Better Depth Control on Hydraulic Press

[INNJunEER]

Hello,

I need to come up with a way to control a very large press to get precise depth control. I need a resolution of 1/1000th of an inch and if not to $$ , 1/10kth of an inch. I was looking at an absolute encoder but I would like to explore other paths. The press is very old ~50 years so the controls on it are nil to none. We did add a Rockford safety device but thats about it.
I need to control har far down the press will go (I will have a block back up to prevent die damage). It seems that the full travel of the press is 3 feet but in use 2 feet would be the max I would think. Any ideas?

 

[unclesyd]

Here is one relatively inexpensive digital readout that is good to 0.0005 with no problem. Though I haven’t operated with one myself a friend has two in his shop and is extremely satisfied. If you are able to physically mount this one it would be the one to try.

http://www.star-techno.com/info.htm

There quite a few optical encoders available at Mill Supply Houses.

 

[INNJunEER]

Thanks a bunch. I will check it out and see. I do not think that mounting will be an issue, its a huge press.

 

[INNJunEER]

Well I checked into it and there is concern with it not being sealed. So I have moved on to looking into using a Sealed Absolute Linear Encoder of some type.

Checking out Stegman and Heidenhain, thing is even if I get the control precision I need, how inthe heck am I going to control it? These presses are 50 years old and there are no "controls" on it I dont think.

If I get a good encoder, what price range would I be looking at to control the thing. hence stop at a certain depth etc.

 

[MintJulep]

It seems unlikely to me that an hydraulic system will ever be able to achieve the type of percision you seek, regardless of the accuracy or precision of the measurement device.

There are just way too many variables in the control loop.

I would think about using a hard-stop. Set up a controller system that will slow the press way down when it gets close, and then reverse it when the position sensor no longer shows any movement.

 

[INNJunEER]

That makes sense, I was wondering how in the heck I would CONTROL the press. At present they use blocks to prevent the press from comming down to far but the control to depth was hard for me to see how, especially in an economical way. I will check into this and see just how important this level of control is. If it is not then a simple limit switch would stop the press travel I hope.

Thanks

 

[unclesyd]

At one time they made a hydraulic valve that had two speeds. We call them creep feed valves. We used them on punch presses where we had a slow travel and then high flow for the piercing process. We had the valves modified for this.
Could you use two valves? One high flow and the other low flow for the positioning. They make this combination of valves.

Please describe a little more about the problem with the Star product if possible.

 

[unclesyd]

Take a look around this site for a possible solution to location. We used a lot of their devices for on the fly inspection machines. The devices are relatively inexpensive and extremely rugged. The people we worked with were all quite competent and extremely helpful.

http://www.keyence.com

 

[funnelguy]

The press is hydraulic? What is the max speed of the press ram? Hydraulic flow at max ram speed? Max pressure? Are there process limitations that would preclude creeping to the final depth?

A simple lever type multiplier may work in conjunction with a suitable limit switch to provide adequate resolution. The signal from the switch would be used to unload the press ram and begin retracting the ram. Assuming that there's no need for a dwell....

Alternatively, contact a builder with experience in closed loop servo hydraulic control?

What is the problem with advancing against the hard stops your shop is currently using?

 

[Barry1961]

You can buy proportional valves from a number of different suppliers which should be able to handle the job depending on the speed you need. You will probably need to creep to the final position.
I have used several Parkers D1FX---- valves in the past that worked well and had all the electronics on self contained on the valve.
I think the most common method is to make an adjustable ram head. This allows the cylinder to stroke full length or against a stop then adjust the ram up or down with a screw or screws. You should be able to some graduated nuts for this purpose for a std. 16tpi thread. This allows you to stroke as fast as possible.

Barry1961

 

[INNJunEER]

OK in discussions with the folks who want this done, I have found out that they do not need to control to such accuracy. What they need is if the "head thickness" of the part is 1/10,000 off, they would like to dial in a compensation factor for the press to stop sooner etc. There are limit switchs and the press is already running at slow speed. I will need to get more info to move on but today is out, I have a project that I need to get done. I really appreciate the help in here folks!

 

[griffengm]

INNjunEEr,
If you have three layers of solid steel(ram, platen, die)bolted together and hanging from the cylinders, there will be significantly more than 1/10,000 compression at the end of each stroke.
Sounds like time to shoot engineers and git on with production.

Griffy

 

[INNJunEER]

Well if I now understand what they want correctly, I think we can do it and get some repeatablilty out of it. But it is on the backburner for now, to busy with other items....

 

[automatic2]

I couldn't see where INNJunEER indicated the press is hydraulic. Regardless, the only way you could control with any accuracy is within the die/anvil fixture.

 

[INNJunEER]

It is a hydraulic press and I know squat about them. It seems that we just need to get repeatability in the distance the press travels. I will gather more info today I hope.

Thanks

 

[automatic2]

And it says so in the tittle. You will not be able to control the repeated positioning of the piston/rod within such accuracies. The only way is in the die fixture. An encoder or any other device will not work due to latency in the valving. The only way that could come close is to drive the cylinder to the end of it's stroke then adjust the die/fixture, or, if you could accurately control the volume of fluid delivered to the cylinder, you could exercise some degree of positioning. Using a pump cylinder to provide the volume would give a degree of accuracey as you could worm drive that cylinder using an encoder on the rotary component. A mathematical ratio could further your needs. Sealing element deflections would play a role at your indicated resolutions.

 

[EdDanzer]

To obtain maximum position resolution an absolute linear encoder connected to a motion control device that drives a servo valve for finish positioning can do what you need. It will cost several thousand dollars for all the parts and may require a separate hydraulic pump to supply oil for the servo valve to reduce cost and heat build up. Newall and Sony make these type of scales, Galil would be a simple motion controller to set up, and flow volume will determine valve suppliers.

 

[INNJunEER]

Well I just found out that these units (press's) have SLC500's. Makes things a bit easier to think about solutions. :0

 

[griffengm]

INNJunEER,
At this point in this discussion, I'm guessing that what you are doing on this press is some type of forging or cold heading operation and your designers are trying to get as close to net form as possible to eliminate some secondary operation.
If this is true, then you are chasing a will-o-the-wisp. Thermal issues alone will drive the operator crazy whether cold or hot.
The best thing you can do if you succeed in mounting such a control is to chain the designer to the machine until he can either run the process reliably to the specs (six sigma) or gives up.
'Nuf said.

Griffy