Periodic engine surge/vibration

[peter-h]

I've noticed this a few times when the engine is running at a very low power setting, way into LOP, during descent.

There is a periodic behaviour, perhaps best described as an air bubble in the fuel, every 3-4 seconds. Very brief; lasting maybe 1/3 of a second.

The RPM doesn't change so it isn't the prop governor. All readings including EGT (EDM700) and fuel flow (Shadin) are stable.

I don't think I have a fuel leak (an air leak into the fuel system), since the Shadin system is consistently accurate to within 1% and I check this at every fillup.

I reckon it is something in the fuel servo - RSA5AD1 on an IO-540-C4D5D engine.

But what suprises me is the long period of this - several seconds. What kind of physical phenomenon would account for this long period?

 

[RotorAndWing]

One problem of the fuel servos you mentioned as they wear they develop air leaks around the throttle shafts thus begin sucking air.

When was this servo last overhauled? We use to overhaul ours every 1200 hours.

 

[peter-h]

It was opened up by a well known firm in Tulsa, OK, a year ago when the engine went to the USA for the SB569A crank swap. They found a leaky fuel valve, apparently, which account for a shudder when shutting down.

I'd imagine that had there been wear elsewhere, they would have noticed, but maybe not?

The TT on the servo is 800hrs, and the above work was done ~ 150hrs ago.

 

[Tom-D]

I've noticed this a few times when the engine is running at a very low power setting, way into LOP, during descent.

There is a periodic behaviour, perhaps best described as an air bubble in the fuel, every 3-4 seconds. Very brief; lasting maybe 1/3 of a second.

The RPM doesn't change so it isn't the prop governor. All readings including EGT (EDM700) and fuel flow (Shadin) are stable.

I don't think I have a fuel leak (an air leak into the fuel system), since the Shadin system is consistently accurate to within 1% and I check this at every fillup.

I reckon it is something in the fuel servo - RSA5AD1 on an IO-540-C4D5D engine.

But what suprises me is the long period of this - several seconds. What kind of physical phenomenon would account for this long period?

Richen the mixture and see if anything changes.

 

[EdFred]

Pretty much what I was thinking too, Tom. My first thought was carb ice until I saw it was an IO-

 

[sba55]

Pretty much what I was thinking too, Tom. My first thought was carb ice until I saw it was an IO-

Well, I don't think Tom is saying that being too lean is causing the problem?

Normally, your engine will just die when it is too lean, and it will do so smoothly. That setup is difficult to achieve, however, so maybe there's something else going on. The engine roughness makes sense to me if the fuel flows aren't perfectly balanced, but the surging is odd.

-Felix

 

[peter-h]

Richen the mixture and see if anything changes.

The problem disappears.

If it didn't I wouldn't be flying with it

It is obviously something that happens only when operating way into the LOP region.

 

[gismo]

I've noticed this a few times when the engine is running at a very low power setting, way into LOP, during descent.

There is a periodic behaviour, perhaps best described as an air bubble in the fuel, every 3-4 seconds. Very brief; lasting maybe 1/3 of a second.

The RPM doesn't change so it isn't the prop governor. All readings including EGT (EDM700) and fuel flow (Shadin) are stable.

I don't think I have a fuel leak (an air leak into the fuel system), since the Shadin system is consistently accurate to within 1% and I check this at every fillup.

I reckon it is something in the fuel servo - RSA5AD1 on an IO-540-C4D5D engine.

But what suprises me is the long period of this - several seconds. What kind of physical phenomenon would account for this long period?

There could be a fluctuation in the fuel flow that wouldn't show on the Shadin as they heavily filter (smooth) the indication. One factor here is that your fuel servo uses manifold airflow in addition to a combination of the mixture and throttle positions to set the fuel flow and at low power settings this airflow can vary significantly with mixture changes. That would allow a "bootstrap" cycle where the mixture get's too lean reducing the airflow which then causes a further reduction in the fuel flow until it get's too lean to run, at which point the airflow might increase enough to start the whole thing again. This was a common problem in car engines when they first started trying to run them lean to reduce pollution.

 

[RMCN172RG]

LOP running requires matched injectors as I understand living in the normally aspirated world with no actual experience to draw on.

Could it be a fuel injector partially plugged causing one cylinder to cut out momentarily?

 

[SCCutler]

LOP running requires matched injectors as I understand living in the normally aspirated world with no actual experience to draw on.

Could it be a fuel injector partially plugged causing one cylinder to cut out momentarily?

I think this is a good possibility.

---

One other thing which is a long shot, but I bring it up anyway:

I had an occasional vibration / shudder which only appeared (to me) on descent at low power, relatively high airspeed, and I had written it off as something like the misfire described above, because nothing at all else showed up.

Then, when my plane was being test-flown for the IFR cert after the 430W upgrade, the test pilot felt a vibration.

Into the shop she went, and nothing expected revealed itself, but I did get to spend a lot of money on "maybes"; rebalanced the ruddervators (well, they needed it anyway, as it turned out), replaced a couple of aileron links, new spark plugs... then finally, a busted motor mount was found, and it seems as though there was, at particular combinations of airspeed, power and RPM, a little harmonic or other vibration made possible by the "slop" the busted motor mount allowed.

May not be relevant for you (my plane has "bed" mounts, bet yours has dynafocals), but it's an "out of the box" possibility...

 

[Tom-D]

The problem disappears.

If it didn't I wouldn't be flying with it

It is obviously something that happens only when operating way into the LOP region.

Your leanest cylinder is trying to quit. but it is getting enough fuel to fire intermittently.

richen it up until it runs smooth, you will still be LOP. If you spread is not too wide.

don't richen it too much as to leave your richest cylinder on the rich side of peak.

 

[peter-h]

Your leanest cylinder is trying to quit. but it is getting enough fuel to fire intermittently.

Interesting possibility. I saw no indication of this on the EDM700 though. I can download the data and see if there is something there but for sure nothing was showing on the EDM which I watch like a hawk all the time. But yes enriching the mixture fixes the issue immediately.

It may have always been there.

The EGTs also show a huge spread - this seems typical of deep-LOP operation at very low power settings. There are two peaks in the EGT as one leans: one is the normal peak we all know about, and further along one gets a second peak but at a uselessly low (for level flight) power output.

I have GAMIs (2003) and the SB569A engine shop (2008) did a lean check to show that the peak EGT still happened together at a specific common fuel flow rate. I have seen their data and it was OK.

Maybe I need a new set of GAMIs?

What puzzles me is what physics could bring about the long period of this oscillatory behaviour. It "must" be something in the fuel servo. There isn't anything else anywhere else around the engine which could have such a long time constant.

This doesn't actually matter but as I say I am curious about the physics. This

That would allow a "bootstrap" cycle where the mixture get's too lean reducing the airflow which then causes a further reduction in the fuel flow until it get's too lean to run, at which point the airflow might increase enough to start the whole thing again. This was a common problem in car engines when they first started trying to run them lean to reduce pollution.

seems a good condidate.

However, I thought the airflow into an engine is a function ONLY of RPM and throttle opening (WOT in this case), and nothing else. The combustion process is isolated from the air pump side of things, by the inlet valves being shut anything combustion is going on. Also, the fuel servo has no control over the airflow, other than by the throttle lever (the butterfly valve).

I have Lord engine mounts - fitted new ones (lots more $$$ than the Barry alternative but the Barry ones looked really "cheap") in 2008.

 

[gismo]

Maybe I need a new set of GAMIs?

If you're able to run significantly LOP at normal cruise power settings I wouldn't even touch the injectors but if your GAMI spread has widened, it might be time to clean the injectors you have and/or check for induction leaks. BTW an induction leak will have the most pronounced effect when the throttle is nearly closed, e.g. when descending at low power.

What puzzles me is what physics could bring about the long period of this oscillatory behaviour. It "must" be something in the fuel servo. There isn't anything else anywhere else around the engine which could have such a long time constant.

There actually is a cycle with up to a few seconds of period involving mixture variations due to changes in the airflow that occur when a cylinder "misfires" but most folks would classify that as a "shudder" or brief roughness.

However, I thought the airflow into an engine is a function ONLY of RPM and throttle opening (WOT in this case), and nothing else. The combustion process is isolated from the air pump side of things, by the inlet valves being shut anything combustion is going on. Also, the fuel servo has no control over the airflow, other than by the throttle lever (the butterfly valve).

Try leaning the mixture while watching the manifold pressure with the throttle set for 1000-1200 RPM on the ground. You will likely see a significant increase in manifold pressure when you get lean enough. That increase in MP is the direct result of reduced airflow through the engine (and intake manifold). The reduction occurs because the pressure in the cylinders at the point where the intake valve opens is higher, and the higher pressure is the result of the less complete and/or slower combustion during the power stroke with an excessively lean mixture.

 

[peter-h]

Try leaning the mixture while watching the manifold pressure with the throttle set for 1000-1200 RPM on the ground. You will likely see a significant increase in manifold pressure when you get lean enough. That increase in MP is the direct result of reduced airflow through the engine (and intake manifold). The reduction occurs because the pressure in the cylinders at the point where the intake valve opens is higher, and the higher pressure is the result of the less complete and/or slower combustion during the power stroke with an excessively lean mixture.

However, adjusting the mixture is going to change the RPM (at 1000-1200 one is below the governor control range) and that alone could account for the MP variation.

IF you see an MP change as a result of a mixture change at a constant RPM, that would be something else. I will try that next time.

How does one clean injectors?

 

[gismo]

However, adjusting the mixture is going to change the RPM (at 1000-1200 one is below the governor control range) and that alone could account for the MP variation.

IF you see an MP change as a result of a mixture change at a constant RPM, that would be something else. I will try that next time.

Normally when you reduce RPM by closing the throttle, the MP goes down not up. I'm pretty certain that if you leaned until the MP increased and simultaneously cracked the throttle open enough to maintain a constant 1200 RPM, you'd still see the increase in MP, and it would probably increase even more.

In any case, the airflow through the engine is affected by the combustion process as I described. The residual pressure in the cylinders at the beginning of the intake stroke is the missing factor in your thinking. And of course, be certain you get the injectors back into the same cylinders they came from. They are marked but the markings don't necessarily differentiate between left and right.

How does one clean injectors?

Remove them, soak them in Hoppes #9 or SeaFoam for half an hour. If you want them really clean, do this in an ultrasonic cleaner, but I've never had a problem with just a plain soaking. Whatever you do, DO NOT try to clean them with anything but liquid. No poking wires into the holes, no wiping with rags, no brushing etc.

 

[Dr. O]

When it starts that, turn the boost pump on and see if there is a change...

denny-o