PASM counter setup question

turbosupra

I'm trying to set up pin 25 to start as a low output and configured for a counter type of DUTY single-ended mode. Then I'd like to adjust the counter pulse width (and in turn average voltage) by setting frqa. I've gotten this working in spin, but cannot get it working in PASM. Pin9 is also an output and doing something different.

dira binary output on the pst yields
00000010000000000000001000000000

outa binary output on the pst yields
00000000000000000000001000000000

ctra binary output on the pst yields
00011010000000000000000000000000

frqa binary output on the pst yields
01111101001010110111010100000000 (not posX2, but still 1050000000 which should register)


But the pin itself doesn't generate any pwm voltage. What am I doing wrong?

                        rdlong  mafPin, mafPinPtr
                        mov     mafPinMask, #1                     ' create mask from pin #
                        shl     mafPinMask, mafPin
                        or      dira, mafPinMask        ' sets direction to 1 for an output
                        andn    outa, mafPinMask        ' sets pin to 0 for a low starting output  (or for high)
                        mov     ctra, mafPinCfg         ' counter to single-ended mode, ctra := %00110 << 26 +    CurrentTripSetPin
                        or      ctra, mafPinMask        ' ctra to use pin 25

                        mov     frqa, posX2

Here is the spin code I'm mimicking

PUB Main

  pst.start(115200)
  dira[CurrentTripSetPin]~~
  outa[CurrentTripSetPin]~
  ctrb := %00110 << 26 +    CurrentTripSetPin
  'frqx := ( centi_volts * ( POSX / 3.3 / 100) ) << 1
  skew := (POSX / 330)          ' the equivalent of 1 single centiVolt
  centiVolts := 27
  '       
  repeat 
  ''''''   test values, produces voltages as shown
    'frqb := -1                         '3.265V***** 4.294B    ' 3.21/3.25
    'frqb := -2_147483646     '1.655V***** 2.147B              ' 1.57/1.62
    'frqb := 2_147483647       '1.655V***** 2.147B             ' 1.57/1.62
    'frqb := 0                           '0.000V                ' 0.00/0.00
    frqb := (((centiVolts * ((POSX / 33) / 10)) + (4 * skew)) << 1)
    pst.dec(frqb)
    pst.char(13)
    pst.dec(centiVolts)
    pst.char(13)
    pst.char(13)
    'centiVolts += 1
    waitcnt(clkfreq + cnt)  

Tracy Allen

I think you want,

or ctra, mafPin

not

or ctra, mafPinMask

turbosupra

You're the man Tracy! Thank you.

I knew it was something stupid ... ugh!

So can you tell me what the difference is? The values are the same, so it's how they were applied. One is the 25th bit being set to "1" and the other is a cumulative numeric value of 25. Are bits 5..0 summed to get the A pin? Is that what I misunderstood about the register?



ctra for mafPinMask
00011010000000000000000000000000


ctra for mafPin
00011000000000000000000000011001



Table 2-5: CTRA and CTRB Registers
31 -
30..26 CTRMODE
25..23 PLLDIV
22..15 -
14..9 BPIN
8..6 -
5..0 APIN

turbosupra

And to calculate which numeric value I need to get a corresponding duty cycle (and in turn analog voltage), I should use the formula explained in this post?


FRQx = Fout / CLKFREQ * 2^32
FRQx = 2^32 / N

http://forums.parallax.com/showthread.php?87905-The-counters&highlight=frqx+pasm

turbosupra

In case someone searches and finds this (I have a 10uF cap and a 100k resistor making up my R/C network/DAC and to give an average of the output of the props pin, will try a 10k/1uFcap later on)

You can have up to 2^32 different pulse width output settings. I did a voltage range (5v) * 100 divided by 2^32 to get a centiVolt value and that gave me 13015052
or 2^32/(3.3*100) = 13015052

After that (this isn't quite exact because of a slight skew, but it is close) I can just do any value between 0 and 330 (which is 3.3v * 100) multiplied by "centiVolt" to get a corresponding average voltage value.

So 300 = 2.95v and 320 = 3.15v (notice there is an approximate .05 skew)


Next I will probably compute a millivolt value for added accuracy, 2^32/(3.3*1000) = 1301505

Tracy Allen

1. Some pasm instructions such as ctra or dira take the specifc pin number and affect that one pin. Sometimes two pins for ctra if it calls for both an apin and a bpin. Other pasm instructions such as muxc use a mask, and that allows them to affect any number of pins at once. You just have to remember for each instruction which type it is.

2. In your previous math, I don't see why you are bringing the factor of 5V into the math. I think you meant to write 3.3V? If you have that simple RC filter on the Prop pin output, and you want a centivolt multiplier, it is
....frqa = 2^32 * 0.01/3.3 = 13015052
as you have. In general, if you want an output voltage of Vx, then calculate,
....frqa = 2^32 * Vx / 3.3 = ????

3. In reference to the 0.05V skew you observed, the 100kΩ resistor in your RC filter is going to form a voltage divider with the input impedance of your voltmeter, and that will reduce the measured voltage proportionally. The RC values of 100kΩ with 10µF are probably higher than they need to be. Try 10kΩ with 1µF. There are tradeoffs of smoothing, speed and jitter, but even smaller values are suitable for most purposes.

turbosupra

Tracy,

That was a typo, it is supposed to be 3.3v thank you for catching that. (I'll edit the post now)

I will change my R/C components tonight and see how that works, the voltage divider makes sense. Thanks for confirming my suspicions about the register bit locations also, I will have to be more careful when reading the prop manual to make sure I don't make that mistake twice!