On Friday, May 27, 2016 at 5:27:48 PM UTC-4, Tim Wescott wrote:
On Fri, 27 May 2016 08:50:26 -0700, rangerssuck wrote:
On Friday, May 27, 2016 at 11:27:39 AM UTC-4, Tim Wescott wrote:
On Thu, 26 May 2016 19:43:33 -0700, rangerssuck wrote:
A customer has a very large, very old (and pretty dirty) machine that
needs some help. It is a multi station drilling and tapping machine.
The production items sit on 4" square pallets spaced roughly 8"
center to center. There is a microswitch that is activated by contact
with the leading edge of a pallet. The switch causes the drive motor
to stop, and the drilling & tapping to begin.
Their problem is that the stopping point is somewhat variable, and
they thought that if I replace the microswitch with a prox sensor,
all would be well. So, I went down there and had a look, and found a
good place to mount the prox sensor and a brush to sweep off whatever
debris was there (and maybe an air jet for good luck). Then, I had a
talk with the plant electrician.
The electrician, who really knows his stuff, pointed out that the
controls stop the motor, but it's really a hydraulic motor that
drives the system, and maybe it coasts a little when the electric
motor is shut off. Also, he pointed out, that the offset changes
during the day, maybe as a result of the fluid heating up. So, my
not-at-all educated in hydraulic motors mind started thinking, what
if, instead of shutting of the pump, what if we installed, right at
the motor, the hydraulic equivalent of a DPDT relay that would
disconnect the motor from the pump and put a short across the motor?
Does that make sense, would it (as would a DC motor) stop on a dime?
I have googled a bit and I believe I have seen such circuits, but I'm
really pretty much in the dark about such stuff. If it DOES work, do
I have to worry about overpressure in the lines? I haven't seen the
machine run, but I'm told that a) it doesn't move fast and b) it
doesn't move far.
The pump motor is 3-phase, so I suppose I could install a VFD (they
probably have extras on site) and use a second prox sensor to cause
the speed to ramp down, but it would be better if I could just stop
the motor. And, there really isn't a good place to mount a magnetic
brake.
As always, all thoughts (especially kind ones) are welcome.
Before I did anything else I would ask both myself and any old-timers
if the machine used to work better. It might even be worth a visit to
the retirement home of a former millwright or two.
Then I would check for signs that something on the table is there but
broken, or has been removed.
It may be that the only thing the "idiots" who designed it did wrong
were to retire and leave it in the hands of _real_ idiots.
The only way you're going to get the thing to have precision in line
with modern machine tools is to either have some positive closed-loop
control of position, no doubt with ball screws and motors and whatnot
involved, or to have some fixed positive indexing (like the "shot pin"
mentioned elsewhere).
Deciding to revamp it with electronic controls may just be jumping onto
a slippery slope leading to replacing the whole damned thing.
For the fixed indexing, you probably want a pin or whatnot to establish
the correct location and then something that holds tension against the
pin (else the whole pallet will be short of the pin, and your precision
will have gone out the window).
Sometimes the quickest and most reliable super-high-performance
modification to an older machine is to put it back the way it was
originally designed.
--
Tim Wescott Control systems, embedded software and circuit design I'm
looking for work! See my website if you're interested
http://www.wescottdesign.com
OK - in no particular order: Iggy's idea of an indexing table would be
fine except that it would require rebuilding the entire machine which is
about 20 feet long, six feed wide and six feet high. It simply isn't
happening.
There is pin indexing, but the current stopping position is so bad that
the pins are totally misaligned with the holes. If I can just get it to
stop close enough for the pins to go into their holes, everyone's going
to be happy.
I am told by the machine operator who has been there 15 years that he
has never seen it work reliably without playing with the microswitch
position, and then it goes out of whack again after some (I'm not sure
how much) running time. I AM sure that he told me the overshoot is worse
in the afternoon than it is in the morning, and he and the electrician
are pretty certain that's due to the oil temperature.
Of course there could be any number of other things wrong - there could
be leaky valves, worn pump, worn motor, sticky relays, who knows what?
But I need to start somewhere, and this looks like a pretty simple thing
to set up.
For test purposes, all we have to do is hook up the hydraulics - the
hydraulic equivalent of a 4PDT switch thusly:
In position A (normal running position) :
Connect pump line 1 to motor line 1 Connect pump line 2 to motor
line 2
In position B (braking position):
Connect pump line 1 to pump line 2 Connect motor line 1 to motor
line 2
along with whatever relief valves or shock absorbers the supplier
recommends and hook it up to the existing microswitch (probably through
a relay). If it works, great. If not, it will have cost a couple of
bucks for the valves and a couple of hours to hook it up.
I know I sound like I'm clinging to an idea, but -- maybe it broke
_before_ that operator got there? Lots can happen in 35 years.
That is certainly true. Unfortunately, the plant has been through a few owners over the years and documentation has been lax or lost entirely. I haven't seen this machine run, nor have I seen prints for it - I don't even know that they exist. When I get back there, I will, for sure, pull off a bunch of covers and trace out the hydraulic circuit. I really have no idea what I might find. It's possible that the work I plan to do has already been done and, as Roger suggested, there are stuck or leaky valves. We'll see next week.
Of course, the idea that wear has made it less accurate is quite
believable: one of the jobs of a _good_ control systems engineer is to
anticipate what happens as a machine ages, and compensate for it. Not
everyone does, or they figure that if it lasts for five years that it's
paid the bills and can be replaced or upgraded.
Remember, this is a pre-PLC, pre-anything modern contraption. It is entirely possible that I'll find excessive wear in moving parts (chain drive, by the way) and will have to go back to the idea of an "almost there" sensor to trigger a slow approach and then a "there" sensor to stop. On the face of it, I have no problem with that, and it may be the easier way overall - all electronics, no oil. Right in my wheelhouse.
Maybe continue to use the microswitch for actuation, but connect it to a
solenoid valve? So hitting the switch simultaneously cuts power to the
motor and brakes the hydraulic system?
I certainly could do that, but I don't know what the benefit would be to cutting the motor power. It's certainly easy enough to do, though, since the controls are already in place.
With the caveats about fluid hammer that have been mentioned by people
who know much more than I about hydraulics, of course.
And that's why I asked here. Even if you know next to nothing about hydraulics, you still know more than I do :-)
It sounds like you just need to get the accuracy to the point where the
existing pins will engage, rather than +/- 0.001", which is nice.
That appears to be correct.
--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com