Pulse Width Measurement
swampie777
Posts: 33
Mike Green was extremely helpful on my last project by showing some mis interpretations i had of the documentation. Ok I'm new to Parallax land. I did run into a bit of strangeness when i tried to upgrade the code to two channels.
The code follows.
This did not work. It would process the first repeat and send out the results which checked against an o-scope ( pulse width measurement), but the second repeat block was ignored completely.
If i commented out the first block, the second block was still ignored.
Any suggestions will be appreciated.
Thanks.
Swampie777
Post Edited (swampie777) : 5/17/2009 3:57:46 AM GMT
The code follows.
CON
_clkmode = xtal1 + pll16x
_xinfreq = 5_000_000
VAR
long stack[noparse][[/noparse]72]
long xxx1
long iii
long Idx
long xsum
long xxx2
long iii2
long Idx2
long xsum2
long iiii
long i5
long Cog
OBJ
Num : "Numbers"
TV : "TV_Terminal"
' uart : "FullDuplexSerial"
PUB Main
xxx1 := 0
iii := 0
xsum := 0
xxx2 := 0
iii2 := 0
xsum2 := 0
Num.Init 'Initialize Numbers
TV.Start(12) 'Start TV Terminal
Cog := cognew(@pulse_,@xxx1)
' TV.Str(Num.ToStr(Cog, Num#DDEC))
' TV.Out(13)
repeat while Idx < 5
xxx1 := -1
repeat while xxx1 == -1
xsum := xsum + xxx1
TV.Str(Num.ToStr(xxx1, Num#DDEC))
TV.Str(Num.ToStr(xsum/(Idx+1), Num#DDEC))
TV.Out(13)
Idx++
Cog := cognew(@pulsea_,@xxx2)
repeat while Idx2 < 5
xxx2 := -1
repeat while xxx2 == -1
xsum2 := xsum2 + xxx2
TV.Str(Num.ToStr(xxx2, Num#DDEC))
TV.Str(Num.ToStr(xsum2/(Idx2+1), Num#DDEC))
TV.Out(13)
Idx2++
DAT
org 0
pulse_ waitpeq zero,mask
:loop waitpne zero,mask
mov xx1, cnt
waitpeq zero,mask
mov yy1, cnt
sub yy1, xx1
wrlong yy1, PAR
jmp #:loop
pulsea_ waitpeq zero,mask2
:loopa waitpne zero,mask2
mov xx2, cnt
waitpeq zero,mask2
mov yy2, cnt
sub yy2, xx2
wrlong yy2, PAR
jmp #:loopa
zero long 0
mask long |<25
xx1 res 1
yy1 res 1
Mem res 1
lll long 0
mask2 long |<24
xx2 res 1
yy2 res 1
This did not work. It would process the first repeat and send out the results which checked against an o-scope ( pulse width measurement), but the second repeat block was ignored completely.
If i commented out the first block, the second block was still ignored.
Any suggestions will be appreciated.
Thanks.
Swampie777
Post Edited (swampie777) : 5/17/2009 3:57:46 AM GMT
Comments
Leon
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle
What one-liner Leon is talking about has nothing to do with your code per se. He is suggesting that, first, you click on the little pencil icon near the upper right corner of your post and then add a title to your Subject line so we all know what this post is about. Second, when he suggests you use Code Tags, he's talking about.... See the little boxes (under the upper emoticons), the little boxes that are labelled LIST, QUOTE, CODE, etc.? Before pasting in your code, click on CODE and paste your code after the code tag that will appear. Then click on the CODE* box and the browser should add a tag at the end of your code. The end result is that your code gets put into a neat little box that looks like this:
For me and my Mac, this only works for one line of code, but for PCs it should work out okay.
Do what Leon says and then maybe people who know more than I can come to your rescue.
Hope that helps,
Mark
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Watching the world pass me by, one photon at a time.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
My new unsecure propmod both 1x1 and full size arriving soon.
Need to upload large images or movies for use in the forum. you can do so at uploader.propmodule.com for free.
CON zero = $1F5 ' outb DAT org 0 pulse_ waitpeq zero,mask :loop waitpne zero,mask mov xx1, cnt waitpeq zero,mask mov yy1, cnt sub yy1, xx1 wrlong yy1, PAR jmp #:loop mask long |<25 xx1 res 1 yy1 res 1 DAT org 0 pulsea_ waitpeq zero,mask2 :loopa waitpne zero,mask2 mov xx2, cnt waitpeq zero,mask2 mov yy2, cnt sub yy2, xx2 wrlong yy2, PAR jmp #:loopa mask2 long |<24 xx2 res 1 yy2 res 1zero is used by both cogs and could either be redefined as local variable (global doesn't work due to the second ORG getting in the way) or we can use a register which we know is empty (outb in this case).
It didn't work.
If you change··xxx2· xsum2.. to xxx1 and xsum1 ( change everything but the mask and pulsea_) then it works.
I thought when you did a Cog·:=·cognew(@pulsea_,@xxx2) that a new object was created with the lable segment of machine code pointed to ( (@pulsea_).
Is this not how you generate different assembly objects for each different cog?
Thanks
DAT [color=red]org 0[/color] [color=#ff0000] ' this resets the adress pointers which are used in PASM. ' So pules_ now points to long-adress 0, :loop now points to long adress 1 [/color] [color=red]' Whenever you use :loop in you PASM it will be replaced by 1[/color] pulse_ waitpeq zero,mask :loop waitpne zero,mask mov xx1, cnt waitpeq zero,mask mov yy1, cnt sub yy1, xx1 wrlong yy1, PAR jmp #:loop [color=red]' here we are at adress 8 for pulsea_ and 9 for :loopa [/color]pulsea_ waitpeq zero,mask2 :loopa waitpne zero,mask2 mov xx2, cnt waitpeq zero,mask2 mov yy2, cnt sub yy2, xx2 wrlong yy2, PAR jmp #:loopa [color=red]' this says: jump to adress 9 [/color] zero long 0 mask long |<25 [color=red]xx1 res 1 ' Here you have a problem because these Labels will have the same adress as the following longs yy1 res 1 ' xx1 = lll and yy1 = mask2 Mem res 1 [/color]lll long 0 mask2 long |<24 xx2 res 1 yy2 res 1What you do here
·Cog·:=·cognew(@pulsea_,@xxx2)
is, that you load a part of PASM code to a COG. This PASM code will always start at COG-RAM adress 0. But your compiled code assumes that it is starting at adress 8. So when executed the jmp #:loopa will simply jump to the wrong place.
Let's have a look at Kuroneko's code:
CON [color=red] zero = $1F5 ' outb This is a trick to spare a long for the ZERO you had in your code. The propeller does not have a B-Port, but this register[/color] [color=red] ' is available in RAM. As it's always initialized with 0 you can use it as ZERO-variable.[/color] [color=red] ' But in your case you wouldn't even need that [/color] DAT org 0 [s]pulse_ waitpeq zero,mask :loop waitpne zero,mask [/s][color=red]pulse_ waitpne mask,mask ' this will do the same job and you can use OUTB for something else :loop waitpeq mask,mask [/color] mov xx1, cnt waitpeq zero,mask mov yy1, cnt sub yy1, xx1 wrlong yy1, PAR jmp #:loop mask long |<25 xx1 res 1 yy1 res 1 [color=red]DAT org 0 ' this resets the internal PASM program-adress-counter, so you can do a cognew(@pulsea_,..)[/color] [color=#ff0000] ' that's OK as such, but you have nearly the same code as for pulse_, so it's a waste of[/color] [color=#ff0000] ' HUB-RAM. Once your pulse_ is loaded into the COG you can do with it whatever you want in[/color] [color=#ff0000] ' HUB-RAM. So it's easy to reuse pulse_.[/color] [s]pulsea_ waitpeq zero,mask2 :loopa waitpne zero,mask2 mov xx2, cnt waitpeq zero,mask2 mov yy2, cnt sub yy2, xx2 wrlong yy2, PAR jmp #:loopa mask2 long |<24 xx2 res 1 yy2 res 1[/s]pulse_··················waitpeq···zero,mask
:loop···················waitpne···zero,mask
pulse_··················waitpne···mask,mask····· ' this will do the same job and you can use OUTB for something else
:loop ··················waitpeq···mask,mask
························mov·······xx1,·cnt
························waitpeq···zero,mask
[noparse][[/noparse]/quote]
The reason I used zero·-- that is the value to compare the pin to so that the extent of the square wave pulse can be measured.
You wait while the pin is zero to catch the start of the pulse. Then you capture the cnt. After the pin goes back to zero, you get cnt again and subtract to get the count width of
the pulse. If you use the mask as both arguements, you'll never see the zero level.
·······
Did I mis-read you?
waitpeq zero, mask
can be translated to
repeat while INA & mask == 0
It waits as long as the Pin is 0
And now have a look at
waitpne mask, mask
can be translated to
repeat while ina & mask != mask
It waits as long as the Pin is not 1 ... and for a single pin not 1 is equal to 0
So, both instructions do the same but in my version you don't need an additional long to compare with.