Brijm
Lifetime Supporting Member
Beej,
Sounds like you have a slow changing system, that you need to compensate for deviation. Depending on how important accuracy is (or if the speed only needs to be in a certain range) I think your prox will work. At these slow speeds, with slow devieation. You might be able to calculate the number of pulses over say 30 seconds, to determine your speed. (The longer your average, the more accurate it will be, but the slower the control) Compare this to your setpoint, and either increase or decrease your reference accordingly. I recenlty did something similar on a fan system. I didn't want to control for fast fluctuations, but needed to increase the fan speed, as the filter gets dirty to maintain a set flow. It's a simple / slow control, that gets the job done.
The only thing that I'm not sure about is if this is a normal reaction of the process, or a component failure. If it's a normal reaction, then it is controlable. If you are simply covering up symptoms, this ussually comes back to bite you.
brian
Sounds like you have a slow changing system, that you need to compensate for deviation. Depending on how important accuracy is (or if the speed only needs to be in a certain range) I think your prox will work. At these slow speeds, with slow devieation. You might be able to calculate the number of pulses over say 30 seconds, to determine your speed. (The longer your average, the more accurate it will be, but the slower the control) Compare this to your setpoint, and either increase or decrease your reference accordingly. I recenlty did something similar on a fan system. I didn't want to control for fast fluctuations, but needed to increase the fan speed, as the filter gets dirty to maintain a set flow. It's a simple / slow control, that gets the job done.
The only thing that I'm not sure about is if this is a normal reaction of the process, or a component failure. If it's a normal reaction, then it is controlable. If you are simply covering up symptoms, this ussually comes back to bite you.
brian