Can P Basic Math be done on the fly with RC TIME command this is for Tracy All

sam_sam_sam · 2009-07-31 21:48

Hi EveryOne

This is a project that I have been ·ask to do·by employer

As of right now I can not give all of the details.......................
................·but when the Project is done I will share with all of you on the Forum

I know that I have·This a another Post but this is a different question
http://forums.parallax.com/showthread.php?p=804156

http://forums.parallax.com/showthread.php?p=804554

Here is the problem that I now have this part resolved but there is a new problem
·
Silo Hold 50,000 Pounds need to keep track of usage need smallest error possible
·
I was using the Seconds */309 to get the Flow Rate··
·
300 Pounds of Sand
248 Seconds to run the·Sand out of hopper
Which give you 1.21 pound per sec or 72.6 per minute
·
So I took 256 x 1.21 = 309.76 rounded to 310

The problem now is that the flow rate changes depending on air flow setting
is there a way to adj the . value on the fly with RC TIME command

.xxxxxxxxxxxx * 65536 base on where RC TIME value is

x value range is from .1xx to .9xx would be nice if this can be done

x is value that want to adj how would I set this up IF this can be done

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Post Edited (sam_sam_sam) : 8/11/2009 3:15:10 AM GMT

phil kenny · 2009-07-31 22:48

What sort of resolution do you want for X?

The Stamp only has integer arithmetic. Depending on the
accuracy you need for the final answer, you may have to
use double precision arithmetic. Tracy Allen has some of
these extended math routines posted on his web site.

You may have to change the Stamp from a BS2 to a BS2P
in order to have enough room.

phil

Tracy Allen · 2009-08-01 22:19

Sam, I'm not sure from your description what you need to adjust.

But suppose you have a fraction N/D<1, and what you really want to compute is x*N/D. That is, the result is a certain proportion of x. N and D may be variables (dependent on air flow?) so you have to compute it at run time, not in advance as a constant. The Stamp can calculate a number, f, such that you can replace x*N/D with x ** f. That will be probably the most accurate way to do it on the Stamp. Here is the algorithm that computes f:

'from http://www.emesys.com/BS2math2.htm#binary_long
' enter with N<32768, D<32768, N<D
'---------binary division loop-----------
for J=15 to 0               ' 16 bits 
N=N//D<<1                   ' remainder*2 
F.bit0(J)=N/D               ' next bit 
next
   '----------------------------------------

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

sam_sam_sam · 2009-08-02 00:24

Tracy Allen

Thank You for your help in this Project

·I think This is just what I was look for

I change the value of N and F gave me a roughly the value that would except

If I use the out put from·the· RC TIME command· which the output from that would be =·N
D = 1% of 65536 which =·655
F =· value .1###########· to .9###########····· ·which is what I want to adj on the fly

I hope that you can understand what I want do

300 pound / 248 seconds = 1.209677411935...........

1 +· .209677411935·x 65536 = 13741..>· This is·what I want basicly dial this·

in so I do not have reprogram the Basic Stamp each time the Flow rate changes

500 pounds / 348 = 1.436781609195................

1 +· .436781609195x 65536 = 28625..> ·and so on

·Will this work Or am I completely wrong with·what I have here

' {$STAMP BS2}
' {$PBASIC 2.5}

J·· VAR··· Byte
N·· VAR··· Word
D·· VAR··· Word
F·· VAR··· Word
·'from http://www.emesys.com/BS2math2.htm#binary_long
' enter with N<32768, D<32768, N<D
'

binary division loop

N = 64000
D = 655

FOR J=15 TO 0·············· ' 16 bits
N=N//D<<1·················· ' remainder*2
F.BIT0(J)=N/D·············· ' next bit
NEXT

·DEBUG HOME,DEC5 ? N, CR, DEC5 ? D, CR, DEC5 ? F, CR
·

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Post Edited (sam_sam_sam) : 8/2/2009 9:20:21 PM GMT

Tracy Allen · 2009-08-02 16:50

Hi Sam, Can you please give it another try at explaining? Lay it out in pseudo-code.

It won't work as you have it. Note that the comment at the start of the code requires N<32768 and N<D, but you have N=64000 and D=655. Why those values? Also in your example you have 1.21 * 65536 = 13741 and 1.41 * 65536 = 28625, but neither of those is correct, even mod 65536. Huh?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

sam_sam_sam · 2009-08-02 20:07

In the post above I left out a step in·how·I got this value· I edit that post ·for you understand what i am trying to do

' {$STAMP BS2}
' {$PBASIC 2.5}
RC              PIN     7
result          VAR     Word

Main:
  DO
    HIGH RC                             ' charge the cap
    PAUSE 1                             '   for 1 ms
    RCTIME RC, 1, [color=red]result[/color]                ' measure RC discharge time
    DEBUG HOME, DEC result              ' display value
    PAUSE 50
  LOOP
  END

I want to take the result value and change ** ---- value on the fly if this can be done
the problem now is that when the flow rate changes I have reload the Basic stamp with the new value I want·to use·a pot to change
this value·

Metered:
 
 GOSUB ReadRam
 
PD = PDsecs + (PDsecs **  [color=red]28625[/color])
SL = SLsecs + (SLsecs **  [color=red]28625[/color])

 
IF oldsecs <> secs THEN
   oldsecs = secs
 
   PDSecs = PDSecs + 1
   SLSecs = SLSecs + 1
 
ENDIF
 
GOSUB WriteRam
 
RETURN

·

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Tracy Allen · 2009-08-02 22:14

OK, so the result from RCTIME is somehow related to air flow and then to (pounds per second). Does result have a maximum and a minimum value? Does it need to be calibrated?

We know that pounds per second is a number around 1.2 or 1.4, but what is the full expected range?
(pounds) = (seconds) * (pounds per second)

But how do you figure this:
(pounds per second) = function_of (result)?

For example, if result=12345, what is (pounds per second)?
or if result = 54321, what is it?
Maybe there is a math formula, or maybe it is empirical, or maybe it needs a table of values to figure it out.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

sam_sam_sam · 2009-08-02 23:48

Tracy Allen

Does result have a maximum and a minimum value?· Yes

1.1 to 1.9 pound per seconds is the excepted value

·Does it need to be calibrated? Yes some what in that if the pound per second is way·off then the amount in the Silo will be way·off

So the result from RCTIME is somehow related to air flow and then to (pounds per second).

·Yes in·that I would like to adj the Flow rate value

I will still would have to figure this part·out manually

Pound / seconds = pounds per seconds then 1 + (. some value) ·x 655365 = P Basic math value

Then take this value and basically dial this in as the new value

·That would be ADJ by RCTIME value so I would not need to reload the Basic Stamp Again with this new pound per second value again
·
·
We still have to check the flow rate from the hopper to see if that value· is still right
I can not change that part

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Post Edited (sam_sam_sam) : 8/3/2009 12:27:12 AM GMT

sam_sam_sam · 2009-08-03 01:23

Tracy Allan

I wrote this Demo to test the end value of 50,000 pound which is how much the Silo hold to see how far off the actual value versus the Basic Stamp
value is

I tried this Demo to see if I could use the RCTIME to do what I have in mind and it seem to work OK

Please give your thoughts on what·I have here if you see any problems with what I have in the Demo

Please feel free to let know what changes·need to make this work

Thank You for all of your help in this Project this has help me a lot I have learned a lot about programing

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Post Edited (sam_sam_sam) : 8/3/2009 1:35:13 AM GMT

Tracy Allen · 2009-08-03 02:50

I see the RCTIME, and there is a capacitor, but what is it hooked up to? What is it measuring?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Tracy Allen · 2009-08-03 22:46

Okay, after the PM I understand better what is going on. You have a resistor variable from 1 kOhm to 11 kOhm in parallel with a 0.1 uF capacitor. The corresponding RCTIME values will fall approximately in the range of 70 to 740. You need to map that range into a multiplier with a decimal fraction of of 0.0 to 1.0, or in Stampese a ** multiplier from 0 to 65535.

Here are the calibration steps.
1) Have PBASIC find the RCTIME value at maximum resistance, call it RTmax
2) ditto the RCTIME value at minimum resistance, call it RTmin
Those two values will be approximately 740 and 70.
3) Have PBASIC calculate the difference RTdiff = RTmax - RTmin, approximately 670.
4) Turn the dial to an intermediate setting, and have PBASIC measure the RCTIME, call it RCset
5) run the division algorithm:

'from http://www.emesys.com/BS2math2.htm#binary_long
' enter with N<32768, D<32768, N<D
'---------binary division loop-----------
N = RCset - RCmin MAX RCdiff
D = RCdiff
for J=15 to 0               ' 16 bits 
N=N//D<<1                   ' remainder*2 
F.bit0(J)=N/D               ' next bit 
next
   '----------------------------------------

Now, the Word value f contains the multiplier you need to use in
mass = seconds + (seconds ** f)
The overall result is something in the calculator range of 1.0 to 2.0 as the mass per second conversion.

You could calibrate the dial of the potentiometer in units from 1.0 to 2.0. If you want to translate the f factor into a decimal fraction to display on your LCD screen, use,
x = f ** 1000
DEBUG "1.", DEC3 x ' or the LCD equivalent.

Note that some of the variables are temporary and can be recycled.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

sam_sam_sam · 2009-08-05 02:30

Here is the problem that I now have this part resolved but there is a new problem
·
Silo Hold 50,000 Pounds need to keep track of usage need smallest error possible
·
I was using the Seconds */309 to get the Flow Rate··
·
300 Pounds of Sand
248 Seconds to run the·Sand out of hopper
Which give you 1.21 pound per sec or 72.6 per minute
·
So I took 256 x 1.21 = 309.76 rounded to 310

The problem now is that the flow rate changes depending on air flow setting
is there a way to adj the . value on the fly with RC TIME command

.xxxxxxxxxxxx * 65536 base on where RC TIME value is

x value range is from .1xx to .9xx would be nice if this can be done

x is value that want to adj how would I set this up IF this can be done

This· give me·the range that I want and it tell what the flow rate pounds per seconds

I want to thank Tracy Allen for all of his help in this project

and by this doing this project i have learned a lot about programing a Basic Stamp

I also want to Thank all of·you·that have replied and offer help in this project

' =========================================================================
'
'   File...... Multiplier with a decimal fraction a value from 0.0 TO 1.0
'   Purpose... To use a POT to adj a value on the fly from .000 to .999
'   Author.... Tracy Allen
'   E-mail.... [url=http://www.emesys.com/]www.emesys.com[/url]
'   Started... 8-2-09
'   Updated...
'
'   {$STAMP BS2}
'   {$PBASIC 2.5}
'
' =========================================================================
'from [url=http:///BS2math2.htm#binary_long]http:///BS2math2.htm#binary_long[/url]

' -----[noparse][[/noparse] Program Description ]---------------------------------------------
'
'
' Suppose you have a fraction N/D<1, AND what you really want TO compute is x*N/D.
' That is, the result is a certain proportion of x. N AND D may be variables
' (dependent ON air flow?) so you have TO compute it at run time, NOT in advance as
' a constant. The Stamp can calculate a number, f,
' such that you can replace x*N/D with x ** f.
' That will be probably the most accurate way TO DO it ON the Stamp.
' Here is the algorithm that computes f:

'
' Part 1
'
' You have a resistor variable from 1 kOhm TO 11 kOhm in parallel with a 0.1 uF capacitor.
' The corresponding RCTIME values will fall approximately in the range of 70 TO 740.
' You need TO map that range into a multiplier with a decimal fraction of of 0.0 TO 1.0,
' OR in Stampese a ** multiplier from 0 TO 65535.
'Here are the calibration steps.
'
'1) Have PBASIC find the RCTIME value at maximum resistance, call it RTmax
'2) ditto the RCTIME value at minimum resistance, call it RTmin
'   Those two values will be approximately 740 AND 70.
'3) Have PBASIC calculate the difference RTdiff = RTmax - RTmin, approximately 670.
'4) Turn the dial TO an intermediate setting, AND have PBASIC measure the RCTIME,
'   call it RCset
'5) run the division algorithm:
'
' enter with N<32768, D<32768, N<D
'
'   Part 2
'
' Now, the Word value f contains the multiplier you need TO use in
' mass = seconds + (seconds ** f)
' The overall result is something in the calculator range of 1.0 TO 2.0 as
' the mass per second conversion.
' You could calibrate the dial of the potentiometer in units from 1.0 TO 2.0.
' IF you want TO translate the f factor into a decimal fraction TO display ON
' your LCD screen, use,
' x = f ** 1000
' DEBUG CR, DEC RC_Set, TAB, "1.", DEC3 X, CLREOL
' Note that is (Tab) to the next column for DEBUG .
'
' -----[noparse][[/noparse] Revision History ]------------------------------------------------
'
'
' -----[noparse][[/noparse] I/O Definitions ]-------------------------------------------------
 
RC              PIN     15
              
' -----[noparse][[/noparse] Constants ]-------------------------------------------------------

RC_Diff         CON     600
RC_Max          CON     613
RC_Min          CON     13
 
' -----[noparse][[/noparse] Variables ]-------------------------------------------------------

J               VAR     Byte
N               VAR     Word
D               VAR     Word
F               VAR     Word
x               VAR     Word
RC_Set          VAR     Word
 
' -----[noparse][[/noparse] EEPROM Data ]-----------------------------------------------------

' -----[noparse][[/noparse] Initialization ]--------------------------------------------------

' -----[noparse][[/noparse] Program Code ]----------------------------------------------------
  DO

    Main:
 
    HIGH RC                                            ' charge the cap
    PAUSE 1                                            ' for 1 ms
    RCTIME RC, 1, RC_SET                               ' measure RC discharge time
    PAUSE 50
 
 N = RC_set - RC_min MAX RC_diff                       ' binary division loop
 D = RC_diff

FOR J=15 TO 0               ' 16 bits
N=N//D<<1                   ' remainder*2
F.BIT0(J)=N/D               ' next bit
 
NEXT
 
 DEBUG HOME, DEC RC_Set, TAB, "1.", DEC3 X, CLREOL
 x = f ** 1000
 
  LOOP

' -----[noparse][[/noparse] Subroutines ]---------------------------------------------------

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Post Edited (sam_sam_sam) : 8/5/2009 3:04:20 AM GMT

sam_sam_sam · 2009-08-11 03:29

The new problem is that we are putting a Air Regulator on this Sand Blast Hopper

Because the·Air Flow Rate now is Controlled with a Ball Value right now

the Flow Rate is suppose to be 300 CFM @ 120 PSI· MAX·and the compressor that we are using put out 700 CFM @ 140 PSI

and the Flow Rate may below 1 pound per second What would I need to change to make this work with less than 1 pound per seconds

I would like to use· **· instead of */ if this can be done

Just so I know how to do it how would ·you do it if your Flow Rate is more than 2 pounds per second

The person that works with this Sand Blaster will have find the right flow rate for the best work to be done

Then I can take this New Flow Rate and go from there

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
··Thanks for any·

·that you may have and all of your time finding them

·
·
·
·
Sam

Post Edited (sam_sam_sam) : 8/11/2009 3:38:02 AM GMT

Can P Basic Math be done on the fly with RC TIME command this is for Tracy All

Comments