Oscilloscope help

[Megahertz]

Hi, I am trying to replicate a waveform. I can see it on my oscilloscope, but I can't get the exact data I need just by looking at it as it is changing very fast. These are pulses from a uC of different periods. I tried the datalogging function on my scope, but I am unable to understand the data. Can someone please help me understand the information which is logged by my scope. I have attached the PDF of the logged data.

The pulses are not continues and only appear for few seconds and then there is a period of no signal and then again the pulses start.

If you want to watch how the pulses appear, I have also uploaded the video.


Thanks

[dhouston]

It's very difficult to judge given that you have not synced your scope but they appear to be RF or IR signals with a fairly wide start pulse followed by a series of shorter pulses which contain the data.

Perhaps you could get a better picture using your soundcard to record the signal. That gives you something you can playback in an editor (e.g. GoldWave). See http://davehouston.org/learn.htm

[Megahertz]

This is neither an IR nor an RF signal. I put the scope directly to the uC pin outputting this signal. This is the waveform which creates an alarm like noise when fed to the base of the npn transistor on which a coil and a piezo buzzer are connected on the collector side in parallel to each other. But since the pulses are changing so quick, visually I can only make out that the minimum pulse is 150uS and MAX is 250uS. These pulses are increasing from 150 to 250 in some number of steps. I tried replicating this in equal 10 increments of 10uS but it does not match as it does not produce the same sound effect. So I logged the data, but need someone to explain me what the two long rows mean. To me it seems some values (between 150uS - 250uS) are repeated more than once, I am hoping the correct interpretation of the logged data will make it more clear.

[kellyseye]

Looks like 'swept' signal.

The square wave frequency is being increased (swept) from x(Hz) to y(Hz) by a modulating waveform. It 'looks' like it's sweeping from around 2kHz to 3kHz by a low (250Hz???) signal.

Configure an output to a 2kHz square wave and arrange for it to increase by 4Hz every 4uS until it reaches 3kHz then repeat.

[Megahertz]

Hello kellyseye, I tried the following, but no sucess:

while 1
		Pulsout PortC.0,25
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,24
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,23 
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,22
@	NOP
@	NOP
@	NOP
@	NOP		
		Pulsout PortC.0,20 
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,19
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,18 
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,17
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,16 
@	NOP
@	NOP
@	NOP
@	NOP
		Pulsout PortC.0,15
@	NOP
@	NOP
@	NOP
@	NOP

Wend

[Megahertz]

Here I have recorded the sound produced by the buzzer. Please rename the attachment to "Memo.m4a" from ".pdf" as I could not upload it in its original format.

[Darrel Taylor]

Try this ...
You'll need to put the piezo circuit on a CCP pin.

Freq  VAR WORD

Main:
    FOR Freq = 3000 TO 4000 STEP 10    
        HPWM 1, 127, Freq
        PAUSE 1 
    NEXT Freq
GOTO Main

It should sound something like this ... http://support.melabs.com/DT/Piezo.wma

[Megahertz]

Thanks Darrel, I will try it tomorrow with 16f877A which I have. But I would like to ask if this task can be made possible using a smaller PIC without an HPWM, I would like to use 16F676 if possible as it is very easily available and where I am and if this takes off then it will be economical as well.

In the mean time I am attaching the datasheet of the Chinese uC which is controlling the base of the NPN transistor connected to the piezo & coil. The signal comes out from PIN 2, which to me seems just an I/O pin.

Thanks

[Darrel Taylor]

Yes, of course it can.
It's only 3-4 khz. You can generate that in software easily.

But see if it drives your piezo first.
I was assumming your piezo has a resonant frequency of ~3.5khz.

[Megahertz]

THANKS Darrel, it worked like a charm with HPWM. Any pointers as to how to do without it.

I tried FREQOUT

    FOR Freq = 3000 TO 4000 STEP 10    
        FREQOUT PortC.2,1, Freq
    NEXT Freq

but the results were not the same, not even near enough, the pulsing sound is not that prominent

[Darrel Taylor]

FREQOUT makes sine waves and must be filtered, which is not what you need.
Try this ...

PiezoPIN  VAR PORTC.2
TMR0IF    VAR INTCON.2
Width     VAR BYTE
OPTION_REG = %11010011 ; TMR0 1:16

Alarm:
    FOR Width = 161 TO 120 STEP -1
        WHILE !TMR0IF
            TOGGLE PiezoPIN
            PAUSEUS Width
            TOGGLE PiezoPIN
            PAUSEUS Width
        WEND
        TMR0IF = 0
    NEXT Width
GOTO Alarm

[Megahertz]

Thanks Darrel and everyone. It's done

[Megahertz]

Just out of curiosity, can this be achieved without using a microcontroller?

[kellyseye]

Commonly driven by a type 555 multi-function timer chip. Google '555 alarm schematic' and take your pick.
Also can be done using standard logic gates such as 2-NANDs in astable mode with another 2 providing the modulation therefore a single dual-input quad NAND package (4011 or 7400) will work. You can use NOR gates too.

[falingtrea]

You can also buy piezo devices that are already configured to output an alarm tone, aka piezo buzzers. Sonalert for one, comes to mind.