analog feedback 7344 0-5V non liniar

[Peter Rene]

Hi.
Having a 7344 left over from an aborted project, I wondered if this controler would accept a non liniar 0-5V position feedback in PID closed loop.
I know the analog input is "only" 12bit, but this gives by far the required position on the travel I need for this application which is to be an automatic throttle control on my enginedyno. It is not possible (nor practical) to mount encoders on the many different throttlesystems, so the 0-5V from the throttlepositionsensor is all I have.

If ok, then what kind of motor should I look into, stepper og servo since the application is more still standing in any position than actual turning (running is expected to be very few turns when geared to move throttle 0-80 grd. I have both a SX and CX Nudrive.

Thanks in advance

Peter Rene

[Jochen]

It should be possible to use the 0-5V signal as an analog feedback. You can configure the ADC settings for a 0 to 5 V range that maps the full 12 bit resolution to your feedback range. For position control it doesn't matter that your feedback signal is non-linear as long as you know how the voltage of your feedback signal corresponds to the position of the motor. Still you will see the same amount of non-linearity in the velocity profile of your axis as long as you don't run your axis always in contouring mode to compensate it.

Even though most time your axes are static which I agree would suggest to use a stepper I recommend to use a servo instead. The reason is the fact that closed loop steppers are not controlled in the same way as servos. In theory, analog feedback for stepper motors should work correctly; however, there are several problems that will arise when you attempt to control a stepper motor from an analog feedback. In this type of system, the analog feedback is converted into a position and compared against the desired position, and then the stepper motor is "pulled in" to that position. One of the problems with analog feedback for stepper motors is noise, another one are non-linearities of the feedback signal. Closed loop stepper mode relies on a fixed ratio between the generated steps and the feedback. This is not the case if the feedback signal is non-linear. Thus you will have to use a servo motor.

As an alternative you should consider using a stepper motor in open loop mode (without feedback). If you can make sure that you don't loose steps this would be the best and easiest option.

I hope this helps,

Jochen Klier
National Instruments Germany

[Peter Rene]

Thanks for the very quick reply! 🙂

I´m not shure if I will loose steps, but most likely I think. The reason for this is I normally will have to hook it up to the existing throttle cable with another cable, and combined with vibrations and the "fleksible" connection I would be surprised if not. But if this tendence settles after a few corrections it could perfectly be an open loop stepper. I would be happy with positiontolerance of +- 0,2 grd on the throttle, achieved through gearing approx 5 times.

Running open loop, and using the encoder for position:
Would it be possible to easy "learn / store" the actual (always individual) throttle travel into 32 breakpoints (positions), which I initially approach manually stepwise from idle to full throttle while reading the desired position and storeing each of the 32 encoder positions when happy?
I´m thinking something like stepping the motor forward with either a real or software potentiometer or / similar in this "learning" phase.

If so, still in open loop, I would most likely need to be able to have easy online acces to trim / offset the individual stored beakpoints due to the floating connection vibration stuff.

I would then have a list of 32 desired breakpoints, throttlepositions. Could a button be atached to each of these 32 positions, making it possible to go directly to any of these and directly from that to any other or home when chosen on the respective button / condition?

Thanks in advance

Peter Rene

[Jochen]

Peter,

sorry I'm a bit in a hurry as I'm going to leave for vacation pretty soon so this will be just a brief answer.
Your idea sounds good. If I understand you correctly you are not referring to the breakpoint feature of the board (this works only in closed loop mode). Everything else shouldn't be a problem to program in LabVIEW.

Best regards,

Jochen