What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Oil Accumulator Burst

wawrzynskivp

Well Known Member
Moroso 1.5 qt Heavy Duty oil accumulator burst inflight. 100 Hrs in service.

Accumulator installed per installation instructions on engine mount Not on engine.

Upper cap sheared its threads and separated from body inflight. Immediate loss of all accumulator oil. Engine oil pressure was preserved by failsafe spring holding accumulator valve shut. Messy but not flight critical.

Returned to manufacturer for engineering investigation. Engineers ignored the helical remnants of sheared threads and pictures of safety wire keeping the cap from turning. Their conclusion was that the cap unscrewed and there is nothing deficient in design or manufacture.

I have put the accumulator back in service using steel plates on top and bottom of accumulator connected by four (4) 1/4-28 grade 8 threaded rods. I estimate this gives me a safe service pressure of over 500 psi.

My accumulator has temperature witness strips and so far has not exceeded 150 degrees surface temperature. Highest witnessed pressure while hot was 120 psi.

I am a retired aircraft accident investigator and without lab testing am fairly certain the threads suffered fatigue cycle failure and were improperly cut for full thread major diameter. This is just my opinion.

I am confident that my accumulator was not safe to operate the way it was received. Highly recommend not using this accumulator without reinforcing its end cap retention.
 
Installation details

No pics to share at this time.

End cap is ported with 1/2" NPT. I used 1/2"-1/4" NPT SS reducer, then 1/4" SS ball valve actuated with push pull cable. Downsteam of valve is #8 braided SS hose to a T between engine block and oil cooler. Fail safe was spring on ball valve to hold it shut if uncommanded.

Primary use for accumulator was start-up oil priming.

Certainly oil pressure captured cold would see an increase in acculumator pressure because of oil and gas expansion when hot. But running some rough numbers showed the increase in pressure to be less than 100 psi for a 150F increase in temperature. In practice the temperature deltas were much closer to 80-100F even in a closed cowling post flight.
 
Expansion

No, I didn't feel the oil expansion would be meaningful. Certainly measurable but probably not meaningful. I ran the numbers based on the trapped gas.

My common practice was to discharge prior to start, recharge when oil pressure was stable, then equalize after takeoff. Normal valve position was closed. Open if zero G transients were expected. Otherwise only open in an oil pressure failure.

I kept the elevated pressure on shutdown in order to retain the desired volume of oil. Post flight pressure would be 100-120 psi, then after cooling I would see 50-60 psi depending on ambient temps. Equalizing on shutdown made preflight oil volumes confusing.

My unit was also clamped (mounted) around the threaded portion of the caps per manufacturer instructions. I can't imagine the extra compression from the clamps provided any meaningful strain on the cap's threads, but whatever stress they provided was in the right place.
 
No, I didn't feel the oil expansion would be meaningful. Certainly measurable but probably not meaningful. I ran the numbers based on the trapped gas.

My common practice was to discharge prior to start, recharge when oil pressure was stable, then equalize after takeoff. Normal valve position was closed. Open if zero G transients were expected. Otherwise only open in an oil pressure failure.

I kept the elevated pressure on shutdown in order to retain the desired volume of oil. Post flight pressure would be 100-120 psi, then after cooling I would see 50-60 psi depending on ambient temps. Equalizing on shutdown made preflight oil volumes confusing.

My unit was also clamped (mounted) around the threaded portion of the caps per manufacturer instructions. I can't imagine the extra compression from the clamps provided any meaningful strain on the cap's threads, but whatever stress they provided was in the right place.

These units really need an inline check-valve/relief valve plumbed in parallel to the ball valve shut-off valve (or solenoid valve) to prevent over-pressure of the accumulator when the valve is closed.

The problem with accumulators is that the actual gas volume can become unknown over time unless the gas pressure is periodically (before each flight) checked with the accumulator known to be empty of oil. If the gas leaks out there may be very little expansion volume remaining for the oil when it's filled.

Skylor
 
Pressure relief

Thank you Skylor,

Considering your input I would have to agree that a parallel pressure relief either back to system or catch container for inspection would make a lot of sense.

Preflight verification of gas volume in addition to pressure would also make a lot of sense. In the past I have used the witness of oil gallery pressure rise to verify the gas chamber had sufficient volume, but that alone doesn't tell the whole story.

Some method of measuring the oil volume, or piston position would be in order. Any ideas? Would a capacitive sender work in this application?

Putting the E back in experimental.
 
HD Accumulator

Hi Dan,

Yes, I needed the CG to move FWD and thought the bulk of the HD unit would solve previous burst issues. But alas not so much. Of note on the oil expansion burst scenario: I had seen normal oil pressure rise on start up that day and the accumulator expelled all of its oil in a determined spray leading me to believe there was a decent volume of air below the piston that day.

But that failure mode is of interest.
 
No, I didn't feel the oil expansion would be meaningful. Certainly measurable but probably not meaningful. I ran the numbers based on the trapped gas.

My common practice was to discharge prior to start, recharge when oil pressure was stable, then equalize after takeoff. Normal valve position was closed. Open if zero G transients were expected. Otherwise only open in an oil pressure failure.

I kept the elevated pressure on shutdown in order to retain the desired volume of oil. Post flight pressure would be 100-120 psi, then after cooling I would see 50-60 psi depending on ambient temps. Equalizing on shutdown made preflight oil volumes confusing.

My unit was also clamped (mounted) around the threaded portion of the caps per manufacturer instructions. I can't imagine the extra compression from the clamps provided any meaningful strain on the cap's threads, but whatever stress they provided was in the right place.

Don't be so quick to discount trapped oil expansion pressure rise. I bet your cylinder had the piston bottomed out with no room to expand, causing the pressure within the vessel to spike exponentially, shearing the threads.

I actually have some experience with this: years ago I was involved with software development for the F-35. We had the actual aircraft Integrated Core Processor (ICP) housing installed in a 19" rack in a lab. PAO (thin synthetic oil) was used to cool the electronics and circulated within the white machined housing shown below. That unit is an LRU, so easily replaceable in the field, with quick-disconnect fittings for the PAO (the blue ones on the RH end). One day we received one shipped to us which had the walls cracked and was PAO-soaked. We determined that the unit was previously full of PAO, and then the QD's were disconnected and the unit air-shipped to us. By heat or altitude or both, that caused the pressure to rise to the point where the housing failed. And that housing is quite stout, given the critical nature of this device.

I think you need to figure out, by math or otherwise, how much air pressure is required to be maintained in the upper end of the cylinder to ensure the piston NEVER bottoms out.

f-35-icp-hardware-dsc_0629-003.png
 
There are surely fixes for the burst problem. That said, the absolute sure cure is to eliminate the accumulator. Parts not used never fail.

It's a risk-reward judgment. What are the rewards?
 
Risk Reward

I think the question may have been rhetorical, but here are some thoughts:

1: Put the E back into experimental. This is why we build these right?

2: So many claims out there that some large percentage of engine wear occurs during start up due to lack of oil gallery pressure. This would address that yes?

3: My build had an aft CG owing to certain chosen components. The heavy accumulator helped fix that.

4: Zero G transient oil pressure assurance,.

5: 1.5 Qt of extra oil for 'just in case scenarios'

But mostly it is for the experiment. The presence of problems can inspire us to retire or problem solve. There certainly are some added risks to this piece of added gear. Sooner or later somebody is going to hit on the elegant solution that addresses the risk and then we all will benefit from the rewards if desired...or not...again I think this is why the FAA allows and encourages this type of certificate.
 
Last edited:
How much air is required?

A useful exercise. I don't mind playing out the numbers.

Manufacturer say 10 psi on the gas side when accumulator is empty of usable oil for normal gasoline engine oil operating pressures. So what does that mean?

The bore of the accumulator is 4" I don't know the exact stroke but it has got to be very close to 10". There is a cavity in the piston face from memory it looks like 3.5" in dia and 1/2" deep.

So there seems to be about 130.4 in3 of volume on the gas side when the oil side is bottomed out. When the piston bottoms out on the gas side it will have 4.8 in3. So if we charge with 10 psi and assume for the purpose of drawing the worst case scenario that the oil will get hot but the gas will not, then also ignoring Boyle's law again to make the worst case we can calculate that the gas pressure will be 270 psi at bottom out.

Our oil will expand a little less than 5% over 100F. I don't get that kind of temperature rise but let's say I do. So the oil will force it's way irrespective of gas pressure another 6.52 in3 due to expansion. If we add that to the piston cavity and assume the pressure will be 270 psi when it bottoms we can calculate that if the gas side was charged to 10 psi empty, we would need to apply 115.2 psi of oil pressure to squeeze the piston to within 6.52 in3 of available volume.

So without Boyle's Law and without Gay Lussac's Law, I see a safety factor of 2. I am comfortable that the manufacture's recommendation is reasonable. BTW I use 15 psi when empty, just so we are all on the same sheet.

Folk's there are lots and lots of these things out there doing exactly this job, this isn't rocket science.

But, what happens' when the gas pressure leaks and how can we determine that before we fire up? This is the big concern.

My method had been to preflight the pressure and observe expected oil gallery pressure rise showing normal piston travel when the accumulator was opened. Not a quantitative approach, but at least one way to look at it.

The pressure relief valve is the obvious solution and well with all of our capabilities as builders. I think I will plumb mine to a catch bottle so I'll have another peace of mind thing to look at.

My Burst: I mentioned before that I observed normal oil gallery pressure rise at start up and when the accumulator burst it expended all of its oil in an aggressive spray. There certainly was meaningful pressure on the gas side for my burst. But how much?

In a past life I flew around with several fluid pressure accumulators and while I knew their pressure's I never once checked their gas volumes. Never had one burst and as the Maintenance Officer for twelve aircraft and aware of hundreds of others I never heard tell of other people's accumulators bursting. Not to say that it doesn't happen, just pointing out that accumulators are a sound part of aviation builds.

I really would like to figure out how to tell where the piston is without opening it up. I have called a few capacitive tank sensor manufacturers and so far none have offered a solution that can live in the described environment. The relief valve is the safety, and now I feel it is needed. But for operational considerations knowing the gas volume beyond it's apparent pressure is called for.
 
Last edited:
Piston position

Interesting thread. My experience with accumulators is on industrial hydraulics, where I likewise have never seen one fail. Would be nice to have a definitive explanation.

As for position sensing, a crude but common approach on industrial hydraulic cylinders is to embed a magnet in the piston, then attaching proximity sensors to the outside. Depending on the sophistication of the sensor and geometry, it can be a digital signal (piston is next to me, yes or no) or it can provide a little actual position information as the piston reaches end of stroke.

Caution: I've been away from the field for 7 years so my knowledge is dated and rusty. There may be better ways now, including variations on string potentiometers.
 
You did a lot of calculations on the gas and oil pressure inside the accumulator. But you didn't relate these to the shearing force at the threads. You know the thread size and can guess the material is 6061-T6. Should be an easy calculation to show if the Moroso engineers opinion is correct?
 
Moroso Engineering Opinion

I don't mean to seem snide, but the Moroso conclusion doesn't involve thread strength. They concluded that the cap simply unscrewed. They ignored the sheared threads, the rounded threads, the absence of helical witness marks under the clamps, and the intact safety wire that prevented rotation of the cap relative to the body.

Before I would try to speculative with math on pressure capacity of the cap as it relates to its threads I would need to examine the elastic and plastic characteristics of the accumulator walls. Then there are going to the thermodynamic expansion deltas between the cap and body based on thickness, heat sinking, and material chosen. Simpler to just apply a more secure retention system.

But, my point from Moroso was that the company was unwilling to address why there was material failure present. In fact they were surprisingly unwilling to formally recognize the obvious helixes of aluminum in the size and shape of the thread that was missing from the parts.

I made that point about Moroso's willingness to examine their product only to encourage that future builders not use this accumulator without additional cap retention.
 
I probably shouldn't have made a reference to the "Moroso engineers" at all. And you're right, there are a lot of variables that may have contributed to the failure. But, as a first order calculation, wouldn't a simple thread shear calculation show a "ballpark" safety factor based on the rated accumulator pressure? The website shows a 160 psi pressure gauge. For a pressure vessel, I would expect at least a 4:1 safety factor to the yield of any part of the accumulator.
 
These units really need an inline check-valve/relief valve plumbed in parallel to the ball valve shut-off valve (or solenoid valve) to prevent over-pressure of the accumulator when the valve is closed.

The problem with accumulators is that the actual gas volume can become unknown over time unless the gas pressure is periodically (before each flight) checked with the accumulator known to be empty of oil. If the gas leaks out there may be very little expansion volume remaining for the oil when it's filled.

Skylor

Thats why I'm planning on running an electronic oil pressure sensor on my accumulator instead of the dial gauge that's included with it. It'll be plumbed into the g3x and the gas precharge will be the reading after opening the valve before engine start. This way I don't have to pull the cowl to check the precharge level every flight.
 
If you really want a preoiler and can afford the space and weight for one, consider the oilamatic preoiler. It's a pump plumbed into the engine. The one on my certified airplane has been trouble-free.

I've noticed that using it on shut-down gives the prop a bounce-back on the final compression stroke. Without it, the prop comes to a stop like an axe into a stump. Of course you can't try this with an accumulator and still get pre-start oil.

Dave
 
Thread strength

If I was to rely on the Moroso design then looking further into the threads would make sense to me. But let's do it just for fun:

Normal expected loads on the end caps are an initial pre-charge, then initial oil pressure charge plus gas expansion due to heat, and the oil expansion due to heat so we should be at around 107 psi. I see numbers around 120 psi so the math is pretty close to practice.

Let's just go with 200 psi to make a point. 2512 lbs on the end caps. Shear strength on 6061 is in the ballpark of 30,000psi. There is about 3/4" of thread overlap so we have around 9 in2 and let's say they cut their threads to 75% so we are down to 3.4 in2, that's well above what's needed for 2,500lbs.

Now I don't know the materials, don't know how well they cut their threads and the stress concentration quality of their work. Lots of unknowns.

Did the accumulator generate more than 8,000psi required to shear its threads? Maybe, it's possible if the piston bottomed out. But we now have to look at the wall thickness of the accumulator and see if that kind of pressure is tolerable without at least some plastic deformation of which I saw none. (the walls are pretty thin) Would my tiny 1/4" ball valve have managed 8,000+ psi without any evidence? I doubt that, but I've never tried it.

So for the sake of playing with numbers there it is. I still am most convinced of fatigue cycle failure rather than ultimate failure, but the lingering threat of blocking the piston still looms. It was probably something in between. Steel has better fatigue properties so I think I have that possibility licked.

Big take away is that we should never let the pressures rise above our expected values by: 1. Using a pressure relief valve such as https://www.grainger.com/product/PARKER-316-Stainless-Steel-Safety-442F40 2. Somehow know that there is sufficient gas volume prior to each heat cycle. This is the part I am hoping someone can point to.
 
Ooh, yes.

I ran a hose to remote my guage where I could see it preflight. Too many opportunities for leaks. I have been looking for an electronic sender at the accumulator so I only have one leak point. agent4573: Could you share what gear you found?
 
Last edited:
I ordered an extra oil pressure sensor from garmin when I ordered my engine sensor kit. I believe the analogue gauge and the garmin sender are both 1/4 npt, so I should be able to just swap em out and mount the sender in the end of the accumulator.

https://www.steinair.com/product/g3x-oil-pressure-sensor/

Also, I just put together my accusump and it has a 175 psi blow off valve to prevent over pressure. It's the same spring loaded valve that's on air compressors. Does the moroso setup not have a safety release in the design?

https://www.amazon.com/Safety-relief-American-Compressed-valve/dp/B00NT78S9G
 
Last edited:
Here is a photo of the Accucump internal piston. Note the deep cavity that creates the minimum air volume. Can you provide the dimensions for the Moroso?

IMG_2958.jpeg
 
Moroso Parts

agent4573: Thanks for sharing. No relief on Moroso units. I am going to go with a safety valve that plumbs its discharge. A bit pricier, but cleaner and will provide quantitative evidence of overpressure discharge if routed to a catch bottle.

Still pursuing a method to determine piston position. The magnet method may be a good fallback.
 
Moroso Accusump Differences

Hey Bill,

The Moroso body is externally threaded rather than internally. I suspect that would enhance thread engagement with increasing pressure rather than reduce. But some destructive testing would be required for me to come to that conclusion.

I don't have any pics of the piston, just going off memory. The fixed space is definitely smaller than Accusump.

I am not silly enough to ignore the idea that my burst was related to a piston bottoming out. Again, I will not close that valve on a warming bottle again until I get my relief valve in place for that very reason.

I lean towards fatigue and improper thread cutting based on experience doing mishap investigations and the random nature of the burst. Many cycles of normal operation including evidence of gas volume loss without manifesting a failure. Then a cycle with all the numbers and behavior in the right place and the cap popped off. Fatigue vs Ultimate failure? Without the right tools I don't know so like I said it is probably somewhere in between. The system is not inelastic so mild bottom out events may have been occurring unnoticed. My pressure display for the closed accumulator was not visible from the cockpit so I would not ever have been aware of when the piston hit bottom.

All of the source of failure discussion is good speculation but in the end not really going anywhere. The shipped Moroso unit is not reasonably safe when operated according to manufacturer's instructions. The engineers at Moroso are not engaged in helping with that problem. But talking about it is good and hopefully somebody going this direction will see the discussion and avoid the messy burst.
 
Hey Bill,

The Moroso body is externally threaded rather than internally. I suspect that would enhance thread engagement with increasing pressure rather than reduce. But some destructive testing would be required for me to come to that conclusion.

I don't have any pics of the piston, just going off memory. The fixed space is definitely smaller than Accusump.

I am not silly enough to ignore the idea that my burst was related to a piston bottoming out. Again, I will not close that valve on a warming bottle again until I get my relief valve in place for that very reason.

I lean towards fatigue and improper thread cutting based on experience doing mishap investigations and the random nature of the burst. Many cycles of normal operation including evidence of gas volume loss without manifesting a failure. Then a cycle with all the numbers and behavior in the right place and the cap popped off. Fatigue vs Ultimate failure? Without the right tools I don't know so like I said it is probably somewhere in between. The system is not inelastic so mild bottom out events may have been occurring unnoticed. My pressure display for the closed accumulator was not visible from the cockpit so I would not ever have been aware of when the piston hit bottom.

All of the source of failure discussion is good speculation but in the end not really going anywhere. The shipped Moroso unit is not reasonably safe when operated according to manufacturer's instructions. The engineers at Moroso are not engaged in helping with that problem. But talking about it is good and hopefully somebody going this direction will see the discussion and avoid the messy burst.

Well, the root cause is important if it is fatigue in a reasonable operating range. Less of a serious factor since you have a shut off valve, I don't. The failed A brand showed a progressive failure (different cause for pressure) and the conclusion was it went solid several times and yielded the outer shell to ultimate failure.

I keep pushing for understanding when the piston goes solid because if that is the root cause for excessive pressure, then the relief valve is a safe solution. If not, then it is a more thorny issue. I did all the same calculations for stresses as you did and came to the conclusion the (failure) pressure had to be much higher. And, talking with the actual engineer at A-Brand, he had many examples of operation and bench testing to support that idea.

Until your event popped up I had Moroso envy :D as it claims a higher pressure capability than the Accusump. The internal thread design for the tube shows a better understanding of the design needs of this component.

Now to your reason for pre-lubing. Actually, this is a factor for large engines (ships) to float the bearings prior to starting. Engines below 10" bone seldom have that feature. The bearing hydrodynamic capability is good enough for a drain down that this failure mode, or contribution is far down the list of life influence on the bearings. I spent a lot of years learning from my former life in engine design/test/analysis to reach this conclusion. And this was on diesel with much higher forces on bearings at start an idle than Otto cycle. Did you know that and idling engine with 10:1 Cr only makes 50 psi compression pressure due to low MAP? Not saying it is worthless, but it will be an interesting experiment after you have a tear down. Worth the investment for that experiment!
 
Thread Strength

Hey Bill,

I think I understand your point, but for me the thread issue is moot. I am of the opinion whether justified or not, that threading the aluminum body and caps is a problematic approach altogether. I have clamped my caps between steel plates with 1/4-28 grade 8 rods. I no longer rely on the threads for anything other than assembly guides.

This was the purpose of my original post. i.e. Don't rely on the threads at all.
 
KISS

There are surely fixes for the burst problem. That said, the absolute sure cure is to eliminate the accumulator. Parts not used never fail.

It's a risk-reward judgment. What are the rewards?

AMEN Brother...
:)

V/R
Smokey
 
Back
Top