Retrofit position control to hydraulic ram

UKB

Member
Join Date
Aug 2014
Location
UK
Posts
56
Hiya -

just hoping someone might be able to point me down the right path here;

I have recently finished a job for a press cutting some carpet like material.
One of the modifications that has come up is stopping the cutting head accurately so that it cuts through the product, but not into the thin rubber conveyor belt below. Must stop within 0.2mm ish I guess.
Recipe system so lots of different cutting depths are required.

Currently the depth of cut is indexed with a few shims and operator input...

I have considered
LVDT --
draw wire encoder
linear encoder

I'm leaning toward a draw wire encoder -

Has anyone done something similar before, any pointers?

Kind Regards
UKB.
 
Resolution of any feedback device will depend on the overall length of travel of the cylinder. Position accuracy will have to do with speed of the ram, and your update speed of your plc, as well as what type of hydraulic valve is used to control the ram, ie: solenoid valve, servo-valve, servo-proportional valve, ect.

For high speed, high accuracy, a dedicated motion controller might be the way to go.

If it was me, I would use a Temposonic inside the hydraulic cylinder, if that is not possible, then an external mount will have to do, but be very careful of damage to the Tempo rod.

As you can see, there are several variables that have to be looked at before you can make a decision on what to use.

I am sure Peter N will weight in on this subject as this is the kind of application he works on every day.

Edm.
 
Thanks EDM hydraulics,

The cylinder is controlled via solenoid valves;
max stroke length is 200mm
operating stroke is ~ 100mm
max speed 400mm / sec


Couple of ideas;
non contact laser via profi interface or analogue input card.

I think the internal ultrasonic will be too expensive for this job,

Kind regards
UKB
 
Last edited:
While the level of repeatability you are looking for is certainly not that great, it still may be beyond the capability of the current valve you are using. If you want to try to do this using your current valve and some analog feedback into your PLC you will almost definitely have to slow down the cylinder to a creep speed before trying to stop it.

Here is everything that will be going on against you as you try to get into a repeatable position:
  • As the cylinder is moving, the transducer is measureing the position but there is some time related to converting its physical measurements to the analog signal to send to the plc
  • Then on the PLC side, there is a time it takes to convert that analog signal back into a number
  • Then there is the scan of the PLC. In the worst case your PLC scan could line up to read the analog value just before the new value is written so you are using the oldest possible data. By the next scan you could be beyond your target.
  • Then you turn on an output to a valve which has its own inherint delay
  • Then there is the reaction time of the valve which besides changing over time as the valve wears will likely change with temperature and pressure as well

Each of these steps along the way will add uncertainty to your final position. If your process time allows you to move slow enough so that all these uncertainties still leave you within your deisred position window then you MAY be able to do it with your current directional valve.

1. You need a way to control the slow speed by switching a flow control into the circuit (you will need a very good one with temperature and pressure compensation).
2. When you get to your position, you have to be able to hold it without actualy contacting hard stops. The knife reaction forces will be trying to push yoru cylinder out of position.

We do have one customer that does use a method like this to get into a window of about 0.25mm but the slow speed is so slow you almost can't see the cylinder moving. Even with the slow speed, we need to test each system and come up with an "approach distance" where we turn the valve off before we actually get to the desired position so that the cylinder coasts in and we don't overshoot. We also have been burned by slow analog cards which take too long to convert the signal to a number. And when we do this, we have positive locking valves to keep the cylinder in position.

You are probably best served by looking at your overall system and re-designing it as a closed loop position system with a prop valve and dedicated controller for cylinder position.
 
Last edited:
Ditto what EdmHydraulics said. 0.2mm is easy for a hydraulic motion controller but difficult for a PLC.
Ditto Norm's comment about slow analog input and output cards.
I prefer SSI MDT rods from Balluff or Tempsonic.
Analog cards are slow, noisy and the sample time is almost random making it impossible to compute an accurate speed.
Valves spools should have 0 overlap.
Valves should be mounted directly on the cylinder with NO HOSE between the valve and the cylinder.

This application should be easy if one doesn't try to cut corners.
 
Can you replace the shims with something controllable?
So I am picturing a cutter that comes down to some height above the rubber conveyor. Currently that stop point is a mechanical stop.
Instead of controlling the cutter as it moves, can you control the mechanical stop instead.
Generally (your mileage my vary) you need to set the stop once per setup. You can move it nice and slow and walk it right into position. Then during production, the knife can zoom up and down all it wants.
 
Thanks all for the replies,

To workaround a slow analogue card issue I thought about getting a laser distance displacement tool which has an RS422 interface, would this have the same time based constraints as the analogue card? (SICK - OD2P 250W150A0 + rs422 card)

Nevis - great idea.

I know that the valve is situated ~ 1M away from the ram via flexible hose. I will look into the other factors,

thanks for your help.

Will
 
To workaround a slow analogue card issue I thought about getting a laser distance displacement tool which has an RS422 interface, would this have the same time based constraints as the analogue card? (SICK - OD2P 250W150A0 + rs422 card)
You will need to look at the update rate and how it synchronizes with the PLC.
The problems is that you have a position but that is not good enough. You need to know when that position was value.


I know that the valve is situated ~ 1M away from the ram via flexible hose. I will look into the other factors,
That hose is a will make the project difficult if there is any changes in load or fast accelerations or decelerations. Basically the hose reduces the natural frequency of the system.
 

Similar Topics

This is a Pharma Plant application. The Plant is highly Automated, but there are still many, many Manual Valves throughout. Is it possible to...
Replies
22
Views
4,576
Hello, I am doing research project focused on the retrofit of a diesel locomotive into a hydrogen-fuel-cell-battery hybrid locomotive. The...
Replies
7
Views
510
Good morning. I'm looking at switching out an older 1794-L34 FlexLogix PLC. My plan was to take out the FlexLogix PLC and add a 1794-AENT then add...
Replies
4
Views
517
I have a Generator Control Module which is designed for D399 Caterpillar engine for understanding this can be considered as the replacement of the...
Replies
0
Views
1,664
I'm looking for opinions on retrofitting a 3-axis gantry robot. Currently the controller is Keba (codesys system out of Austria), talking to 3...
Replies
4
Views
2,194
Back
Top Bottom