Self-excited oscillations, caused by variable displacement
hydraulic pump and motor controls are the one topic I could discuss for
hours, and I surely will devote a number articles to this very
interesting, but also extremely extensive subject. Not the theory and
math, but rather some (the key word is some) practical stuff
taken from my personal experience, like what can be attempted to "fix"
them controls when they go swingy on you.
I would like to begin the first chapter of my self-oscillating chatter with focusing on the control fluid leakage as a factor, that can condition self-oscillations in pump and motor displacement controls. By control fluid I mean - hydraulic fluid, that is directed from a control spool to the servo-cylinder, and by the leakage I mean the controlled (with a calibrated orifice) leak of this fluid to tank (or pump case).
You will find this in many pump control
assemblies - that tiny orifice that creates a leak in the servo-pressure
line. It is there for a reason, in case you wondered, for several
reasons actually. For one - the leakage flow helps cooling the pump,
especially in stand-by or dead-head modes. But there is another benefit
to its presence - the control fluid leakage can often resolve control
instability issues, because it shifts the working position of the
control spool to a more stable zone, since it "forces" the spool to
supply more oil. The drawback of the efficiency losses due to the added
leakage comes with the benefit of a stable behavior.
From theory to practice! Below IH is
presenting a practical example of how controlling the amount of
servo-fluid leakage solved the problem of unwanted self-oscillation of
high pressure cut-off control in a Rexroth A7VO open circuit pump
(extensively used and reasonably warn-out). I also created a video that
contains the pressure readings at different stages of the "experiment",
but you should read the description before watching it, otherwise it
won't make much sense.
The problem - Rexroth A7VO55LR3E/61L open circuit
pump, equipped with torque limiter and high pressure cut off control,
both with hydraulic overrides (I placed a detailed description of this model in the back-engineering section, also you can find the schematics and the detailed cut view
of the main control spool to the left), was presenting self-oscillating
behavior in the dead-head condition, when the cut-off was kicking in
(seen at the beginning of the video). The self-oscillation was present
both with and without the cut-off setting override pilot. In the video
the 400 bar pressure gauge is connected to the A1 port (the outlet
pressure), and the 250 bar gauge is connected to the X3 port
(servo-pressure).
In such situations many people assume that these
self-sustained oscillations can be fixed by means of
strangulation/dampening of the control flow with a "strategically
placed"orifice, but despite seeming like a perfectly logical next step,
it often doesn't work or even brings undesirable consequences - like the
increase of the control response time. This pump actually has a thread
in the control fluid passage (cut-view),
so this theory was easy to test. Placing a 0.8mm orifice in this
passage didn't solve the self-oscillation issue, and also resulted in an
unacceptably slow response time of the cut-off control, which can be
seen by the 350 bar pressure peak, easily visible even on the analogical
pressure gauge in the second part of the video.
If you check the schematics,
you'll see that there is a small orifice in the T1 port, that provides
the control fluid leakage to an external tank line. Since these pumps
have no drain line (the case is internally connected to the suction
line) - the first thing that comes in mind is - it is there to allow the
pump to dead-head for longer periods of time before overheating like
hell, as the T1 flow carries away some of the heat, generated by the
pump. However, by looking at the schematics, one might assume that,
aside from the overheating rate, this control should behave in the
same manner with the T1 plugged. Well, if you try it - you will
see that the cut-off does cut-off, but it self-oscillates even more
violently than before. (seen in the third part of the video)
What we got so far is - no control fluid leakage -
violent outlet pressure self-oscillation, some control fluid leakage -
less self-oscillation. Let us play around this tendency a little,
shall we? Let us take out the 0.8mm orifice the T1 port was carrying,
and connect it to the tank line via an adjustable restrictor valve to be
able to infinitely control the control fluid leakage. The result can be
seen in the fourth part of the video - two things happen as the rate of
the leakage increases - the dead head pressure rises (slightly), and
the behavior of the control becomes more and more stable, with the
self-oscillations eventually disappearing.
So, we are finally getting somewhere - increasing the
amount of the control fluid leakage results in a more stable dead-head
condition of our pump. By the way, a "filed" variant of the added
leakage experiment can be easily performed by means of a gauge test
hose, connected to the X3 port (which already has the conveniently
placed test fitting there), with the other end disconnected (I would
point it to an adequate vessel, though...). The behavior of the control
with this "improvised" added leakage is seen in the next scene - again -
no self oscillations here.
In case you wonder how much leakage the disconnected test
hose added to the already existing one - observe the next part, where
you can see a one liter volume, filled by the test hose with the pump
dead-headed at 230 bar. The added leakage is around 4-5 liters per
minute.
Several orifice sizes were tested in the T1 port, and in
the end the original 0.8mm orifice was replaced with a 0.9mm orifice,
which didn't remove the oscillations completely, but reduced them to an
acceptably low level, with a relatively small increase of the leakage
flow. The final result can be seen in the last part of the clip.
Conclusions:
1) Control fluid leakage, along with the obvious benefits
of creating an additional cooling flow, can solve control stability
issues and reduce self-oscillations.
2) A 0.9 mm orifice is different from a 0.8 mm orifice, although the difference is barely seen with the naked eye.
3) In filed conditions, gauge connecting hoses can
serve as a rough replacement of an orifice or a restrictor valve.
