Hydraulic drive is not repeatable

B

Thread Starter

Bill Code

My company is involved in a debugging project on a:

Cut to length system producing 10 ft to 20 ft 'sawlogs' from a 'stem'. A stem is what arrives on a logging truck (most people would call them logs) and is up to 80 ft long.

The stem is carried through an arrangement of conveyors and two spiked rolls. The spiked rolls are driven by hydraulic motors. One roll is fitted with a home brew encoder: two inductive proximity switches sensing lobes on two
steel discs. Quadrature signal derives from angle of one disc with respect to the other. The encoder is mounted on the high speed side of a gearbox, between hydraulic motors and gearbox. The encoder provides position feedback. A PLC closes the position loop. The PLC has a VHSC module that counts the encoder signal. And a simple proportional control ramps the velocity down to a stop at the requested sawlog length. The proportional control is implemented via math blocks in ladder logic. The hydraulic drive is throttled by an analog output from the PLC that signals two Rexroth VT cards. The VT card provides
some kind of PWM signal to proportional valve coils controlling oil flow to the motors.

My company did not supply machinery or hydraulics. Nor did we supply the home brew encoder. But we have agreed to do what we can to improve the accuracy and availability of the equipment. I was recently at the plant and witnessed seven hours of down time over two shifts while maintenance staff adjusted the proximity switches to bring the quadrature signal into the specs required by the VHSC module. That could see the signal on an o-scope and it was having phase problems causing the VHSC to fail to count.

I thought I would ask the members of the list:

Is anyone aware of an encoder which will survive high (exactly how high is unknown) G forces that exist as the rolls contact the stem? If no device meeting the requirements exists, then I will rebuild the home brew encoder to add an adjustment to directly affect the phase of the two proximity switch signals. The duty cycle is adjusted via the air gap between the proximity switch faces and the discs. In the present configuration, the phase is adjusted by playing
with the duty cycles.

Is anyone aware of a 'motion controller' or 'position controller' that can accept a quadrature encoder signal and provide output to Rexroth VT cards or Rexroth proportional valves directly? Ideally, it can also accept an 'at position' signal. My customer has installed photo-electrics at the sawlog length positions to cut
the signal to the proportional valves in an attempt to mitigate the accuracy problems. The motion controller would be under the supervisory control of the PLC in that it would receive 'requested positions' from the PLC and return a 'position done' signal to the PLC.

Any response is greatly appreciated.

Bill Code
MPM Engineering Ltd.
4-6240 202nd St., Langley, B.C., Canada, V2Y-1N2
Phone: 604-534-6605 Fax: 604-534-6693
E-Mail: [email protected] WWW: http://WWW.MPM-ENG.COM
 
L

Leighton Toews

I do process control primarily in the forest sector in BC Canada. This kind of solution is very typical of sawmills. There is a very good quad encoder available which I use in conjunction with the VHSC all the time. It is manufactured by Danaher Controls, the brand is Dynapar. They have a steel disc instead of glass, very robust and you can get a flex coupling to join the shaft of the encoder to the device, although i just use a short length of rubber hose, which is cheaper and seems to be more tolerant of shaft misalignment which is the primary problem with an encoder solution. The part number is H231000104244. As for the accuracy problems, this is an inherant problem with the AB (or most other) PLC's due to scan time. The only way we have been able to get around the accuracy problem with encoder positioning in this type of application is by putting the VHSC completely in
charge of driving all the positioning devices through its own outputs. This seperates the encoder from the PLC scan time, unfortunately the VHSC is not capable of proportional control. One idea may be to replace the "math rungs" with PID instructions using the encoder count as your PV. In this application that may be enough.

Leighton Toews
Automation Technologist
Deltech Industries Inc.
http://www.deltech.ca
Corporation.
 
Hi Leighton,

Thank you for your response.

Have you ever used a Dynapar HS35 Hollowshaft Encoder? The specs indicate it is able to operate over a wide temperature range. It has a metal disc and it can slide over, and clamp to the shaft between hydra motor and gearbox. I am concerned about it's ability to withstand shock.

Bill Code
MPM Engineering Ltd.
4-6240 202nd St., Langley, B.C., Canada, V2Y-1N2
Phone: 604-534-6605 Fax: 604-534-6693
E-Mail: [email protected] WWW: http://WWW.MPM-ENG.COM
 
M
I'd use a Delta Tau PMAC motion contoller with their hydraulic servo accessory. I have successfully replaced an AB PLC5, and Rexroth valve electronics with the PMAC and hydraulic servo amp. Your PLC would communicate with the PMAC to command lengths and then monitor the PMAC to see when the length was cut. The output from the amplifier would directly drive the proprotional valve.

Since you describe the valve as a proportional valve and not a servo valve you need to be aware that it doesn't have the high bandwidth of a servo. I doubt that makes a big difference in this application.

For an encoder you could try an Encoder Products heavy-duty series encoder, but you'll still need to isolate the encoder shaft from the side loads
imposed by the stems.

Mark
 
M

Mike Asztalos

From what I read in your email, it sounds like your problem has nothing to do with your subject. It sounds like this motion system is just not robust... Some things that I would need to know in order to think of a solution, is this proximity switch encoder mounted on the input or the output of the gearbox? You say the high speed side, I can only assume that you mean the input. If you want accuracy, you should mount the encoder on the end product, in this case the
spiked roller. If you want a tip, loose the AB
counter and analog output, this system has a low
bandwidth. I am sure you can find some servo
controller that can give you what you want, I know
Reliance Electrocraft's controllers have encoder
inputs and +/-10V outputs. Also AB has motion cards for this also. The IMC line will do this.

Here is a suggestion.. Why not use optics instead of inductance (proxes) for position feed back. You can make a wheel with holes in it and then have a fiber optic photoeye look thru the holes to determine your position. I'm sure this would with stand the G forces.

Good Luck,
Mike Asztalos
Sr. Control Engineer.
 
Bill,

We have done this exact application for Welco Lumber and Willamette industries both in ACVector versions and servo hydraulic versions. I would be happy to discuss your app with you. I make it up to Sedro Woolley and Bellingham once a month or so and would be happy to drive to Langley to go over the application with you.

As far as a robust encoder is concerned, Northstar (also known as Lakeshore) makes a magneto restrictive encoder that we have used in this environment with good success. We have also put them on sharp chain drives feeding shape saw systems that are hosed down twice a day and have had no failures in 2 years. The encoders put out 1024 quadrature counts per rev. (4096 counts / rev when decoded on each encoder edge)

As far as motion controllers go, we have used the Delta Tau PMAC for this application, but it depends a lot on what kind of PLC you are using
as far as connectivity is concerned. I assume you are scanning the log and optimizing for lengths with the PLC and could pass the lengths to
the motion controller via the PLC network? The PMAC can talk serially or Profibus or DeviceNet or DH+ or any number of other protocols. It can drive the Rexroth amplifier cards directly.

In the applications we have done in the past, we have done soft-stop bucking lines that hold about 1/4 inch accuracy depending on the fiber and the triggering prox switch. These systems run 8000+ lb logs at speeds to 450 fpm.

Again, I would be happy to chat with you about your needs and help in any way we can.

Best Regards,
Ken Brown
Applied Motion Systems, Inc.

Voice 360 686 1133
Fax 360 686 1166
 
J
I have been successfully applying the Giddings and Lewis PLC/Motion Controller in cut to length applications for a number of years. The current
G&L family of controllers includes a 2-1/2 axis MMC controller that will accept an encoder signal and supply the +/- 10 VDC reference signal to the
VT valve drivers. The algorithms for just about any type of motion control you need to perform are built in to this controller. As for communicating with the AB PLC, if your CPU is one of the models with an RS232 port, G&L has a function block for communicating using the DF1 protocol.

I beleive this will help, but it sounds like you need to solve a few other problems before you can operate reliably. Adding a good motion controller to your existing system would be like buying a Mercedes and using the old retread tires.
 
L

Leighton Toews

No, I haven't used that model of encoder, The series H20 (which is the one i speak of) is rated to 50 g's for a duration of 11 msec and can withstand vibrations of 5 - 2000 hz at 20 g's. I have to wonder, though, if perhaps it is possible to move the point of measurement? Say for instance to the motor shaft itself, is there a spot which does not receive the impact?

Leighton Toews
Automation Technologist
Deltech Industries Inc.
http://www.deltech.ca
 
C

chris wright

Hi bill
without plugging our company too much
we produce encoders that use magnetoresistive technology, they are more durable and resistant than optical. that could be a solution. check us out on "www.northstar-tec.com"
hope this helps
chris wright
northstar technologies
 
Top