Gate Timer

Brian_B

Hi All,
I'm working on a dual gate timer and have it somewhat working. This code works except that I can't figure out how to get hundredths of seconds to display properly. Any thoughts would be much appreciated.

Thanks Brian

JonnyMac

I have no idea what you're doing, but I think what I've attached is a cleaner way to do it.

Unless you know your gate time is going to be very short, avoid using calls to the cnt register. In this version the ADC reader cog also increments a millis register that is used to provide timing references -- same way you're doing.

No need for floating point, either.

Brian_B

Jon,
It looks good ,but adc seems like it's not working.

Thanks'
Brian

Brian_B

Jon,
If I comment out "waitcnt (t +=MS_001)" it runs.


Thanks'
Brian

JonnyMac

You can't simply remove the waitcnt as that is controlling your milliseconds timing. I did a quick test and it turns out in Spin, that ADC is stupid slow (yes, I know I wrote the code, but I never measured until today). It takes about 1.5ms to read an ADC channel so you can change your milliseconds resolution to 2 like this:

pri gates | t, ch

'' Scan analog channels every 2 milliseconds
'' -- update global millis value

  adc.start(ADC_CS, ADC_CLK, ADC_DI, ADC_DO)                     ' start adc

  t := cnt                                                       ' sync time point
  repeat
    repeat ch from 0 to 3                                        ' loop through channels
      long[@chan0][ch] := adc.read(ch)                           ' read adc
      waitcnt(t += (MS_001 << 1))                                ' allow 2ms window to expire
      millis += 2                                                ' update millis register

This is still 5x the desired resolution you want to display, so I think you'll be fine.

Sorry for the confusion.

If you know for a fact that your timing is always going to be under 26 seconds, you can get to clock ticks pretty easy -- this is what I did to determine the time to read an ADC channel.

elapsed := -cnt
{ code under timing test }
elapsed += cnt - 544

The "- 544" removes the Spin overhead (if there is no test code between the timing lines the result will be 0). In my program I have constants for microseconds and milliseconds that I use for converting clock ticks to human-friendly time units.

Finally, you now have an object called jm_time.spin that can do timing for you. You can do something like this

time.start
{ code under timing }
elapsed := time.millis

Because I do so many timing-oriented things, I developed the object. The only caveat with it is you must call it periodically (no less than every 53 seconds at 80MHz). If you had a very long timing event you could do it like this

repeat until (start_condition)
time.start
repeat
  elapsed := time.mark
until (stop_condition)

Brian_B

Jon,
Works SUPER ! It has a little glitch if you try it really fast , but that should never be a proplem.

Here are some photo's of my 3d printed gates and homemade air track.

Thanks' a million,
Brian

JonnyMac

Cool!

I'm curious as to why you're using the ADC for inputs. What is the output of the device you're reading? Based on your threshold, the sensor seems to be pulled up and then goes low. If this is the case and it drops below 1.6v, you can use a simple digital input.

Brian_B

Jon,
That's a good question. When I started messing with this IR sensor (BPW41N) I couldn't get it to drop below 4.8 volts. Now that I have it wired right it goes from 5 volts to .5 volts , so yes I could just use a simple digital input.

It's funny that once you start heading one way with a problem you overlook the simple solution

Thanks'
Brian

JonnyMac

If you use a couple pins as inputs, you no longer need the background cog (remove it). And as I know you're doing very short-term events, I would modify the main code like this to display gate-to-gate time with 1ms resolution.

pub main | elapsed 

  setup

  repeat
    repeat while ina[START_PIN]
    elapsed := -cnt                                              ' start timing
    outa[LED_26] := IS_ON
     
    repeat while ina[STOP_PIN]
    elapsed += cnt - 544                                         ' stop timing
    outa[LED_26] := IS_OFF

    elapsed /= MS_001                                            ' convert ticks to ms

    term.str(string("Elapsed time: "))                           ' report 
    term.dec(elapsed / 1000)
    term.tx(".")
    term.rjdec(elapsed // 1000, 3, "0")
    term.tx(term#CR)

Now that I have it wired right it goes from 5 volts to .5 volts , so yes I could just use a simple digital input.

To be on the safe side, put a 3.9K (or higher) resistor between the sensor output and the pin.

Brian_B

Jon,
Now I remember why I went with a ADC. During daylight hours the beam break only drops the voltage to 4.5 volts. If wrap my hand totally around the sensor it will drop to .5 volts.


Thanks'
Brian

Brian_B

Jon,
I finally got the logic code to work in sunlight. I printed some "sun glasses" for my gates , just some pinhole covers on the infrared sensors.

Thanks' for the help !

Brian