What's new
Van's Air Force

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

IO-540 at 2500rpm with EFII - drops cylinder

Jesse

Well Known Member
I am trying to work through a problem we are having with an RV-10 with an IO-540 and EFII. We have gotten to the point where we can get Cylinder 1 to drop (complete EGT drop and engine runs rough) by running the engine at 2500 rpm. The manifold pressure doesn't matter. We have replaced the injector. We have swapped the injector power wire with cylinder 3 (to verify it isn't a bad crimp). We have changed ECU's vie the ECU 1/2 switch. We have run on both and either mag individually. Nothing gets rid of this problem. At 2530 rpm there is no problem. At 2450 rpm there is no problem. At 2500 on the money, it drops cylinder 1. No other cylinder, and no exception. Every time, cold engine or hot engine, lean or rich, in flight or on the ground.

My main questions is, what could possibly be magic about 2500 rpm that could make this happen.

We though about a stuck valve at some point, but that would not just be at one specific rpm, at least not that I could imagine.

Thoughts?
 
I am trying to work through a problem we are having with an RV-10 with an IO-540 and EFII. We have gotten to the point where we can get Cylinder 1 to drop (complete EGT drop and engine runs rough) by running the engine at 2500 rpm. The manifold pressure doesn't matter. We have replaced the injector. We have swapped the injector power wire with cylinder 3 (to verify it isn't a bad crimp). We have changed ECU's vie the ECU 1/2 switch. We have run on both and either mag individually. Nothing gets rid of this problem. At 2530 rpm there is no problem. At 2450 rpm there is no problem. At 2500 on the money, it drops cylinder 1. No other cylinder, and no exception. Every time, cold engine or hot engine, lean or rich, in flight or on the ground.

My main questions is, what could possibly be magic about 2500 rpm that could make this happen.

We though about a stuck valve at some point, but that would not just be at one specific rpm, at least not that I could imagine.

Thoughts?
I would blame it on software and a bad map, or a weird injector.
 
Last edited:
Jesse,

Since you've done the fuel injector trouble-shooting, I'd suspect that you're losing spark on #1. If you have a timing light or engine analyser with an inductive pickup, you can clip it on a #1 spark plug lead to check.
 
Jesse,

Since you've done the fuel injector trouble-shooting, I'd suspect that you're losing spark on #1. If you have a timing light or engine analyser with an inductive pickup, you can clip it on a #1 spark plug lead to check.

The problem with that is that it only fails at 2500 rpm.
 
Whats Robert say> maybe the computer doesnt know the engine is suppose to run at 2500. Over rev protection like an MSD Chip?

Tom
 
The problem with that is that it only fails at 2500 rpm.

Exactly. If the timing light or engine analyser with an inductive pickup on the #1 ignition lead, shows spark at 2400 RPM and no spark at 2500 RPM, you've probably isolated the problem to the EFII unit.
 
Last edited:
Exactly. If the timing light or engine analyser with an inductive pickup on the #1 ignition lead, shows spark at 2400 RPM and no spark at 2500 RPM you've isolated the problem to the EFII unit.

It runs great at 2700 rpm. I don’t think it’s a over-rev issue.
 
Which could actually be a hardware problem in the EFII unit - bad memory corrupting the map.

I was thinking maybe a bad current driver, but don't know if they are shared, and/or why it would shut off at 20.83 hz. Maybe if they injection pulse is timed, then each injector would have its own driver, but then we get back to why the lookup table would have a zero or it is a computing error. Thus the reason for suspecting the ubiquitous software problem.

If the PWM signal at the injector were measured via break-out harness via o-scope reading, then it could be definitively verified as no current to the injector.

Yes, corruption of some table could have happened yielding a zero on-time for that box, but wouldn't it interpolate and trail off and on with an RPM sweep?
 
Last edited:
A simple "spark" test would be to use any tachometer which is capable of sensing a P-lead impulse and wire it to the spark plug lead in question. You may have to wrap the sense wire around the silicone spark plug wire a few turns to get enough energy to trigger the tachometer. Of course the RPM reading will be totally incorrect but if the RPM reading goes to zero when you hit 2500 real engine RPM then you know you have a spark problem.
 
Do you have the full EFII system with Electronic Fuel Injection and Electronic Spark?

If you have both, it would be helpful to try to determine if the problem is fuel or ignition. I was wondering if there could be a loose connection that just happens to resonate at 2500 RPM

But in reality, I'm a Mags and Carb luddite, so please feel free to laugh at my expense!:p

I can't believe I'm even commenting on a Jesse Saint post!!:eek:
 
A simple "spark" test would be to use any tachometer which is capable of sensing a P-lead impulse and wire it to the spark plug lead in question. You may have to wrap the sense wire around the silicone spark plug wire a few turns to get enough energy to trigger the tachometer. Of course the RPM reading will be totally incorrect but if the RPM reading goes to zero when you hit 2500 real engine RPM then you know you have a spark problem.

We have a Dynon Skyview system in the plane, and it gets RPM from the EFII ECU's. Again, the RPM doesn't change, and 5 of the 6 cylinders are still running normally. Just Cylinder #1, and always Cylinder #1, is dropping out. We know it's not a sensor on the EGT because the engine runs rough (not excessively so, but about like a 6-cylinder would run if it lost one cylinder). We have replaced or isolated every issue we can think of and it doesn't change the situation. We have changed mixture from ROP to LOP with no change. The only single point of failure would be the fuel delivery, but we have replaced the injector and swapped injector power wires with another cylinder with no change. Dual ignition and dual ECU's seem difficult to imagine being the problem, especially since Cylinders 1 and 2 share a coil on each ignition.

Can anybody think of a cylinder issue that would or could cause a cylinder to stop running just at a specific RPM?
 
Certainly sunds like a s/w problem or possibly an electronic component failure one the main board . Just need to confirm injector or ignition. Run at 2475 and observe fuel flow. Then run at 2500 and observe fuel flow. If the injector is failing to open at 2500, you should be able to observe a reduced fuel flow.

Both ignition and injector pulse width have tables that are rpm dependent.

Ignition seems unlikely, as both coils would have to fail the same cylinder at the same time. Also, likely wasted spark coils that would fail to fire in cylinder pairs. Likely that the unit has a discrete mosfet driver for each injector and that would b the likely failing component, though in that case, the problem would have followed the injector swap IF the wiring followed the injector.

If you have been in the s/w configuration area an error is most likely.
Larry
 
Last edited:
Certainly sunds like a s/w problem or possibly an electronic component failure one the main board . Just need to confirm injector or ignition. Run at 2475 and observe fuel flow. Then run at 2500 and observe fuel flow. If the injector is failing to open at 2500, you should be able to observe a reduced fuel flow.

Both ignition and injector pulse width have tables that are rpm dependent.

Ignition seems unlikely, as both coils would have to fail the same cylinder at the same time. Also, likely wasted spark coils that would fail to fire in cylinder pairs. Likely that the unit has a discrete mosfet driver for each injector and that would b the likely failing component, though in that case, the problem would have followed the injector swap IF the wiring followed the injector.

If you have been in the s/w configuration area an error is most likely.
Larry

I have talked to Robert Paisley and he says it can't be a software problem because it is the same software as every other unit he has shipped.

Fuel flow won't tell, because fuel flow is a data point that is sent from the ECU to the Dynon, and the K Factor adjusted to get it accurate, so if the ECU is telling the injector to open, it will say it is flowing fuel whether it is opening or not. Also, I hooked the wires from Cylinder 3 to Cylinder 1 and vice versa and that problem stayed with Cylinder 1, so the wiring is working fine. We replaced the injector on Cylinder 1 and it didn't fix the problem. I may switch an injector anyway to make sure.

Thanks for the thoughts so far. We are going to look at the intake pipe and exhaust pipe as well as pull the valve cover and see what we find.
 
This May Sound Weird...

Have you looked at the spark plugs? Try replacing one or both on Cylinder 1 to see if that has any effect.
Frequencies, resonance, and harmonics can have some crazy effects in things electronic as you well know.
 
Make a break out harness to plug inline with the offending injector. Take the wires and hook to an oscilloscope and look at the injector power pulse. It should be a square wave with varying on time. Run through the 2500 RPM point then validate when the issue occurs and what the pulse looks like.

If that shows the injector pulse has disappeared or changed in shape, take a photo or screen shot of it. Then get back to Robert with the issue. He should know the architecture of the controls, sensors and internal hardware of the ECM's.

Then - connect to the ECM maps, look at all the possibilities that could occur exactly at 2500 rpm. If a large step change in a map occurs then change it to something in-between the adjacent cells.

BTW, the software guys always say it isn't software. Sometimes they are right, but no one will know the inner workings like the code writer.
 
My 2 cents:
Far from being an expert, logic says that it would be improbable that software is an issue. In one ECU, possible yes, in both individually the SAME software glitch ???
And since Jesse swapped the injector connectors, the issue would then transfer to #3 cylinder...
I have no idea of a solution.

Definitely a strange problem...
Keep on searching
 
BTW, the software guys always say it isn't software. Sometimes they are right, but no one will know the inner workings like the code writer.

+100

I spent decades in the s/w business. They are pretty conditioned to say that it can't be s/w, as no one else reported the problem. In the s/w world, it is 100%"innocent until proven guilty!" and that burden typically falls on the end user. Certainly not saying they aren't right often, as they are routinely accused when h/w or config issues turn out to be the problem.

Just ask the guys with the plane power problems. Have all confidence that when hartzell answers the phone they say "odd, we have never seen this before" even though we see hundreds of posts with the same issue.

If it is not a s/w or config issue, I would be looking at the injector. I suspect it is possible that it only fails at a specific Pulse width, though seem implausible, as the pulse width is a function of both RPM and MAP and wouldn't expect it to be exactly the same PW at 2500 in all conditions. It should also be adjusting PW based upon IAT, which shouldn't have been the same through all of your testing. It just feels like a s/w or config issue with the fuel table. Could also be a failure in a memory chip or EEPROM, with a corrupted sector.

I would definitely look at the fuel table entries in the 2500 RPM column. That data is likely saved to an EEPROM and corruption could change the value to something that the CPU doesn't understand and cuts the injector off. It's possible Robert is building tables for each injector, vs a batch fire protocol. He seems to brag about semi-sequential, so who knows what he is doing in s/w.

Larry
 
Last edited:
You mentioned that you moved the injector power wire between #1 and #3 with no change.

I'm not familiar with the current hardware Robert is selling, but I have the SDS module in my airplane - and it does not switch the power for the injectors, it switches the ground. The power is supplied to each injector constantly through a fuse, and the ground return path comes back to the ECU where the drivers will momentarily ground them to cause the current to flow for the pulse width. If Roberts hardware is the same, then swapping the power for the injectors does nothing to diagnose that particular fault path - you'll need to swap the ground.

If you have an oscilloscope probe on the injector wiring, you should still see the square wave either way and it will tell you the story.
 
Here we go...

"... It's possible Robert is building tables for each injector, vs a batch fire protocol. He seems to brag about semi-sequential, so who knows what he is doing in s/w..."

Do we really need to start bashing, yet again?

I have the System 32 and it has been flawless...
 
"... It's possible Robert is building tables for each injector, vs a batch fire protocol. He seems to brag about semi-sequential, so who knows what he is doing in s/w..."

Do we really need to start bashing, yet again?

I have the System 32 and it has been flawless...

No bashing. Just reiterating what Robert wrote on this forum (he was bragging that his sequential injection is better than all the other batch fire systems on the market) and why it might be a potential hypothesis for troubleshooting. Glad you are happy with your system, but it would seem that doesn't mean 100% perfection for all. Sequential injection implies individual injector control and helps to explain why only one injector would misbehave where batch fire systems would fail in pairs or more. Sorry that you didn't see the value in my observation, but possibly the OP does.
 
Last edited:
We have isolated the wiring and the control of the injector as not the problem because we connected the harness and connector going to cylinder 3 onto cylinder 1 and vice versa. So, cylinder 1 is getting fuel when cylinder 3 is supposed to, and cylinder 3 is getting fuel when cylinder 1 is supposed to. That would eliminate the software in that part of the equation from being the problem.

We now have swapped injectors between 1 and 3 and the problem remains.

Cylinders 1 and 2 get spark from the same coil at the same time, and there are 2 independent ignition coils. so that is not the problem.

We have switched between ECU 1 and ECU 2 and the problem remains.

We have run ROP and LOP, 10-26" MAP and it still happens.

Again, the craziest part of this, it ONLY happens at 2500 rpm. Not 2530. Not 2470. Just 2500. I feel confident that I can go fly for 20 hours and never see it if I stay away from 2500 except just transversing fairly quickly. I also know from experience that I can go run it on the ramp or fly it and get it to happen every time.
 
Here's a wild idea - is it possible the prop control cable, when it's exactly at the position for 2500 rpm, is somehow interfering with the spark to that cylinder?
 
No bashing. Just reiterating what Robert wrote on this forum (he was bragging that his sequential injection is better than all the other batch fire systems on the market) and why it might be a potential hypothesis for troubleshooting. Glad you are happy with your system, but it would seem that doesn't mean 100% perfection for all. Sequential injection implies individual injector control and helps to explain why only one injector would misbehave where batch fire systems would fail in pairs or more. Sorry that you didn't see the value in my observation, but possibly the OP does.

I never said that I didn't see value in your observation, quite the contrary...

The part I take issue with is, "...He seems to brag about semi-sequential, so who knows what he is doing in s/w..." which adds absolutely no value to the discussion, and is nothing more than a jab at a competing system...
 
Here's a wild idea - is it possible the prop control cable, when it's exactly at the position for 2500 rpm, is somehow interfering with the spark to that cylinder?

It is on the other side of the engine and never gets close to the plug wires.
 
It is on the other side of the engine and never gets close to the plug wires.

Hi Jesse I hope you’re doing well! When I installed a efii system32 on my RV-10 I noticed the spark plugs supplied were such that the tips were screw on instead of solid. Is it possible that the tip on the cold cylinder spark plug has unscrewed? What does the EGT spread look like on the cylinders during normal operation? Another item that may be worth checking is the fuel supply line for that particular cylinder….perhaps some junk is caught at the junction of the T. Just thinking out loud here.

God bless
Bill
 
Last edited:
Hi Jesse I hope you’re doing well! When I installed a efii system32 on my RV-10 I noticed the spark plugs supplied were such that the tips were screw on instead of solid. Is it possible that the tip on the cold cylinder spark plug has unscrewed? What does the EGT spread look like on the cylinders during normal operation? Another item that may be worth checking is the fuel supply line for that particular cylinder….perhaps some junk is caught at the junction of the T. Just thinking out loud here.

God bless
Bill

Bill,

I was just going to add this same thought. I had the EFII ignition only on my -7 and was trying to track down an intermittent erratic RPM issue. It turned out that the NGK screw terminal plugs were an issue on #2. It happened at the same time on climb out at higher RPMs as the frequency of RPM, prop combo was causing just enough vibration to excite the loose plug terminal cap. It took a while to track down but replacing with solid terminal NGKs fixed this particular issue. I didn't realize NGK offered the screw terminal plugs until further investigation.
 
We have isolated the wiring and the control of the injector as not the problem because we connected the harness and connector going to cylinder 3 onto cylinder 1 and vice versa. So, cylinder 1 is getting fuel when cylinder 3 is supposed to, and cylinder 3 is getting fuel when cylinder 1 is supposed to. That would eliminate the software in that part of the equation from being the problem.

We now have swapped injectors between 1 and 3 and the problem remains.

Cylinders 1 and 2 get spark from the same coil at the same time, and there are 2 independent ignition coils. so that is not the problem.

We have switched between ECU 1 and ECU 2 and the problem remains.

We have run ROP and LOP, 10-26" MAP and it still happens.

Again, the craziest part of this, it ONLY happens at 2500 rpm. Not 2530. Not 2470. Just 2500. I feel confident that I can go fly for 20 hours and never see it if I stay away from 2500 except just transversing fairly quickly. I also know from experience that I can go run it on the ramp or fly it and get it to happen every time.

So, are you saying that after swapping the injector wiring between 1 and 3 that #1 continued to stop firing at 2500? Of did the dead cylinder move to #3 at 2500?

I remain convinced that the only way for you to have one cylinder go completely dead at only 2500 RPM must be an issue with the s/w or the h/w that executes its will. Also possible that bad sensor data is causing the s/w to get confused. Also, just because you haven't been in the configuration area doesn't necessarilly mean that data hasn't changed. Corruption in an EEPROM, or whatever data storage method is used, can cause a data change or produced unreadable data. Ever had a hard disk sector go bad and make a file unreadable on your computer?

In all my years, I have never seen a mechanical engine issue cause such a unique, specific and isolated fault, like this one, that is 100% repeatable. IMHO, it is just way too specific to be a mechanical issue.
 
Last edited:
So, are you saying that after swapping the injector wiring between 1 and 3 that #1 continued to stop firing at 2500? Of did the dead cylinder move to #3 at 2500?

I remain convinced that the only way for you to have one cylinder go completely dead at only 2500 RPM must be an issue with the s/w or the h/w that executes its will. Also possible that bad sensor data is causing the s/w to get confused. Also, just because you haven't been in the configuration area doesn't necessarilly mean that data hasn't changed. Corruption in an EEPROM can cause a data change or produced unreadable data.

In all my years, I have never seen a mechanical engine issue cause such a unique, specific and isolated fault that is 100% repeatable.

The problem stayed with Cylinder 1 even after moving the harness. The thing that tells me that it is not a software issue is that the software doesn't know that it is giving fuel to Cylinder 1 when it thinks it's giving to Cylinder 3, and that it is giving fuel to Cylinder 3 when it thinks it's giving to Cylinder 1. On the spark, the coil for each ignition sparks Cylinder 1 and Cylinder 2 at the same time, all the time. So if the ignition timing is 25 BTDC, it is firing 1 and 2 at 25 BTDC on the Power/Compression Stroke and the Exhaust Stroke. The only thing that could possibly be there is that at 2500 it may be missing every other spark, and it just happens that that's the compression stroke every time on Cylinder 1 and the exhaust stroke every time on Cylinder 2. It's hard, however, to imagine that it would be so consistently wrong on both ECU's and both ignitions.

I suppose it is possible that the frequency of vibration right at 2500 rpm could cause the coil that sparks 1 and 2 to somehow lose connection to 1, but it is pretty hard to imagine, and that it is so consistent, exact and repeatable.

As for the data, the Dynon Skyview HDX logs the data, and it agrees completely with what we remember from having the problem both in flight and in ground runs.
 
The problem stayed with Cylinder 1 even after moving the harness. The thing that tells me that it is not a software issue is that the software doesn't know that it is giving fuel to Cylinder 1 when it thinks it's giving to Cylinder 3, and that it is giving fuel to Cylinder 3 when it thinks it's giving to Cylinder 1. On the spark, the coil for each ignition sparks Cylinder 1 and Cylinder 2 at the same time, all the time. So if the ignition timing is 25 BTDC, it is firing 1 and 2 at 25 BTDC on the Power/Compression Stroke and the Exhaust Stroke. The only thing that could possibly be there is that at 2500 it may be missing every other spark, and it just happens that that's the compression stroke every time on Cylinder 1 and the exhaust stroke every time on Cylinder 2. It's hard, however, to imagine that it would be so consistently wrong on both ECU's and both ignitions.

I suppose it is possible that the frequency of vibration right at 2500 rpm could cause the coil that sparks 1 and 2 to somehow lose connection to 1, but it is pretty hard to imagine, and that it is so consistent, exact and repeatable.

As for the data, the Dynon Skyview HDX logs the data, and it agrees completely with what we remember from having the problem both in flight and in ground runs.

I can't see spark as an issue, as the cylinder wouldn't go dead unless BOTH coils stopped sparking. The likelihood of both wasted spark coils failing to spark only one side of the coil at the same time and at a specific RPM is so far off the charts to be almost impossible. If you confirm that only one sparkplug is firing on #1 due to some issue, then a odd harmonic becomes a possibility. Just run it up to 2500 with a timing light on each plug wire to eliminate the spark as an issue

I Would definately swap the injectors between 1 & 3 to eliminate it as a cause. A unique harmonic causing the injector to stop working is at least in the realm of possibility. It has a VERY thin copper winding in it and a compromised segment of that wire could be subject to vibration.

Larry
 
Last edited:
I Would definately swap the injectors between 1 & 3 to eliminate it as a cause.
Larry

See post #29 made yesterday: We now have swapped injectors between 1 and 3 and the problem remains.

This is an interesting conundrum and it may turn out that it gets solved by switching to mags, but I imagine that would be quite a bit of work. I wonder if Barrett installed the EEFI system and dyno tested the engine with it, or if Barrett used mags and then pulled them for the owner to install his own ignition/FI system.
-Marc
 
Last edited:
See post #29 made yesterday: We now have swapped injectors between 1 and 3 and the problem remains.

This is an interesting conundrum and it may turn out that it gets solved by switching to mags, but I imagine that would be quite a bit of work. I wonder if Barrett installed the EEFI system and dyno tested the engine with it, or if Barrett used mags and then pulled them for the owner to install his own ignition/FI system.
-Marc

Check for continuity of ground on the #1 & #2 spark plugs with the cyl head. An interesting thing about wasted spark coils, is that the ground return is through the opposite side of the coil and not the engine, though the engine transfers the energy from one plug to the other. If the continuity from plug to cyl head is not good on both plugs, then one of them may not fire. When I test my EFI on the bench, I have to ground the two plugs bodies together, else neither will fire.

I would definately put a timing light on one of the #1 plug wires to see if the sparking stops at 2500. Pretty easy test. If the light is not flahing, there is no spark at the plug. No need to hold the light near the spinning prop. It is a good way to identify intermittent spark failure events.

Larry
 
Last edited:
You switched to the backup ECU-----no joy.

You swapped injector wiring-----no joy.

You swapped the injector itself----no joy.

I think it may be time to stop looking at the electronics.
 
You switched to the backup ECU-----no joy.

You swapped injector wiring-----no joy.

You swapped the injector itself----no joy.

I think it may be time to stop looking at the electronics.

I wouldn't have expected it to be mechanical, for sure - but it's looking more likely.
 
I wouldn't have expected it to be mechanical, for sure - but it's looking more likely.

Only two mechanical elements to support combustion with an installation like this. Proper compression and valves opening or closing properly. If you find either of those to be cause here, I will be more than shocked. Don't see how either of those can cause total failure at 2500 RPM, yet perfect operation at <2470 and >2530 RPM. Peripheral mechanical components like oil pump and fuel pump or even camshaft timing can't really affect just one cylinder with a total power loss, while the other 5 work perfectily. Ring problems, a hole in the piston, or head leakage would create problems in a much broader RPM range.

I even considered valve float from a broken valve spring, but in that case, the cyl shouldn't come back to life at 2530.

The incredible RPM precision and repeatability really only points to the EFII controller, its s/w or one it it's sensors. I simply can't see how compression or valve train issues can only manfiest in problems at a very specific RPM.

Larry
 
Last edited:
You switched to the backup ECU-----no joy.

You swapped injector wiring-----no joy.

You swapped the injector itself----no joy.

I think it may be time to stop looking at the electronics.

Three elements to combustion: compression, fuel/air and spark. I see only one listed above. Once both fuel and spark are confirmed it may be time to move on to the mechanical components, but not before. Mechanical issues rarely create such a narrow fault range.

Larry
 
Last edited:
Three elements to combustion: compression, fuel/air and spark. I see only one listed above. Once both fuel and spark are confirmed it may be time to move on to the mechanical components, but not before. Mechanical issues rarely create such a narrow fault range.

Larry

Agree-----I did leave out spark.

So, move the plugs and see if the problem follows the plugs.

Next move the coil and see if the problem follows the coil.

Sound good?
 
Just to give a final update. We have solved the problem. It turns out that either the intake or exhaust hydraulic lifter bodies was causing the problem. I don't know fully what was happening, but apparently at 2,500 rpm there was some resonance in the spring and the hydraulic unit was either not deflating or not inflating or over inflating, causing a valve to not fully close. The best flight we got with the existing (newly overhauled) hydraulic bodies was after I removed them and deflated them manually. It was not perfect, but that cylinder did not fully drop out. The next flight it did, so we kept checking other things. Finally I bought 2 hydraulic bodies and the problem has been gone for 3 or 4 flights now. Kudos to LJ at Zephyr Aircraft Engines for suggesting that.
 
Lifters?

How did you replace the lifter bodies? Dont you have to split the case to do this on that engine?
Ed
 
The tappet stays in the engine, but for the hydraulic unit, you just remove the valve cover, rocker arms, pushrods, pushrod shrouds and just pull it out. Very simple, actually.
 
I don't think this is related. If the cam was spalling so the valved didn't open all the way, the problem would likely be at all RPM's, not just one. Either way, the problem appears to be gone. Only [flight] time will tell.
 
Back
Top