 32 MHz, so one has a lot more clock cycles with which to process the data between data result outputs!
Also, some of the micros have both the programmable logic unit, and the Event System, both of which can sometimes be very helpful in pushing the limits of the hardware.

Finally, for a frequency meter, which likely doesn't require the use of the analog modules, (ADC, DCA, AC, etc.), know that for a non-commercial product one might overclock the Xmega well beyound 32 MHz ! That becomes trivial with the internal clock PLL, you simply give it a new multiplier.

Good luck with your project!

JC

Edit: Typo

MOR_AL · Posted: Mon Dec 07, 2020 12:39 pm Post subject:

albertsm · Joined: 09 Apr 2004 Posts: 5917 Location: Holland

there are 2 methods to measure frequency. the simple one is to use a known time base and count for say 1 sec the number of pulses. the gate period you can vary depending on the input frequency.
the other method is where you count the internal periods of the clock. and now the input signal turns on/off the measurements.
since you need to calculate : f= 1.0 / period it is called reciprocal method.
but just as the simple method there are things to consider. with the simple gate method it is obvious that a measurement takes 1 sec. you can make the gate 100 ms to shorten that but that also changes the resolution.
for example when you make it 1 uS in order to be quick you can only grab input signals with a high frequency.
with the reciprocal method the internal clock is in most cases very high. say 10 MHz. when you count this high freq. signal in a 16 bit time it will overflow quick. of course depending on the input signal.
you could cascade timers.

MOR_AL · Posted: Tue Dec 08, 2020 2:44 pm Post subject: