Injector fouling

[Dorfie]

Fuel system is Bendix RSA 5 servo with Bendix fuel divider. Complete system is 6 years old from complete overhaul by Airflow Performance and about 550 flight hours. I have done the fuel flow distribution check, and all 6 fuel samples were exactly the same. I mainly fly LOP during cruise and suspect that any further decrease in fuel flow to a cylinder will manifest itself with rough running.

It has now happened twice that I likely had an injector partially blocked. First one No 5 on startup (shut down and changed injector to same size) and second No 2 on take-off recently (Few seconds duration then recovered itself). No indication of a magneto issue. Engine data appears consistent with impaired fuel flow to those cylinders. The two events occurred about 30 hours apart with no indication that there were such episodes during those 30 hours. Injectors are from Airflow Performance with sizes 0.0245-0.0255". After the first episode I ultrasonically cleaned all the injectors, even though I could not see any abnormalities.

Mechanical fuel pump functions normally with no discharge from the drain tube.

The main fuel filter in the tunnel (74 micron which is 0.0029 inch)) was clean 6 months ago. I have never had dirt or water in strainer during preflight fuel sampling, nor visible dirt in the main fuel filter during CI. I will check the filter again.

Assuming that the dirt did not get past the main fuel filter in the tunnel, where can dirt come from downstream from the main filter? (Mechanical fuel pump, Fuel servo (RSA 5), Fuel divider??)

Where downstream are filters/strainers located that I might inspect for dirt accumulation? (I think there is one on fuel inlet to servo, and maybe one at inlet to fuel divider??).

Anything else I should do or check?

Appreciate your help.

Thank you
Johan

 

[Carl Froehlich]

I would first look at the spider. A small chunk of stuff trapped in it resulted in a similar issue on an RV-8A. The data that had the problem change between cylinders was the give away.

You might want to discuss with Don Rivera on how to open and inspect the spider.

Carl

 

[DanH]

An RSA-5 does have a tubular inlet filter.

 

[lr172]

Fuel system is Bendix RSA 5 servo with Bendix fuel divider. Complete system is 6 years old from complete overhaul by Airflow Performance and about 550 flight hours. I have done the fuel flow distribution check, and all 6 fuel samples were exactly the same. I mainly fly LOP during cruise and suspect that any further decrease in fuel flow to a cylinder will manifest itself with rough running.

It has now happened twice that I likely had an injector partially blocked. First one No 5 on startup (shut down and changed injector to same size) and second No 2 on take-off recently (Few seconds duration then recovered itself). No indication of a magneto issue. Engine data appears consistent with impaired fuel flow to those cylinders. The two events occurred about 30 hours apart with no indication that there were such episodes during those 30 hours. Injectors are from Airflow Performance with sizes 0.0245-0.0255". After the first episode I ultrasonically cleaned all the injectors, even though I could not see any abnormalities.

Mechanical fuel pump functions normally with no discharge from the drain tube.

The main fuel filter in the tunnel (74 micron which is 0.0029 inch)) was clean 6 months ago. I have never had dirt or water in strainer during preflight fuel sampling, nor visible dirt in the main fuel filter during CI. I will check the filter again.

Assuming that the dirt did not get past the main fuel filter in the tunnel, where can dirt come from downstream from the main filter? (Mechanical fuel pump, Fuel servo (RSA 5), Fuel divider??)

Where downstream are filters/strainers located that I might inspect for dirt accumulation? (I think there is one on fuel inlet to servo, and maybe one at inlet to fuel divider??).

Anything else I should do or check?

Appreciate your help.

Thank you
Johan

last year I had numerous injector clogging issues after 800 trouble free hours, including two dead stick landings. Tore everything apart and found a WHOLE bunch of some kind of white deposits all over the inside of the fuel pump. My best guess if a non standard additive was added to the avgas somewhere, but no conclusive data. Never anything but 100LL.

 

[rmarshall234]

Aren’t white deposits indicative of water?

 

[hgerhardt]

.....After the first episode I ultrasonically cleaned all the injectors, even though I could not see any abnormalities.........

Did you ever find any foreign objects in the injectors?
If not, sounds more like exhaust valve sticking. Have you done SB 388C yet? You have the right amount of hours for this condition to start.

 

[airguy]

Remember with the injection system that the servo continues to send the same amount of fuel regardless. If you have one cylinder injector plug up, now the other 3 (on 4-cylinder engine) will absorb all the remaining fuel and will have 4/3 of their fuel load compared to before the restriction (or some fraction of that for a partial restriction). If you have full engine monitoring you'll be able to see one cylinder go very lean and the other three going very rich, that's a sure sign of an injector plugging, and it's obvious which one it is.

If you are losing power in one cylinder, but aren't seeing the mixture variation on the other cylinders of an injected engine, then you need to look at ignition or an exhaust valve hanging open.

 

[Bill.Peyton]

It sounds like a sticking valve. You have the right number of hours for this to start showing up. You should be performing sb388c every 400 hours regardless. There are a couple of good threads on VAF and you tube videos on how to perform the task. I do this religiously every 400 hours

 

[lr172]

while symptoms vary, generally a partially blocked injector will show reduced EGT and CHT; significant, but not catastrophic. Often a stuck exh valve will have the EGT and CHT drop off completely, as not enough compression to ignite charge and you generally cannot miss the unpleasant vibrations associated with running on three cylinders.. A fully blocked injector will do the same, but those are less common. Careful analysis of the data logs should help point you in the right direction. Partially blocked injectors occasionally clear themselves.

 

[bjdecker]

...you generally cannot miss the unpleasant vibrations associated with running on three cylinders...

...Along with extreme sphincter tightening...

 

[airguy]

while symptoms vary, generally a partially blocked injector will show reduced EGT and CHT; significant, but not catastrophic. Often a stuck exh valve will have the EGT and CHT drop off completely, as not enough compression to ignite charge and you generally cannot miss the unpleasant vibrations associated with running on three cylinders.. A fully blocked injector will do the same, but those are less common. Careful analysis of the data logs should help point you in the right direction. Partially blocked injectors occasionally clear themselves.

Hehe, true that.

I lost an injector once on climbout from New Orleans, IMC in moderate rain at about 5000'. The ensuing IFR approach to 400' ceiling with 3 cylinders was interesting.

 

[bjdecker]

last year I had numerous injector clogging issues after 800 trouble free hours, including two dead stick landings. Tore everything apart and found a WHOLE bunch of some kind of white deposits all over the inside of the fuel pump. My best guess if a non standard additive was added to the avgas somewhere, but no conclusive data. Never anything but 100LL.

Friend of mine on field here noticed some interesting material in the bottom of both of his RV-9's tanks. We believe it came from the filters in our fuel island; some kind of clay-like filter media...

His -9 didn't have any issue (O-320, MA4 carb) ingesting the stuff.

 

[Dorfie]

Thank you for all the replies. Much appreciated.

Regarding sticky exhaust valve:
This engine was split about 60 hours ago to replace a corroding camshaft and cam followers. At the time IRAN was done and all cylinder and valve, valve guides and seats were inspected and cleaned. (At least that is what I have been told as standard procedure. Work was done by Signature engines in Cincinnati.) I did wobble test on No. 1 cylinder after engine overhaul and before any of these issues surfaced, and that was within normal range. This made me think that sticky exhaust valve to be less likely. The first episode happened directly after startup, smooth running engine with high EGT and cooler CHT No. 5 cyl and both mags working. Shut down, replaced the injector with same size, and no further issues with that cylinder. The second was at take-off after spending about 17 min on ground taxing etc. Mixture at lean all the time. My initial though was fouled plugs.

Regarding fuel flow and partially blocked injector:
Looking at the engine data there is a brief increase in EGT prior to decreasing which I interpreted as leaner mixture immediately before more blockage of the injector. There is also drop in fuel flow as the EGT starts going down. The amount by which the fuel flow decreased is about 1/6 of the initial fuel flow which was 25gph. I thought that is somewhat in line with one injector not delivering its share of the 25gph. I also pulled the mixture back a little when my thinking was soiled plugs, and I believe that was the second drop in fuel flow. At that time the cylinder came back online. But this is reason why I am asking this forum.

I am trying to post the graphics of these two episodes from Savvy, but not sure how to. Guidance will be appreciated. Hopefully doing it that way will allow zooming into the data.

Thank you all.
Johan

 

[airguy]

The amount by which the fuel flow decreased is about 1/6 of the initial fuel flow which was 25gph. I thought that is somewhat in line with one injector not delivering its share of the 25gph.

The RSA5 has no way of knowing that an injector is plugged - it will not decrease the fuel flow for a plugged injector. It only knows how much air is going into the engine, and sends the appropriate amount of fuel to match that amount of air up to the injector divider. If one out of six of the injectors are plugged, the other 5 get that amount of fuel fed to them and each of those cylinders runs 6/5 of the original mixture.

 

[Dorfie]

The RSA5 has no way of knowing that an injector is plugged - it will not decrease the fuel flow for a plugged injector. It only knows how much air is going into the engine, and sends the appropriate amount of fuel to match that amount of air up to the injector divider. If one out of six of the injectors are plugged, the other 5 get that amount of fuel fed to them and each of those cylinders runs 6/5 of the original mixture.

Interesting. I believe the servo supply pressure (not constant volume) to the flow decider. If one injector stops flowing, the pressure to the flow divider remains the same (which is determined by the air volume through the servo) and flow to the functioning injectors remain unchanged. In this case fuel flow will decrease accordingly.

If the servo delivered constant volume flow, then the pressure will increase with plugged injector and the functioning injectors will deliver more volume.

This understanding and reasoning led to my interpretation that fouled injector will lead to decrease in fuel flow. I am very open to correction though.

Hope this link works. I had RPM with EGT graph, and fuel flow with CHT graph. I hope the zoom function remains available.

https://apps.savvyaviation.com/flights/shared/flight/7692491/b7d6a7c7-b120-4dc4-9914-c12e23d667a9

Johan

 

[rv8ch]

Interesting. I believe the servo supply pressure (not constant volume) to the flow decider. If one injector stops flowing, the pressure to the flow divider remains the same (which is determined by the air volume through the servo) and flow to the functioning injectors remain unchanged. In this case fuel flow will decrease accordingly.

If the servo delivered constant volume flow, then the pressure will increase with plugged injector and the functioning injectors will deliver more volume.

This understanding and reasoning led to my interpretation that fouled injector will lead to decrease in fuel flow. I am very open to correction though.

Hope this link works. I had RPM with EGT graph, and fuel flow with CHT graph. I hope the zoom function remains available.

https://apps.savvyaviation.com/flights/shared/flight/7692491/b7d6a7c7-b120-4dc4-9914-c12e23d667a9

Johan

This works. Adding FF shows a lot of changes during this event.



That must have been an exciting 30 seconds.

 

[DanH]

Greg's description of the RSA's response to nozzle blockage is correct.

I've attached a snip from the Bendix training manual. The added notes are mine. Old school flow gauges were actually pressure gauges with the dial face labeled as GPH rather than PSI. Although all our dividers still have the pressure port, that approach to fuel pressure measurement is rarely seen in current RV's.
.

 

[lr172]

Greg's description of the RSA's response to nozzle blockage is correct.

I've attached a snip from the Bendix training manual. The added notes are mine. Old school flow gauges were actually pressure gauges with the dial face labeled as GPH rather than PSI. Although all our dividers still have the pressure port, that approach to fuel pressure measurement is rarely seen in current RV's.
.

I am not sure that is universally correct. If I reduce the nozzle size on my 320 from .028 to .024, my total flow will drop and all cylinders will be leaner even though the pressure went up; Still plenty rich, but no longer filthy rich. This is why I am not allowed to go much lower than .024 or .023 on injector size, as I wouldn't get the flow that I needed at WOT. Therefore, I believe that a blocked injector will give some indication of reduced fuel flow. Yes, with 1 of 4 injectors plugged, the other three will get richer due to the higher pressure, but will not get the full volume that would have flowed through the now missing injector. DOn't believe it is linear like that, as this is all a balance between pressure and orifice volume. THink about it. on a 4 cyl at WOT, I am sending 4 GPH to each cylinder. If I now send 5.3 GPH (25% increase) (dividing up the blocked cyl volume across the other 3), we're getting to REALLY rich territory where we are at least going to see a power drop off and maybe misfires. Your average RSA with .028 injectors is already down around 10:1 at full rich and can only go so far into richer territory before it stops combusting.

 

[DanH]

I am not sure that is universally correct. If I reduce the nozzle size on my 320 from .028 to .024, my total flow will drop and all cylinders will be leaner even though the pressure went up; Still plenty rich, but no longer filthy rich. This is why I am not allowed to go much lower than .024 or .023 on injector size, as I wouldn't get the flow that I needed at WOT. Therefore, I believe that a blocked injector will give some indication of reduced fuel flow.

The note says you won't get an indication of increased fuel flow.

Don't want to drift the thread, but I am curious. Jetting down with nozzle size would be...unusual. Did you actually make the change, and if so, how much did fuel flow drop?

Returning to topic, whatever the figure might be, the OP's would be less. Moving from 0.028 nozzles to 0.024 nozzles is a total flow area reduction of 26.5%. Totally plugging one 0.028 nozzle on the OP's 540 reduces total flow area by 16.6%.

 

[Dorfie]

Greg, Dan,
Thank you for pointing me in the right direction. I now understand much better.

With the mentioning of possibility of sticky exhaust valve into play now as well as my initial perceived fouled injector, does the engine data help in finding the real cause? I must admit that I did not consider sticky exhaust valve due to the low hours since IRAN and camshaft replacement.

I attach a link to the No.5 cylinder anomaly at startup that occurred about 30 hrs before the dead cylinder on takeoff recently. That was when I changed the injector with no subsequent problems. I have a picture of the G3X display screen at that time, but so far unable to get from my I-phone to attach here. I basically saw high EGT and low CHT compared to the rest.

https://apps.savvyaviation.com/flights/shared/flight/7616003/88004497-187d-4e29-856a-27c5dfd3938e

Thanks.
Johan

 

[lr172]

The note says you won't get an indication of increased fuel flow.

Don't want to drift the thread, but I am curious. Jetting down with nozzle size would be...unusual. Did you actually make the change, and if so, how much did fuel flow drop?

Returning to topic, whatever the figure might be, the OP's would be less. Moving from 0.028 nozzles to 0.024 nozzles is a total flow area reduction of 26.5%. Totally plugging one 0.028 nozzle on the OP's 540 reduces total flow area by 16.6%.

Unfortunately I didn't test before / after. Got this from Don at AFP. Don recommended 024's for my 320 and 025's for my 540 over the more standard 028's.. He indicated that the 028's flow a lot (well more than necessary) and while the smaller nozzles will flow less, it was still above the required amount, but the increased pressure involved helps to solve several issues. He did say that you can only go small, otherwise the flow rate will drop to an unsafe level. He used to design these things, so take him at his word.

 

[lr172]

Greg, Dan,
Thank you for pointing me in the right direction. I now understand much better.

With the mentioning of possibility of sticky exhaust valve into play now as well as my initial perceived fouled injector, does the engine data help in finding the real cause? I must admit that I did not consider sticky exhaust valve due to the low hours since IRAN and camshaft replacement.

I attach a link to the No.5 cylinder anomaly at startup that occurred about 30 hrs before the dead cylinder on takeoff recently. That was when I changed the injector with no subsequent problems. I have a picture of the G3X display screen at that time, but so far unable to get from my I-phone to attach here. I basically saw high EGT and low CHT compared to the rest.

https://apps.savvyaviation.com/flights/shared/flight/7616003/88004497-187d-4e29-856a-27c5dfd3938e

Thanks.
Johan

You stated above that just before these symptoms appeared you did a wobble test and were in spec. It takes 100's of hours to build up enough coke to make a valve stick, so seems quite unlikely that is your problem.

As I look at the chart, I see the symptoms of a piece of debris floating around the restrictor. Blocks, opens, blocks, etc. Stuck valves don't tend to create this type of ping ponging of EGTs.

 

[Dorfie]

The note says you won't get an indication of increased fuel flow.

Don't want to drift the thread, but I am curious. Jetting down with nozzle size would be...unusual. Did you actually make the change, and if so, how much did fuel flow drop?

Returning to topic, whatever the figure might be, the OP's would be less. Moving from 0.028 nozzles to 0.024 nozzles is a total flow area reduction of 26.5%. Totally plugging one 0.028 nozzle on the OP's 540 reduces total flow area by 16.6%.

Dan,
When Don at Airflow Performance did my injector tuning, he replaced all the 0.028 injectors with around 0.025 size injectors. I now have smallest at 0.0245 and biggest at 0.026.
I never questioned his reason for doing this.
Johan

 

[DanH]

Unfortunately I didn't test before / after.

Full throttle fuel flow is unlikely to have changed.

When Don at Airflow Performance did my injector tuning, he replaced all the 0.028 injectors with around 0.025 size injectors. I now have smallest at 0.0245 and biggest at 0.026.
I never questioned his reason for doing this. Johan

Just simple flow balancing so they all peak together, or close to it.

 

[Janekom]

Something erratic like this can be very frustrating. Two weeks ago the guy across my hangar asked my help. He said that he had a rough running engine for a few minutes and did notice an EGT rise on #1. This cleared but a week or so later the same happened on run up but this time on #2. It is an angle valve IO-360. This time it would not clear by itself.

I am attaching a few pictures. The last one shows the debris found and in picture is the end of a pencil in order to give an idea of scale. It was not magnetic and feels like rubber.

Since then he has down a number of flights without any hassle. Of note maybe is that this happened after his regular AME has replaced all four injectors with new Gami injectors. I assume that the debris somehow got in there during this, but who knows.

A few years ago I had a similar event on another plane with a Superior IO-360 where no work has been done on the fuel system whatsoever.

This is not helping Dorfie but shows that debris getting in there is real and does happen.

 

[airguy]

When I put my plane in the paint shop, I also took that time to do some work on the flow divider and injectors. I took precautions to keep everything clean and introduce no debris - and guess what happened?

After paint and reassembly, I did a standard runup to 1700, all looked good. Taxied to the hold-short line, another standard runup to 1700, all good. Onto the runway and poured the coals to it - and about three seconds later as I was starting to pull the nose up, all **** broke loose. I yanked the throttle as the nosewheel was lifting, let aerodynamic braking do the trick and took a mid-field turnoff.

One EGT was cold, engine was running like dogmeat on the other three. Taxied back to the paint hangar and pulled that cold-cylinder injector and found something very similar to the pictures above - some kind of unknown something that get in there and plugged the restrictor. The other 3 cylinders were then running so rich as to be barely running. The 1700 rpm runups with a constant speed prop did not pull enough power (and fuel flow) to dislodge whatever it was - but full takeoff power did, and it migrated from somewhere to the injector.

Cleaned it out, reassembled, full power runup, and an uneventful departure.

 

[rmarshall234]

I’ve mentioned it here before but worth repeating. I think paint shops are a very concerning place when it comes to potential fuel contamination issues. They need to remove fuel caps to do their work and who knows what happens after that. I wouldn’t suggest negligence, just a failure to connect the dots based on there own level of expertise. I would be very careful with an aircraft coming out from paint. YMMV