Lycoming IO360 “Roll-backs”

[PeytonM]

https://avbrief.com/faa-investigates-io-360-rollbacks/
I’m sure you’ve seen this already. One step away from an AD. And I’m in the final stages of factory rebuilt installation in my J! OMG!

[76Srat]

I'd love to hear from the experts on here about the specifics of the Avstar fuel servos and how broad those might be (if at all) in the IO-360-equipped Mooney fleet.  At first glance, I don't think there's much exposure.  But I've been wrong before when it comes to the FAA and its infinite wisdom and operator-friendly propensity to have very narrow application of its directives . . .

[kortopates]

I'd love to hear from the experts on here about the specifics of the Avstar fuel servos and how broad those might be (if at all) in the IO-360-equipped Mooney fleet.  At first glance, I don't think there's much exposure.  But I've been wrong before when it comes to the FAA and its infinite wisdom and operator-friendly propensity to have very narrow application of its directives . . .

Lycoming’s current‑production IO‑360 engines use AVStar fuel servos, not Precision Airmotive (RSA‑type) units.
Precision Airmotive RSA‑type servos are still widely used in the field and in overhauls, but new Lycoming engines ship with AVStar units.


Sent from my iPhone using Tapatalk

[PT20J]

Lycoming switched to AvStar some years back so if it was new, rebuilt, or overhauled  by Lycoming in the last 10 years or so, it will have an AvStar servo. 

[PeytonM]

Why are we just hearing about this now??

[76Srat]

It just seems dubious to me that the FAA would go the ACS route so quickly for what it essentially a fuel-starvation issue at idle in what I'm extrapolating to be mostly a training-based fleet.  Obviously, though not so obvious just yet to the larger public, there's been talk between the servo OEM, Lycoming and the FAA.  I could easily argue that any engine cutting out at idle in a training environment is because the student pilot forgot to enrich the mixture, fuel servo notwithstanding.  Again, easy to jump to all kinds of conclusions here and I'm certainly not trying to, but there's alot more going on than meets the eye here, so far.

 

[PeytonM]

I can confirm my factory rebuilt has the AvStar. Why are we just hearing about this now??

[N201MKTurbo]

I would imagine there have been SDRs about it. If you subscribed to SDRs you would already know about it. 
 

Go here and enter Avstar in the part make field.

https://sdrs.faa.gov/Query.aspx

[N201MKTurbo]

It seeme there is an issue with the idle mixture shaft. I read a few and they would not hold a consistent idle mixture.

[Shadrach]

On 12/17/2025 at 4:35 PM, kortopates said:

Lycoming’s current‑production IO‑360 engines use AVStar fuel servos, not Precision Airmotive (RSA‑type) units.
Precision Airmotive RSA‑type servos are still widely used in the field and in overhauls, but new Lycoming engines ship with AVStar units.


Sent from my iPhone using Tapatalk

Paul,

I’m confused by your post. I thought that the Avstar servo was the clone of the Bendix/Precision Airmotive RSA series. It’s my understanding that AVstar will overhaul Bendix/Precision Airmotive RSA servos.

[kortopates]

1 hour ago, Shadrach said:

Paul,

I’m confused by your post. I thought that the Avstar servo was the clone of the Bendix/Precision Airmotive RSA series. It’s my understanding that AVstar will overhaul Bendix/Precision Airmotive RSA servos.

The post was only about how pervasive the use of Avstar servo's are - they have entirely replaced in Bendix/Precision servo's in Lycomings production (as well as Flow dividers and injector nozzles) . But your comment is correct, AVSTAR, and many other fuel repair stations, will overhaul and both Bendix/Precision and Avstar servo's. But the Avstar is not a duplicate clone of the RSA servo but a FAA-PMA replacement for the Bendix/Precision servo's. It would probably be more correct to say Avstar matches Lycomings flow curves with their servo's. They utilize a different manufacturing process as well.

[skykrawler]

On 12/17/2025 at 6:50 PM, N201MKTurbo said:

It seeme there is an issue with the idle mixture shaft. I read a few and they would not hold a consistent idle mixture.

My understanding of the RSA system is the flow divider is the principle means of controlling mixture at idle.  The reason is the airflow at the throttle is too low to provide impact pressure to the (4) pitot tubes located at the opening of the servo.

The Avstar, design-wise seems to be a clone of the RSA in most respects.

From the Avstar document:

"In most installations the flow divider is fed with metered fuel from the fuel injection servo - fuel is then supplied to each nozzle. The flow divider ensures that equal amounts of fuel are fed to the nozzles at engine idle. The flow divider maintains fuel pressure, and this guarantees positive fuel shutoff when the servo is switched to idle cutoff. After entering the flow divider, fuel flows through the central bore of the flow divider spool valve to the diaphragm’s lower side. At idle, fuel pressure from the servo, acting on the lower side of the diaphragm, generates a force to overcome the flow divider spring. This lifts the spool permitting fuel to flow through the precisely machined outlet ports in the valve bushing. Fuel then flows directly to the nozzles. The fuel injection servo sends a fixed amount of fuel to the flow divider in response to airflow through the servo; hence, the valve opens only far enough to allow that amount of fuel to the nozzles. At idle, the spool valve opens only very slightly and the discharge pressure from the flow divider is negligible. Therefore, the fuel to individual cylinders at idle is divided by the flow divider. When fuel flow increases in response to engine demand, fuel pressure rises in the nozzle lines, and the flow divider valve is fully open. At this point, fuel distribution to the cylinders becomes a function of the pressure drop in each fuel line."

[N201MKTurbo]

36 minutes ago, skykrawler said:

My understanding of the RSA system is the flow divider is the principle means of controlling mixture at idle.  The reason is the airflow at the throttle is too low to provide impact pressure to the (4) pitot tubes located at the opening of the servo.

The Avstar, design-wise seems to be a clone of the RSA in most respects.

From the Avstar document:

"In most installations the flow divider is fed with metered fuel from the fuel injection servo - fuel is then supplied to each nozzle. The flow divider ensures that equal amounts of fuel are fed to the nozzles at engine idle. The flow divider maintains fuel pressure, and this guarantees positive fuel shutoff when the servo is switched to idle cutoff. After entering the flow divider, fuel flows through the central bore of the flow divider spool valve to the diaphragm’s lower side. At idle, fuel pressure from the servo, acting on the lower side of the diaphragm, generates a force to overcome the flow divider spring. This lifts the spool permitting fuel to flow through the precisely machined outlet ports in the valve bushing. Fuel then flows directly to the nozzles. The fuel injection servo sends a fixed amount of fuel to the flow divider in response to airflow through the servo; hence, the valve opens only far enough to allow that amount of fuel to the nozzles. At idle, the spool valve opens only very slightly and the discharge pressure from the flow divider is negligible. Therefore, the fuel to individual cylinders at idle is divided by the flow divider. When fuel flow increases in response to engine demand, fuel pressure rises in the nozzle lines, and the flow divider valve is fully open. At this point, fuel distribution to the cylinders becomes a function of the pressure drop in each fuel line."

The flow divider will stop fuel flow and trap the fuel in the injector lines when fuel pressure is removed from its input.

It has nothing to do with the idle mixture. The idle mixture is controlled by the position of the spool valve in the servo. It is set up to give a fixed fuel flow when there is not enough air pressure to open the servo control valves that are opened by the diaphragms. 
 

The problem seems to be associated with the machining of the spool valves.

[N201MKTurbo]

If the flow divider got stuck open, from a broken spring or some crud that jammed it open, the idle fuel distribution could be uneven.

If it was jammed closed, it would be in the cutoff position and the engine would not run.

 

[PeytonM]

Why would the design of the AvStar fuel servo make it more susceptible to “roll-backs” than the RSA?

[PT20J]

I looked at the 12 AvStar SDRs and didn't find a lot of consistency. Most read like "engine did something funny once and so we replaced the servo and it hasn't repeated." So, there must be more to the story.

At high power (maybe above 1500 rpm or so) fuel is metered by a combination of the position of the mixture valve (controlled by the mixture control) and the airflow through the venturi. At low power, there is not enough airflow to produce the necessary differential pressure on the regulator air diaphragm to accurately meter fuel. In this case, fuel is metered by the position of the mixture valve, a spring in the regulator and throttle position. The throttle is connected to an idle mixture valve that controls flow through the main jet by means of a linkage. The idle mixture screw adjusts the length of this linkage. 

The major difference I have noticed between the RSA and AvStar servos is that the AvStar has a much tighter seal on the mixture valve when in idle cutoff compared to the RSA. The effect of this is that fuel pressure will build in the plumbing between the firewall and the servo after shutdown due to residual heat in the engine compartment. The pressure will quickly rise off scale of the fuel pressure gauge on my airplane and others have reported the same effect. This shouldn't damage the servo because it is in ICO. However, someone here reported that they did have a failure that AvStar said was due to excessive pressure. 

 

[PeytonM]

Thank you everyone for your collective thoughts!

As I first posted, I am in the final stages of having a factory rebuilt IO360A3B6 installed in my 1989 J. At 4200 hrs. TT, including a Poplar Grove field OH at 2100, I decided to go Factory with this one, replacing the original 'D' variant with the 'non-D.' After about a 19 month wait(!) and $55K(!!), the engine arrived at Oasis, and they were ready to do the initial ground run when the ACS was published. So we have taken a pause to collect data, but it's mostly just opinions. I've looked at SDRs and don't find a consistent story, either. Of course we do not know what the denominator is, so it is impossible to calculate the frequency. Calls and emails to Lycoming have gone unanswered.

I am not sure what my options are; I went so far as to put a Precision/RSA servo on hold at Spruce.  That would be on my 'nickel.' Probably a little crazy. OK, a lot crazy!

My former engine went many hours and many landings. I've owned this airplane since 1993, so most of those hours & landings are mine. About a year ago, the engine quit during rollout after landing, and no issues were found, and it has not repeated since. I guess that is a "roll-back" and that engine had a Precision/RSA fuel servo.

I will admit, this has kept me awake, wondering what I should do, if anything. I figured if I was going to do anything, now would be the time. Maybe it's a nothingburger, or maybe it will be a AD.  I want to make sure that I have done as much as I can to assure that the engine is safe. I just don't want to worry about this when I take Mary flying.

[PT20J]

I installed a rebuilt IO-360-A3B6 that came with an AvStar fuel injection system (servo, flow divider, nozzles) in Oct 2018 and it has been flawless. I do find that it idles more consistently if I set the idle mixture a little rich. I wonder if some of these cases are due to having the idle mixture set too lean. If it’s too rich, you can always lean it with the mixture control, but there is nothing you can do if it is too lean.

[N201MKTurbo]

8 minutes ago, PeytonM said:

Thank you everyone for your collective thoughts!

As I first posted, I am in the final stages of having a factory rebuilt IO360A3B6 installed in my 1989 J. At 4200 hrs. TT, including a Poplar Grove field OH at 2100, I decided to go Factory with this one, replacing the original 'D' variant with the 'non-D.' After about a 19 month wait(!) and $55K(!!), the engine arrived at Oasis, and they were ready to do the initial ground run when the ACS was published. So we have taken a pause to collect data, but it's mostly just opinions. I've looked at SDRs and don't find a consistent story, either. Of course we do not know what the denominator is, so it is impossible to calculate the frequency. Calls and emails to Lycoming have gone unanswered.

I am not sure what my options are; I went so far as to put a Precision/RSA servo on hold at Spruce.  That would be on my 'nickel.' Probably a little crazy. OK, a lot crazy!

My former engine went many hours and many landings. I've owned this airplane since 1993, so most of those hours & landings are mine. About a year ago, the engine quit during rollout after landing, and no issues were found, and it has not repeated since. I guess that is a "roll-back" and that engine had a Precision/RSA fuel servo.

I will admit, this has kept me awake, wondering what I should do, if anything. I figured if I was going to do anything, now would be the time. Maybe it's a nothingburger, or maybe it will be a AD.  I want to make sure that I have done as much as I can to assure that the engine is safe. I just don't want to worry about this when I take Mary flying.

I would stop worrying about it. These problems are pretty rare and only affect idle. Your worse nightmare would be the engine quiting on rollout and you have to restart to taxi off.

[201er]

24 minutes ago, N201MKTurbo said:

I would stop worrying about it. These problems are pretty rare and only affect idle. Your worse nightmare would be the engine quiting on rollout and you have to restart to taxi off.

Pulling idle on final?

[kortopates]

Pulling idle on final?

The airspeed will keep the prop turning over and engine running till you’re at least exiting the runway or you give it some throttle.


Sent from my iPhone using Tapatalk