I'm an electronics engineer working with a machine that tests hydraulic actuators and my knowledge of hydraulics is a little limited at the moment.

If I have a pump supplying pressure at 5000psi and there is a significant amount of air trapped in the system, will the air act as an energy sponge an ultimately prevent the system from reaching the 5000psi? I have a gauge well downstream of the pump that will only hit about 4300psi and my hunch is the air is constantly absorbing the energy and converting it to heat and it shows up as a loss of pressure in the system.

12/18/2010 10:14:31 AM

I can usually tell that my brakes on the mountain bike need to be bled when the lever gets mushy. Sure sign of air in the system.

12/18/2010 10:34:52 AM

Right, because when you squeeze the energy gets transmitted to the air as heat and liberated instead of being applied to the brakes (where it also gets converted to heat).

I suppose I'm wondering - if the system is strong enough can it get into a state where the air is absorbing energy (and transmitting it to the surrounding oil, to the hose, etc) at the same rate the pump is putting it in such that the system will sit there at a reduced pressure for minutes on end unable to come up to the full supply side pressure.



[Edited on December 18, 2010 at 10:56 AM. Reason : .]

12/18/2010 10:55:24 AM

so uh have you thought about getting a technician to check the system for air

12/18/2010 12:05:59 PM

The question isn't whether the system has air in it. The question is what does air in the system do to the pressure you're attempting to apply. It's more a discussion of physics I suppose.

12/18/2010 2:56:36 PM

Quote :

"Right, because when you squeeze the energy gets transmitted to the air as heat and liberated instead of being applied to the brakes (where it also gets converted to heat). I suppose I'm wondering - if the system is strong enough can it get into a state where the air is absorbing energy (and transmitting it to the surrounding oil, to the hose, etc) at the same rate the pump is putting it in such that the system will sit there at a reduced pressure for minutes on end unable to come up to the full supply side pressure."

The reason for the lower pressures is the fact that gasses are compressible, unlike liquids. Thus when some volume in the system is taken up by a gas, there will be lower overall pressure. Hope that helps.

12/18/2010 3:34:03 PM

But they aren't infinitely compressible and the pump is constantly delivering energy to the system. The fact that the pressure reads lower means energy is being lost. Where?

12/18/2010 4:48:57 PM

from air IN the pump maybe. Also depending on how you control the pressure (bypass), it may be allowing too much to go by because of the air.

[Edited on December 18, 2010 at 5:09 PM. Reason : d]

12/18/2010 4:57:48 PM

Unlikely. We have a single pump supply flow through a very large ball valve first into a big manifold that splits it into 4 channels with ball valves controlling these individual channels. In this particular instance, one channels return will come up to full pressure and the other won't. If it were a pump issue, we'd see it in both channels.

[Edited on December 18, 2010 at 5:02 PM. Reason : .]

12/18/2010 5:01:56 PM

i would check for a leak or blow by in the valve. even with air, a static system should come up to pressure. however, it will behave erratically when put under load or actuation. i used to work with a big hydraulic dyno rig, if you can provide a little more detail i may be able to offer some insight.

12/18/2010 7:31:57 PM

Quote :

"I have a gauge well downstream of the pump"

pressure drops due to frictional losses in the piping between the pump discharge and the gauge

12/18/2010 7:53:27 PM

^Only dynamic pressure drops...think velocity of fluid vs friction. Static pressure on a system served by a positive displacement fluid pump? Think again. You know all this.

Provided we're dealing with a static system here...no fluid moving anywhere except maybe through a popoff or unloader valve at the pump, it doesn't matter how much air is in the system...the system WILL eventually reach equilibrium at max pressure, provided that the air in the system isn't being passed through the pump.

Now let's change gears here. If there's dynamic flow through the system continuously, then one of two factors can be limiting your displayed pressure: 1. Fluid friction as stated above...dynamic friction losses in the system. 2. Lack of either restriction to create a pressure head or adequate pump capacity to produce the desired pressure. OR...3. Maybe the pump is worn and there's too much internal bypass.

12/18/2010 9:22:40 PM

The test at issue is a proof pressure test, no flow. Let me see if I can describe again the path from pump to actuator a little better.

1) Pump supplies up to 8000psi through a main valve into a manifold that splits the supply into 4 seperate channels. Each of these channels has its own valve to control flow to it's channel.
2) From here the flow goes through hard lines to a manifold where it gets converted to flexible rubber/steel lines to another manifold.
3) At this manifold there is a node that takes a hard line down to the actuator, a hose to a gauge, and a manual air bleed valve.

4) The supply goes through the actuator which has it's solenoids on and its eh valves opened fully in one direction. The actuator actually has two systems configured in this way, supplied with two separate hydraulic channels.

5) We have the return lines in the same configuration, one to the part, one to the gauge, one with an air bleed and the main hose/steel line going back to another manifold that will have a valve connected to it. I believe it is this valve that we close to proof the return side of the system.

There is no leakage in the part that we can see. The supply side gauge will read full pressure but the return will not. If there is leakage after the part back at the return valve, would we not see that on both supply and return gauges? Is it possible that the leakage is in the part and if we let it sit for a very very long time we'd see it eventually come to full pressure?



And picture this, we have hangers in this sink area to support the large supply/return hoses and up until Friday, both supply and return hoses went over these hangers and their highest point was higher than the manifold where our air bleed is. We installed a new actuator and had to re-arrange these large hoses and took them off the hangers such that the highest point was now the air bleed. When we did our actual bleed step, more air came out of the system than I've ever seen. Yes, we had the system open switching actuators, but we do this every couple of days and bleed afterwards and I've never seen this amount of air come out - I'm speculating that a very large volume of air was trapped at the top of this hose.

This is what had me wondering if just trapped air would prevent the system from coming up to full pressure.

Quote :

"i would check for a leak or blow by in the valve."

And we have flowmeters. I guess you'd need to know the area of the lines for an idea of how much flow might be flowing for a given pressure drop? I'm just wondering if this is something that will be readily obvious from the flow reading.


[Edited on December 19, 2010 at 12:42 PM. Reason : .]

12/19/2010 12:30:41 PM

Zx covered most of it. It should theoretically get to your 5000psi as long as the air wasn't being recirc'd back through the pump. I want to say eventually that heat would dissipate into the hydraulic fluid and it would pressurize, but the one thing I'm thinking is that if it pressurized it to that high of a pressure the air trapped in the system could be roughly distributed through the fluid and not be like a normal air bubble at a high point. Also, be sure that your pressure gauge is in calibration and check for leaks. You said you just rearranged some of these lines and even tiny leaks on a small pressure supply line to a gauge can lose a lot of pressure.

12/19/2010 1:51:48 PM

The gauges were just calibrated a week ago. I swapped gauge hoses so that the problem return side (call it A) was now going to B (which had no problem). When doing this, the pressure in A would go to mostly full rail and the pressure in B was now not coming up. Originally, not knowing dick about hydraulics we were speculating that an air bubble was trapped right up next to the gauge. But we switched the hoses and the problem followed telling us it wasn't a gauge issue. So...this is where I was thinking perhaps a large volume of air was trapped and absorbing energy.

Is it possible the actuator itself could act like some sort of a pressure valve where the supply side would see the full rail with a ~700psi drop across the actuator to the return side? And assuming that the return side valve is completely closed after the actuator, and with no leaks in the actuator or anywhere else in the system, with enough time wouldn't the return eventually equalize to the supply side?

12/19/2010 4:15:39 PM

Got a chance to try a different actuator today, no problem. So...I suppose flow is leaking from one system in the actuator to the other somehow. The bizarre thing is why have tests that check inter-system leakage that don't fail.

12/21/2010 10:03:08 PM

At least you've got the testing gear back to operating normally. What kind of actuator is it? Is it a linear actuator with a piston head design; how old is it and was it recently reassembled? Those style sometimes use pressure-loaded profile seals on each side of the piston to fully engage against the cylinder walls in each direction as the pressure builds. If one of those seals has been chewed up during reassembly, incorrectly assembled, or from contamination/premature wear it can lead to normal pressurization on one side/stroke direction of the piston and subpar pressurization on the other even if the actuator still operates normally, especially at the travel stops or ends of the actuator strokes.

Piston seals shown in red above, but hydraulic actuator designs and styles vary widely. This model only has one seal but the one below showing the "U-Cup Piston Seals" would be a likely scenario in your issue if one of the backing rings was not installed correctly or the seal was damaged.

This stuff probably won't matter if you outsource the actuator mfg to another company for repairs or orders but if you do it in-house the techs would/should like to know about it.

12/24/2010 1:28:28 AM

Quote :

"What kind of actuator is it? Is it a linear actuator with a piston head design; how old is it and was it recently reassembled? T"

It is a linear but so far it's a little bit of a black box to me as I've been working on the assumption (up to now) that it was known good and we were developing tests for it. It was brand new and I can't really fathom how many hours it has on it, likely approaching 1000.

Quote :

"If one of those seals has been chewed up during reassembly, incorrectly assembled, or from contamination/premature wear it can lead to normal pressurization on one side/stroke direction of the piston and subpar pressurization on the other even if the actuator still operates normally, especially at the travel stops or ends of the actuator strokes"

Interesting, this is how we are doing the proof pressure testing, with at least one hydraulic channel (and I think all three) servovalves opened fully in the extend and retract direction for the supply and then again in the extend direction for return (that is, we open the servovalves fully one direction, there is a main control valve between them and the ram that will be opened fully, and then the ram itself is extended/retracted fully and everything is left this way for the test). Any ideas on how this is manifesting as a pressure difference? I'd (ignorantly, for sure) assume that even if flow was leaking past these seals, the pressure would eventually equalize as we have the return to tank valve completely shut and as best I can tell, nothing is leaking anywhere else in the system. But left for a long time this doesn't seem to happen.

[Edited on December 29, 2010 at 3:36 AM. Reason : .]

12/29/2010 3:34:19 AM

I'm more of a visual person so forgive me but I still can't visualize the way you have the system piped up. I deal more with pneumatics but only slightly in hydraulics. What's the reason you have 3 servovalves in parallel? Assured Redundancy? Also, that last statement confuses me; does that mean it eventually reaches equalized pressure after some time? Lastly, do you fully extend/retract the actuator before or after you shut off the main return line?

12/29/2010 8:59:32 PM

This is a swashplate actuator. Flight critical. So yeah, lots of redundancy.

During normal flight, the three servovalves (ehvs) work together to control the actuator. All three ehvs operate a main control valve which ports fluid to an aft chamber and a forward chamber. The two chambers are of course summed and move the main ram. Two of the ehvs are fed from one hydraulic manifold and I believe the main control valve has working ports for this hydraulic channel that feds the aft chamber. The other ehv is supplied from a different manifold and the main control valve has ports for this supply feeding the forward chamber.

We proof the supply side of the actuator with it fully extended and the ehvs cranked fully one way (which also has the main control valve fully one way). We then retract the actuator and have the ehvs cranked the other way. The return is open to tank and we don't see any flow.

We then extend the actuator again with ehvs positioned fully one way, block the valve to tank and allow the return side working ports to come up to proof pressure. On a working actuator we typically will see the supply side reading ~5700psi on both hydraulic channels (expected as they are supplied from a common pump) while the return will read 5000psi on both channels. We ramp fairly slowly and it takes about 3-4 minutes to get to these pressures. We hold here for 5 minutes.

On our bad actuator, will see one hydraulic channel come to 5000psi on return as expected but the other channel tends to top out around 4700psi and then starts falling back down and will eventually settled near the 4300psi range.

12/29/2010 9:32:56 PM