Code is randomly not working

qsuscs

Hello,

this afternoon, I wanted to check which frequencies I can hear, but I wanted to do this by myself. So I wrote a bit of spin which should give me the tones from 10Hz to 20kHz in different intervals.
It works, sometimes, but not every time. There are two problems:

1. It stops at a random point, there is no sound anymore but the frequencies are printed all at once
2. The frequencies are off too much

I've put the code here: https://gist.github.com/4504149 Feel free to fork, I haven't specified a license but WTFPL should do (if it is compatibele with the pst-license)

BTW, I use a 5MHz crystal (two legs), but I also have a 4.9152MHz oscillator and some other microcontrollers that I could use as a clock.

qsuscs

EDIT: I also tried different frequencies (i.e. pll1x instead of pll16x), it stopped later, but it stopped nevertheless, and the tone frequencies were still off.

Phil Pilgrim (PhiPi)

It's stopping because of this statement:

repeat until cnt > counter+clkfreq*2

That's because cnt doesn't keep going up forever; it overflows from $7FFF_FFFF (biggest positive number) to $8000_0000 (smallest negative number). Do this instead:

reepat until cnt - counter > clkfreq * 2

Your frequencies are off (too low) becasue of this statement:

waitcnt(clkfreq/i*2+cnt)

It doesn't account for interpreter overhead. Also, you need parens in the denominator. Better would be this:

waitcnt(k += clkfreq / (i * 2))

where k is initialized to cnt at the beginning of the loop.

I'm not certain, but Spin may be too slow to reach 20 kHz this way. But you can use one of the counters to generate your output without having to resort to PASM.

-Phil

qsuscs

It works better, but unfortunately not good enough.
But the counters sound interesting, how do I use them? AFAIK the Propeller doesn't have interrupts, this is how the counters are used in the world of AVR.

Just found that: http://www.parallax.com/tabid/780/Default.aspx. I think it’ll be a strange feeling when I do something with AVR again, one core, interrupts, …

StefanL38

Hi qsuscs,

for example the object http://obex.parallax.com/objects/422/ creates frequencies.
So take a look into the OBEX which frequency-object suits best to your needs

best regards
Stefan

JonnyMac

You could easily use the a Propeller counter to generate the frequency, in fact, you could have two at once. This is a method I have in my template code for doing that (and yes, you can change frequency on the fly, or stop output by setting the frequency to 0).

pub set_freq(ctrx, px, fx)

'' Sets ctrx to frequency fx on pin px (NCO/SE mode)
'' -- fx in hz
'' -- use fx of 0 to stop counter that is running

  if (fx > 0)                             
    fx := ($8000_0000 / (clkfreq / fx)) << 1            ' convert freq for NCO mode    
    case ctrx
      "a", "A":
        ctra := ((%00100) << 26) | px                   ' configure ctra for NCO on pin
        frqa := fx                                      ' set frequency
        dira[px] := 1                                   ' make pin an output
     
      "b", "B":                         
        ctrb := ((%00100) << 26) | px  
        frqb := fx                    
        dira[px] := 1

  else
    case ctrx
      "a", "A":
        ctra := 0                                       ' disable counter
        outa[px] := 0                                   ' clear pin/driver 
        dira[px] := 0                                  
     
      "b", "B":                         
        ctrb := 0 
        outa[px] := 0  
        dira[px] := 0

kuroneko

JonnyMac wrote: »

  fx := ($8000_0000 / (clkfreq / fx)) << 1            ' convert freq for NCO mode    

You might want to change that to

fx := ($8000_0000 / [COLOR="#FF0000"]-[/COLOR](clkfreq / fx)) << 1

OK, not critical, you just get inverted phase. Must be the weather ... it's snowing here which happens - if you're lucky - once a year, for a day.

prof_braino

Please post you code when you get you hearing test working. Many of us are curious if we are also losing our hearing, or just going senile.

Also, are you adding very high frequencies? I have always imagined I can "hear" frequencies that are supposed to be near ultra sonic, from working with CRTs, marginal TVs, etc. If convenient, please include a couple brands above what were are supposed to be able to detect.

qsuscs

prof_braino wrote: »

Please post you code when you get you hearing test working. Many of us are curious if we are also losing our hearing, or just going senile.

Also, are you adding very high frequencies? I have always imagined I can "hear" frequencies that are supposed to be near ultra sonic, from working with CRTs, marginal TVs, etc. If convenient, please include a couple brands above what were are supposed to be able to detect.

I will always push the code to GitHub Gist. If you just want to test your ears, there are hundreds of music files everywhere, one minute of google should do. But I thought that it's boring this way.

I planned to prompt a frequency (PS/2+TV or serial) after all the frequencies were played. There is a sticker on my CRT that says fH: 27-68kHz, I don't think you are able to hear that, but, I don't know, maybe you are, or the CRT emmits other frequencies.

qsuscs

After a long time, I finally sat down and read a bit and changed the code. Here you go: https://gist.github.com/qsuscs/4504149