Servo not reaching full travel

[ruijc]

Thanks for the warning Jumper,

i have to be more carefull with that from now on

skimask,

i've checked my pbp manual and you are right. Osc 3 is really 3.58 Mhz.

I will try with lower speeds...maybe using the internal oscilators and see what happens

.

[skimask]

skimask,
i've checked my pbp manual and you are right. Osc 3 is really 3.58 Mhz.
I will try with lower speeds...maybe using the internal oscilators and see what happens
.

That's the thing...you shouldn't have to do that.
I can run servo's just as happily running 48mhz on the 18F4550 as I can at 4mhz on the same PIC, same program, same everything, just a different DEFINE OSC at the beginning.
You've got something else going on that's preventing you from doing so.
Port not set right, interrupt's bugging something, register not set up right or something...

And if I remember right, I don't think I've seen your FULL code yet...just fragments.
That doesn't work so good.
Time and time again, somebody says, 'well, the problem can't be over there, 'cause that's got nothing to do with anything over here'.
Well, guess what.....the problem was over there because it had everything to do with all of it over here in one way or the other...

If you run the servo to an endpoint (at least an endpoint that you've found using the PIC and it's software), how much farther (if any) can you make it move after removing power by manually turning it?

[ruijc]

skimask ,

the full code with settings is in my 2nd post.

I ran the program again and moved the servo to the far right and then i moved it with my fingers and yes...it moved a bit more !

this means that it can move a bit more to the right but it's limited with the low 25 ( 1025 ms ).

To the left it goes all the way !

This also means that center is not where i though it was .

.

[skimask]

This also means that center is not where i though it was .

How are you quantifying 'CENTER' on the servo right now? I mean how are you determining 'center'?
Are you trying to let the program do it for you?
Are you moving it manually from end to end and splitting the difference and making a mark at that 'center'?
If you're looking at center and saying that's 'CENTER' because you want it to be center, well, that ain't gonna work.
Take that manual method described above of moving the servo and write a program to send the servo a 1.5ms pulse and see how far off the two marks are. Let the servo tell YOU where center is.

If you are still running that code in your second post, I think you should abandon that code until you figure out how to get full left and right out of your servo's first, before you start adding in buttons and stuff...
Go back to that code in Post #9 and play with that.
Using those values, the servo should move equally left and right (not just a little left and a lot right, 'cause that would mean that your idea of center and the servo's idea of center are two different things, which I think is wrong) and it should do that without binding at either end.

[ruijc]

skimask,

i ran my code and yours ( as transcribed bellow ) and this was what i got ( please see attached pic ):

Manually found left and right limits of this servo and marked it's true center

Fisical center matches the 1500 ms impulse

With my code:
max right = +- 2150ms
max left = +- 1025 ms

With your code :

max right = +- 1950 ms
max left = +- 1025 ms

both are far from servo fisical limits ( fisical limits marked with in red )



your code :

'************************************************* ****************************
'19/12/2007
'
'Skimask's servo code
'
'PIC 16F628 @ 4Mhz

'************************************************* ****************************
'Config

PIE1 = 0 'off
PIR1 = 0 'off
CMCON = 7 'comparator off
VRCON = 0 'Vref Off
T1CON = 0

' internal osc selected in IC prog
' MCLR ON selected in IC prog

'************************************************* ****************************
'DEFINEs
'************************************************* ****************************

'OSC

DEFINE OSC 4

'************************************************* ****************************
'
'LCD

' DEFINE LCD pins
DEFINE LCD_DREG PORTB
DEFINE LCD_DBIT 4
DEFINE LCD_RSREG PORTA
DEFINE LCD_RSBIT 2
DEFINE LCD_EREG PORTA
DEFINE LCD_EBIT 3
DEFINE LCD_BITS 4

DEFINE LCD_COMMANDUS 2000

'************************************************* ****************************
' variables

va1 var word
va2 var word
ser var word
pos var word
na var word
nb var word
stepval var word
center var word
leftend var word
rightend var word

CLEAR

'************************************************* ****************************
'I/O's

PORTA = %00000000
PORTB = %00000000

TRISA = %00000011
TRISB = %00000011

'************************************************* ****************************
'PINS

bu1 var PORTA.0
bu2 var PORTA.1
R_S var PORTA.2
E var PORTA.3
a4 var PORTA.4
MCLR var PORTA.5
a6 var PORTA.6
a7 var PORTA.7

DATA1 var PORTB.4
DATA2 var PORTB.5
DATA3 var PORTB.6
DATA4 var PORTB.7

bu4 var PORTB.0
bu3 var PORTB.1
led var PORTB.2
servo1 var PORTB.3


'************************************************* ****************************
'************************************************* ****************************

pause 200 'init LCD
lcdout $fe,1 'clear LCD

Start:

leftend = 1000
center = 1500
rightend = 2000
stepval = 20
pos = center

moveleft:
pause 19 'pause 19ms
if pos > leftend then
pos = pos - stepval
else
goto moveright
endif
servo1 = 1 : pauseus pos : servo1 = 0 : goto moveleft

Moveright:
pause 19 'pause 19ms
if pos < rightend then
pos = pos + stepval
else
goto moveleft
endif
servo1 = 1 : pauseus pos : servo1 = 0 : goto moveright

end




EDIT:

I just changed your settings to :
leftend = 500
center = 1500
rightend = 2500

and i got more travel to the right ( more that i got before with mine )
This means that the servo is working bellow 1ms

[ruijc]

Just an update of my previous post:

I changed the values as far as :

leftend = 200
center = 1500
rightend = 3000

( the servo moved a bit more but stopped before reaching it's fisical limits. Also, these values do not represent the true extreme limits )

The new max travel made by the servo is marked in this new picture.

.

[skimask]

I'm thinking (supposing really) that maybe just because a servo has a physical limit, doesn't mean that the electronics inside will let it reach those physical limits. I don't know.
Those marks you made, the black ones, look like a halfway decent 90 degree angle.

How about trying this out for grins:

....all the setup you've already got...it looks like it should work just fine....
Start:
leftend = 100 : center = 1500 : rightend = 2900 : stepval = 20 : pos = center

moveleft:
pauseus 20000 - pos  'pause 20ms - servo pulse width, should keep a more constant refresh rate closer to 50hz
if pos > leftend then
pos = pos - stepval
else
goto moveright
endif
servo1 = 1 : pauseus pos : servo1 = 0 : goto moveleft

Moveright:
pauseus 20000 - pos  'pause 20ms - servo pulse width, should keep a more constant refresh rate closer to 50hz
if pos < rightend then
pos = pos + stepval
else
goto moveleft
endif
servo1 = 1 : pauseus pos : servo1 = 0 : goto moveright

end

I changed it a bit, added a bit of code to keep the refresh rate a bit closer to 50Hz. Maybe the newer servo's are a bit more sensitive to that sort of thing.
Also, I changed leftend and rightend to EXTREME values. They're almost GUARANTEED to bind your servo's at the extreme's.
See what happens.
Are you actually using a real 16F628 or a 16F628A?

[ruijc]

Hi skimask,

i'm using a 16F628A.

With your new code the servo moves all the way up to the new red marks ( as showned in pic # 2 ).

It is still not at it's fisical limit but i think that probably these are the electronic limits of the servo.

.