Looking for a strategy to decelerate a hydraulic cylinder

Tubularfab

Member
Join Date
Jan 2012
Location
Atlanta
Posts
37
Ok, first of all I'm completely new to PLC's, but at least have a basic understanding and a computer programming history. Right now I'm looking for ideas on the best (economical) approach to a problem. Here is the machine info:
It's a home built mandrel tube bender. 20hp motor driving a large hydraulic pump. 6 directional control valves, 5 separate cylinders. Direct logic 205 Plc controlling it all. The main bend cylinder is actually a dual piston design with 4 ports so it can either run fast as a single cylinder or get twice the force as a dual cylinder.

Here's the issue. The bend cylinder speed is adjusted via a flow control valve on the "b" port. The problem is that at normal bend speeds the delay from the Plc hitting the set point to the bend arm stopping equates to up to 3 degrees, and it's completely dependent on the temp of the oil and the bend speed. This is a big problem when trying to hold fractions of a degree tolerance.

I want to devise a way to slow it down a couple of degrees before the setpoint then creep up on it. One idea I had was adding a second flow valve in series with the existing one that also has a normally open solenoid in parallel with it. At bend angle minus a couple degrees close the new solenoid forcing he new flo valve to take over at a much reduced setting. I've also thought about adding some kind of actuator to the existing flow control that the Plc could engage to slow it down.

What would the most elegant way to do this be? There is a tight budget here... I assume all of the digital / proportion controls would be way too expensive?

Thanks for any suggestions!
 
Check and see if you can get mechanical actuated deceleration valve. Otherwise two solenoids both open for high speed close one for slow and the other has a throttle valve in it to set speed. Just like you were thinking

Dan Bentler
 
Here's the issue. The bend cylinder speed is adjusted via a flow control valve on the "b" port. The problem is that at normal bend speeds the delay from the Plc hitting the set point to the bend arm stopping equates to up to 3 degrees, and it's completely dependent on the temp of the oil and the bend speed.

What controls the setpoint? Is it a encoder, or limit switch, temposonic?

It's a home built mandrel tube bender.

This would imply that it does not have a tempo.

This is a big problem when trying to hold fractions of a degree tolerance.

Your forum "name" implies you are in the fabrication business. I'm guessing you are recognizing a problem with your equipment and are looking for a solution.


I want to devise a way to slow it down a couple of degrees before the setpoint then creep up on it.

I'm also guessing that you have discovered that slowing down the bend speed solves the accuracy issue but the speed that is required is not practical in a business sense, correct? If this not the case can you elaborate?

What would the most elegant way to do this be?

A Delta Computer systems RMC controller and temposonic cylinder, and servo valve. But this will be in conflict with this next statement

There is a tight budget here

In the end it will boil down to a decision between production speed, cost of defective materials, and equipment investment.

I'm sure that the title of this thread will attract Peter.
 
Thanks for the input guys!

Dan - I looked up the decel valve you mentioned, but they appear to be a mechanical arrangement. I am looking for an electronic option as the setpoint needs to be controlled by the PLC.

Milldrone - you are correct that I am trying to work around deficiencies in an existing machine. I have a tube fabrication business plus an engineering background. For a little more background on the machine, the guy that built it was an engineer for Pines (company thant pretty much invented and perfected mandrel benders) way back. He built this one for himself at home based mostly on the design they built but with a Tim Allen more power and more options mentality. It's a beast... At the time he relied heavily on sequence valves and an old drum relay to control it. In the early 2000's he went through it giving it a good refresh and swapping in the PLC. The annoying thing is he just mimicked the relay and did not take full advantage of the PLC. The bend angle setting was determined by the length of adjustable rods hitting a limit switch attached to the shaft of the bend cylinder. It was a real pain to dial in the bend angle with any precision.

So, when I bought the machine from him I added a quadrature encoder to the bend spindle and a red lion controls gemini 1000 counter. That let me digitally set the bend angle (and read it) to about a 1/4 degree accuracy (using a 360 div encoder in 4x mode). I then simply used the relay in the counter to replace the old limit switch.

Yes, slowing it way down solves the accuracy but makes it way too slow for practical use. I want it to run fast up to a couple of degrees before the setpoint, then slow down to the stop.

The cylinder this control is working on is fairly complicated. It's a 5" bore with 2 pistons inline and a seal plate in between them. One end has a much bigger shaft diameter so it can move much faster on the return stroke after the bend is made. The shaft comes out of both ends and attaches to a chain that loops around the spindle, back to the opposite end of the machine, around an idler, and to the opposite end of the cylinder shaft. It usually uses just one piston so the speed is good, but for really heavy bending the controls switch so pressure is routed to both pistons, thus doubling the force and halving the speed. I'll attach some pics in a minute...
 
Last edited:
Here is a pic of the bend cylinder and a link to the machine running some really small tube on youtube

http://www.youtube.com/watch?v=pfzVEb2Mhkk

and

http://www.youtube.com/watch?v=MALs4XURgOI

In the video it's running about 1/2 it's max bend speed. You can see the bend arm "snap" as it overruns the stop in the video. At that speed I have the stop set about 2.5 degrees below the actual target bend angle. Also in the video you can see the drive chain that attaches to one end of the cylinder shaft and wraps around a sprocket on the bend spindle.

photo.JPG
 
So, when I bought the machine from him I added a quadrature encoder to the bend spindle and a red lion controls gemini 1000 counter. That let me digitally set the bend angle (and read it) to about a 1/4 degree accuracy (using a 360 div encoder in 4x mode). I then simply used the relay in the counter to replace the old limit switch.

Yes, slowing it way down solves the accuracy but makes it way too slow for practical use. I want it to run fast up to a couple of degrees before the setpoint, then slow down to the stop.

The first gut response is that you will need a encoder with more pulses per revolution. The next gut response is that the encoder needs to be connected to a controller that has control over the cylinder speed. Not a controller that just indicates when the setpoint has been reached.

In the video it's running about 1/2 it's max bend speed. You can see the bend arm "snap" as it overruns the stop in the video.

Do you ever run it at full speed?

If not, you mentioned the diameter at 5 inches what are the shaft sizes ( I'm looking for displacement) I do not have a stopwatch with me, how long does it take for a "1/2 speed" bend and how long for a "good controlled speed" bend?

You mentioned earlier a 20 hp motor and pump. What kind of pump? Pressure compensated, fixed displacement? Do you have an accumulator? Does the PLC have any empty slots?
 
Last edited:
and it's completely dependent on the temp of the oil and the bend speed.
There is a tight budget here...
The annoying thing is he just mimicked the relay and did not take full advantage of the PLC.
These are the 3 key factors. Take full advantage of the PLC. Use PLC abilities to offset the dependency on the temperature of the oil and the bend speed.

The cheap way is to create a PLC internal table that calculates the Bend Stop point based on speed and oil temperature. The PLC already knows the speed, and the oil temperature can be easily measured and put in as an analog input. Now look at these two factors, do a table look-up to find how much to subtract from the normal bend stop to cause the bender to stop at the right place.

This look-up table can be created by making many trial runs and measuring the results. It may be that the relationship is a straight-line function. If so, you will not need a table, but only a formula that calculates the amount to subtract from the bend stop point.
 
Last edited:
So, when I bought the machine from him I added a quadrature encoder to the bend spindle and a red lion controls gemini 1000 counter. That let me digitally set the bend angle (and read it) to about a 1/4 degree accuracy (using a 360 div encoder in 4x mode). I then simply used the relay in the counter to replace the old limit switch.

.

but only a formula that calculates the amount to subtract from the bend stop point.

Lancie,

I'm thinking the encoder is not hooked up to the PLC.
 
I'm thinking the encoder is not hooked up to the PLC.
You are probably correct. An encoder reading in the PLC would make things easier. However, it may still be possible to calculate the stop point without an encoder.
So, when I bought the machine from him I added a quadrature encoder to the bend spindle and a red lion controls gemini 1000 counter. That let me digitally set the bend angle (and read it) to about a 1/4 degree accuracy (using a 360 div encoder in 4x mode).
If the Bend Angle can be sent to the PLC as an anlog input (or a digital number), then the PLC can calculate a reduced angle based on the two dependent factors, and send that new value to the bend angle controller.

Another method is to move the calculated offset based on speed and temperature to the front end by adjusting the bend point where the pipe bend is first started. Here is a spreadsheet that I did years ago to calculate the Bend Point on a pipe bender. This Bend Point calculation could be adjusted to include factors for temperature and speed.
 
Last edited:
From photo and videos I get the idea
1. This is essentially a job shop - no large production runs
2. Therefore quick change setup is mandatory
3. My idea of a mechanical decel valve is out the window - just wont work in this setup given #2.
4. All hydraulic systems leak and hydraulic fluid ALWAYS goes down - something to do with gravity I think. I also theorize from submarine service that like water it goes anywhere it likes - more so if you do not want it there.

I think
1. Sensing the actual bender (rotating part) "degrees" is optimum provided you can keep the sensor out of the way ie below the working surface
2. Sensor must be hydraulic proof (not leak tite just able to tolerate hydraulic fluid)
3. If you can get 1/4 degree accuracy on display AND PLC is fast enough then you should easily be able to control to 1 degree.
4. Given you can measure to 1/4 degree I would not bother to figure out oil temperature - yes it is a factor in speed - but with the resolution you have it should not be a problem.
5. The major obstacle I see now is to find a fast acting and ideally a throttling valve to slow unit or use a solenoid (or multiples) with either manual set throttle valves or orifice plates to slow down that last little bit to turtle speed. I like the multiple valve and orifice plates (or throttle valves) because
1. that is what I am most familier with
2. I can get ASCO valves off the counter across town in a couple hours.

Nice looking tube bending on machine by the way

Dan Bentler
 
Thanks again for all the input!

Yes, it is a small jobshop, and my niche is specializing in small runs. The job I was running in that video above was 150pcs and it comes up every few months. In fact, it was part of my motivation to make improvements. If you watch the video you'll notice I'm adding a bend to atube that already has one bend in it. When using the redlion control entering the bend angle one digit at a time is slow - to the point that it was faster to make one bend on all of the parts, then go back and dial it in for the other. It just isn't efficient to hit the button for each decimal place up to 10 times each bend.

So that brings me to a second part of my plan. I intend to do away with the redlion counter and connect the encoder straight to the plc. The encoder is not visible in the pics of the video, but its directly under the bend spindle. It's out of the way as far as the working side of the machine, and I made a deflector shield over it to make sure no oil found it's way in. It appears that I have 2 options for connecting the encoder to the DL205 (240 processor, btw) - an H2-CTRIO or the D2-CTRINT. From my reading on the modules it sounds like the CTRIO is the easier one to work with? Plus, the CTRINT won't do 4X counting on a 240 processor... Anyone with experience feel free to share info on one versus the other... I also have an optimate display / keypad to add to the mix that I picked up with an entire spare matching plc a while back. Again - very green on PLCs, so I figured I would start by setting up the second PLC as a testbed to read the encoder and integrate the HMI without messing up the working one in the bender right now. Then, once I get that part working I could incorporate it into the PLC on the machine. My goal is to have a way to enter up to 10 bend angles into memory. The first bend cycle would go to the first angle set point, then pop to the next setpoint on it's own for the next bend. If it had a zero for the setpoint it would reset back to the start of the list.

I'm thinking the dual flow controls with an extra solenoid would be the way to go. Have the PLC close the solenoid at "bend angle minus 3 degrees" which forces the flow through the almost closed flow control so it creeps up on the setpoint. I just wondered if there were any cooler, more elegant ways to do it from the opinions of guys that have seen way more machine controls than I.

Lancie - that's an interesting approach to the problem - multiple ways to get it done. I never would have thought of that one, though. The one thing I didn't mention is the bend force required also has a big factor in the bend speed, and therefore the error angle. You would have to calibrate the formula or data points for each tube size as well as the temp.
 
It appears that I have 2 options for connecting the encoder to the DL205 (240 processor, btw) - an H2-CTRIO or the D2-CTRINT. From my reading on the modules it sounds like the CTRIO is the easier one to work with?

The H2-CTRIO is the preferred unit. The D2-CTRINT is nearly obsolete.

But before a H2-CTRIO can be considered, a space in the backplane needs to be available. Is there a available space?

When using the redlion control entering the bend angle one digit at a time is slow - to the point that it was faster to make one bend on all of the parts, then go back and dial it in for the other. It just isn't efficient to hit the button for each decimal place up to 10 times each bend.

I also have an optimate display / keypad to add to the mix

Presuming you have a way to deal with aligning the start point of each bend and if you have any spare inputs they can be configured along with the optimate to give a discrete bend for each input (pushbutton). I'm recommending using a separate push button not the keypad on the optimate for this. The reason for this is if a button fails from use (abuse) it can be replaced on an individual basis. If a keypad button fails it's a different story.

I'm thinking the dual flow controls with an extra solenoid would be the way to go.

We need to know if you have any outputs that are unused. Because you would need one for this.
 
Last edited:
All the right people are responding since I have no experience in hydraulic systems.

My question to everyone is can this (slow speed) be accomplished with a VFD and lowering the RPM of the pump? I realize that the OP probably does not have a VFD on the motor and that a 20 HP VFD would be a couple of thousand $, but would that concept work?
 
My question to everyone is can this (slow speed) be accomplished with a VFD and lowering the RPM of the pump? I realize that the OP probably does not have a VFD on the motor and that a 20 HP VFD would be a couple of thousand $, but would that concept work?

Yes it can, presuming the accuracy and speed demands are within a relatively low area of expectations. I would class the OP's situation in this area.

I have also done this with a 20hp pump (constant displacement with an unloader I know, not my design) and a 1hp pump.
 

Similar Topics

I downloaded v24 for studio 5000 but can’t find where the download manager put it! Any help? I’ve done it before but can’t remember. Thanks
Replies
7
Views
200
Good Afternoon , We all must of done this before ...... Installing a Frequency Drive in an enclosure , you forget how deep your selector...
Replies
1
Views
103
Problem: Our PLC can only output 4-20mA, but the actuators it needs to control, modulate based on a 0-135Ohm signal. Buying 4 or 8 individual...
Replies
7
Views
219
Good Morning , Not sure how many use Markem-Imaje SmartDate X60 Printers , but I'm having trouble finding a good manual for...
Replies
3
Views
353
Would anyone know what software is needed to program a SAIA PCD1 controller? Or know where to find it online?
Replies
1
Views
116
Back
Top Bottom