Need help with Servo Drive analog...

Discussion in 'The Projects Forum' started by franzschluter, Jun 6, 2009.

  1. franzschluter

    Thread Starter Active Member

    Jun 1, 2009
    95
    0
    Available connections on the system is..

    1. Reference Voltage can be 2,2V up to 12V (1st parameter) .. It is adjustable with a poti..(Not linear)
    2. Linear Potentiometer... (2nd parameter) 12V when fully open and 2,2V when fully closed..

    What happens is:

    The motor will run Linear Potentiometer slider up and down....
    Now you can set how much down it should go...

    Essentially this is easy by using a comparator circuit... But this isn't precise enough because of the momentum of the machine..
    This isn't also how the boss wants it...

    What we want is..

    As it approaches the set point (set by poti)...

    Initially it will drive the motor very fast using PWM (80% DC is enough) .... But as it reaches the point of destination it will get slower (30% DC) first before coming to a complete halt (0% DC).. Or when it goes over the limit it will reverse to the set point...

    This must go very very precise up to 5 micron tolerance... So reverse must not happen! Or if it does happen it must not happen that frequently because of mechanical backlash...

    The current setup I am using is with a comparator and is always running at 30% DC. To make it precise as possible.. But they aren't satisfied and want it to run initially fast.. But as it approach goal it will go slower and then stop.

    I used the term "motor" but in reality it is a lesser load.. It's a hydraulic system that has a 12V input and 100mA consumption.. The slightest voltage applied will already move the system up or down by the microns...

    Any advice? Like what components should I use? I have PID running always in my mind.. But I like to avoid it because I have no experience with PIDs only in textbooks and theory. It should be satisfactory enough if the distance was linearily proportional to its speed.

    Best Regards
    Franz
     
  2. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    I can't point to any published circuits, but devices like chart recorders used to have servo amps that acted much as you wish. That 5 micron accuracy could be a problem, though. How far must the motor slew? If it's on the order of 10's of mm, then you will have a hard time getting that final resolution. The amp and the sensor will have to be very good.

    If the mass is not too great, a stepper motor with suitable gearing can give the resolution, but only over a small distance.
     
  3. franzschluter

    Thread Starter Active Member

    Jun 1, 2009
    95
    0
    Yeah it's the amp that must be damn good... The sensor I have no doubts that it will be able to accomplish this task. It's a linear Potentiometer shaft that has basically "infinity resolution" and very much dependent on the driver. It is also very expensive... Same like those glass linear scales from heidenhain...

    I don't need a published schematic for it and I don't believe there would be one since most are patented if not confidential since a lot of machine builders would be using them in their machines.. But perhaps only the idea would be enough for me to create one.

    Distance is 500mm.

    And yes it is a challenge. The mass is about 200kg stepper motors is out of the picture.. But it does have a hydraulic system with a high precision valve that is electrically controlled. According to it's datasheet it can do infinity resolution also but is dependent only on the driver...

    Movement has to be initially very fast. But will start to de-accelerate when approaching the target mark.

    Microcontroller,Comparators basically anything goes to this circuit. Budget is 1,000$ for electronic components and circuitry.


    I have some stuff pop up in my dream last night... But this still wouldn't be what I want...
    It's when it passes by the limit/set point a comparator circuit will trigger.. Then it will reverse for 2 seconds... The reverse phenomenon will trigger as 1 count.. This count will reduce Duty Cycle by lets say 25% each time it passes over the set point. These inputs are then fed to a MCU. It is 1 solution but still not what I want. It would be a sort of motion oscillator if I did this lolz.
     
  4. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    try moving from a comparator to an opamp. The farther from reference, the higher the drive.
     
  5. franzschluter

    Thread Starter Active Member

    Jun 1, 2009
    95
    0
    I'll look into it.. You mean making a differential op-amp amplifier? With negative feedback I suppose. Will the op-amp be able to drive 500mA? Or should I make some interface in between or something. Remember it isn't a motor but a coil with low current consumption (500mA). It doesn't have any initial "not wanting to move phenomenon because it's too heavy"
     
  6. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    Yes. The application I had experience with was a 100ma coil on a proportioning valve that drove swash plates on variable displacement pumps. Another application was in the North American jet valve that directly operated a 5" bore cylinder for furnace stack pressure control.
     
Loading...