Servos not serving

I have two servos running off a 16F887 with no issues. Both toggle back and forth at roughly a 90 degree angle, in about 2 seconds.

I changed commands for a 12F675, mounted same on a prototype board with hefty 5 volt power supply/converter. With one servo connected (to GP1), everything is fine. With TWO connected, it's as if the time period was extended by a factor of 100: each servo does a little "tick" in about 1 second: at that rate, it will take 10 minutes or more to swing 90 degrees. Not only that but there is chattering where they can suddenly move 10 degrees. If I swap servos, same thing occurs. There is NO voltage drop on the MCU or servos when #2 is added. There is no NOISE on the power lines. Different servos create the same problem.

On a scope I cannot see any change whatsoever. The pulses are there, varying 1.0 to about 1.5 mS, and the overall period is 20 mS. (It's a little lower, frequency is about 65 Hz instead of 50 Hz). The program is attached for the 12F675. Even when the servos are chattering, the waveforms look okay.

Am I missing something? Is this a software or a hardware (ie MCU) problem?

Did you make sure the combined current is not too much?

Robert

Is your MCLR tied high?

See this link.
http://support.melabs.com/threads/10...ighlight=servo

When I was exercising I was thinking of my problem, checked the spec sheet but saw that GPIO can handle 25 mA each, total 125 mA for all I/O. I was way below that with two servos.

After more checking, I was shocked to see that NOW, neither of my servos worked, so programmed the chip again with (apparently) no problems, except servos no work. Checked power: 12 volts onto board, 7.3 volts coming off 5 volt regulator! Geez. Don't know why I blew a regulator. It's a heat-sunk T0-220, all my circuitry checks out as not having shorts or anything and my 5 volt regulator never got hot.

So I popped the 12F675 into my MikroElektronika Easy Pic V7, programmed the chip: didn't work. Threw the chip out, put in a new one and all is well: both servos work as specified.

Thanks Tabsoft for pointing out my earlier question, same problem, 6 months ago. I'm losing it. BTW, MCLR is tied high through a 10K resistor. I will investigate WHY my circuit doesn't work.



Obviously my prototype board - which I designed, built, and populated - is not...done right. Mea Culpa.

Thanks for those who responded.

may be just try this one ...

Code:

---

' Name        : ServosDumpTruck675X3.pbp
' Compiler    : PICBASIC PRO Compiler 3.074
' Development PIC  : 40-pin 16F887
' Target PIC  : 12F675 8 pin
' Hardware    : Mikro Elektronika EasyPic V7
' Oscillator  : 4MHz internal clock
' Keywords    : servo, dump truck, X3
' Description : PICBASIC PRO program for control of 3 (or more) servos

#CONFIG
    __config _INTRC_OSC_NOCLKOUT & _WDT_ON & _PWRTE_ON & _MCLRE_OFF & _CP_OFF & _BODEN_ON
#ENDCONFIG

 DEFINE OSC 4
 DEFINE OSCCAL_1K 1

'****Aliases*****
servo1  Var    GPIO.1 ' Alias servo pin
servo2  var    GPIO.2
servo3  var    GPIO.4 ' can't use gp3...input only.


'***Constants and Variables ***
PWCCW  con    120    'CCW Pulse Width:  100 mS * 4 Mhz period (10uS)=1.0mS  def 100
PWCW    con    240    'CW pulse width:  200 mS * 10uS=2.0 mS      def 200
Refresh con    20      '20 mS refresh= 50 Hz
pulse  var    byte 


'****Initialize****

  ADCON0                = 0
  ansel                = 0          'all digital i/o
  cmcon=%00000111  'turn off comparators
  GPIO                = 0 
  trisio                = %00001000 

 '  Low Servo1            ' Initialize servos
 '  Low Servo2
 '  low servo3
     
'****Main Program****
Start:
   
For Pulse= PWCW TO PWCCW STEP -1

        Low Servo1
    PulsOut servo1,pulse
        Low Servo2
    pulsout servo2,pulse
    Low Servo3
    pulsout servo3,pulse
                 
    Pause Refresh - pulse - pulse - pulse

next pulse
   
           
for Pulse= PWCCW TO PWCW step 1

    Low Servo1
    PulsOut servo1,Pulse
    Low Servo2
    pulsout servo2,pulse
    Low Servo3
    pulsout servo3,pulse
                                           
    Pause Refresh - pulse - pulse - pulse
                                                                                               
next pulse         

goto Start

---

NOW ...

1) be sure your servos supply can hold ... 3 Amps short peaks ( not a typo !!! ) ... the EP7 7805 looks a bit weak for that !!!

2) see where the LOW commands have to be placed.

3) your " pause refresh " value looked a bit strange ...

Alain

Alain,

I guess I'm missing something in your "Pause Refresh - Pulse - Pulse - Pulse" statement.
The "Pause" amount for loop # 1 will be "20 - 240 - 240 - 240 = 68ms"
For loop #2 it will be "20 - 239 - 239 - 239 = 71ms"

So to turn 90deg it will take 15.16 seconds?
Is this right?

Code:

---

Total Pause (ms):          15.16
Total Pause(Value): 15160

Pulse        Pause
Value Value
----- -----
240        68
239        71
238        74
237        77
236        80
235        83
234        86
233        89
232        92
231        95
230        98
229        101
228        104
227        107
226        110
225        113
224        116
223        119
222        122
221        125
220        128
219        131
218        134
217        137
216        140
215        143
214        146
213        149
212        152
211        155
210        158
209        161
208        164
207        167
206        170
205        173
204        176
203        179
202        182
201        185
200        188
199        191
198        194
197        197
196        200
195        203
194        206
193        209
192        212
191        215
190        218
189        221
188        224
187        227
186        230
185        233
184        236
183        239
182        242
181        245
180        248
179        251
178        254
177        1
176        4
175        7
174        10
173        13
172        16
171        19
170        22
169        25
168        28
167        31
166        34
165        37
164        40
163        43
162        46
161        49
160        52
159        55
158        58
157        61
156        64
155        67
154        70
153        73
152        76
151        79
150        82
149        85
148        88
147        91
146        94
145        97
144        100
143        103
142        106
141        109
140        112
139        115
138        118
137        121
136        124
135        127
134        130
133        133
132        136
131        139
130        142
129        145
128        148
127        151
126        154
125        157
124        160
123        163
122        166
121        169
120        172

---

Quote:

---

Originally Posted by Tabsoft

Alain,

I guess I'm missing something in your "Pause Refresh - Pulse - Pulse - Pulse" statement.

---

you're right ! I missed time scaling :wink:

should be

Code:

---

Pause Refresh - ((pulse + pulse + pulse)/100)

---

note Refresh could simply be 15ms ( as precise 50 Hz repeating signal is not compulsory - only the 1.2 - 2.4 ms "1" level - )

your travel time will simply be ...

(240-120)* 20ms = 2.4 seconds ...

Just wanted to say thanks to all for your insightful comments and programming hints. I printed off the stuff for later perusal and comparison to my program. I'm a "minimalist", I program with the least amount of code required (thus don't always initialize) and even build circuits with low parts counts. That's where I get into trouble...

Thanks again.

If you want the duty cycle to be as close to 50Hz you can do this to account for the Pulsout statements, Low xx statements, For/next logic and the Pause statement.

Change the Refresh constant value to 20000 'microseconds

Then use this for the pause.

Pauseus Refresh - (((pulse + pulse) + (pulse / 3)) * 10)

This should keep the total duty cycle time to 20ms +/- (~100) microseconds.

Queenidog, Here is a program I wrote some time ago for a couple of kids to test some servos for there robot in the FIRST competition. It is a minimalist program I wrote during lunch one day. It might help. Oh I forgot, Also attached is a schematic I updated a couple of years ago. Enjoy...