Fluctuating EGTs (IO550)

[Robert C.]

Didn't know a data logger could be so interesting!

F.e., discovered (courtesy of savvyanalysis.com) that on most of my recent flights the EGTs on #4 and sometimes #6 fluctuate more than the other EGTs. The attached screenshot shows one example of EGTs graphed against FF. The LOP pull is very obvious and the rest is a cruise segment. On other flights #4 has also fluctuated more but wasn't higher than the other cylinders.

GAMI spreads have been between .8 and 1gph.

Any hypotheses that I should discuss with my A&P/IA?

Robert

[M20S Driver]

How does CHT look?  

 

[Robert C.]

Just now, M20S Driver said:

How does CHT look?  

 

CHTs look steady; don't fluctuate much including for #4. They range from 220 (#6) to 300 (#5) with #4 at 280 after starting at 285 after leaning.

[Cruiser]

a GAMI spread of .8 - 1.0 is too high for the IO-550. 

Which is your leanest cylinder?

[Robert C.]

3 minutes ago, Cruiser said:

a GAMI spread of .8 - 1.0 is too high for the IO-550. 

Which is your leanest cylinder?

On this flight # 3 and 4 which both peaked last at 14.5 gph (at an altitude of 9500ft).  #6 was the 1st to peak at 15.6gph.

[Jim F]

Did Savvy give you any recommendations?

i would bet you have a piece of scaled off varnish in the orfice that is bouncing around and this causes very quick changes to the fuel spray pattern out of the injector nozzle.  You might be able to see on the babe jar fuel test where the fuel stream will jump back and forth from the piece floating in the nozzle. Ultrasonic clean and test the spray pattern. 

Jim

[Robert C.]

Hi Jim, I was wondering about injector cleaning so good to hear from someone else that it's a possibility. I'm using their "free" service at the moment, so no recommendations from them. Oddly enough I got an e-mail from them just a few days ago trying to persuade me to become a paying customer by flagging that my oil pressure was higher than 75% of the peer group and the oil temps below average. Both are solidly in the green range though and I'm not ready to sign up yep.

[Cruiser]

1 hour ago, Robert C. said:

On this flight # 3 and 4 which both peaked last at 14.5 gph (at an altitude of 9500ft).  #6 was the 1st to peak at 15.6gph.

that spread is 1.1 gal which is way too much. I would focus on improving the balance of FF first. you should be in the < .5 gph consistently.

looking at the graph it appears that all the spikes are to the low side, if this is real, it would indicate that cylinder (LOP) is not getting enough fuel to complete a burn cycle. 

[kortopates]

You won't be able to get an accurate gami spread until you correct the EGT5 problem. If the fluctuations are uncorroborated by CHT4 I would be checking the EGT4 probe connection - make sure it's secure and tight. I wouldn't conclude an injector issue yet without CHT corroboration.


Sent from my iPhone using Tapatalk

[Robert C.]

Thanks Paul. I don't see any anomalies in the behavior of CHT #4 (over multiple flights) so I'll ask my shop to start with checking the probe connections. Any thoughts on the GAMI spreads? Number #4 is never the 1st nor the last to peak. When seeing a .8-1gph spread can adjustments narrow that or should I lean towards buying GAMI injectors?

[Txbyker]

45 minutes ago, Robert C. said:

Thanks Paul. I don't see any anomalies in the behavior of CHT #4 (over multiple flights) so I'll ask my shop to start with checking the probe connections. Any thoughts on the GAMI spreads? Number #4 is never the 1st nor the last to peak. When seeing a .8-1gph spread can adjustments narrow that or should I lean towards buying GAMI injectors?

Robert, I saw something very similar a couple of months ago but with CHT. It looks like you have the graph at 15 second intervals.  I expanded my graph and saw 30-40 degree variations over a couple of seconds, something that would be hard to do even with a flame torch.  Not sure but worth a check. Paul helped diagnose too.

Russ

[Piloto]

I had a similar problem with my UBG-16 EI analyzer. Turn out to be the EGT barrel connectors. I decided to solder wires and problem solved. The voltage coming out from the probes is very low and any small resistance on the connectors will cause a voltage drop and low reading.

José

[carusoam]

I'm only kind of familiar with a few things...

1) an IO550 with a good Gami spread is in the area of 0.1 gph... (my first IO550...)

2) an average one may be a few tenths... (my second IO550...)

3) Thermocouple failures.... TCs when they fail they tend to go off the chart or go to zero... a wire breaks, or it falls out of the mount.

4) the shape of these curves seem to vary a bit... #4 is peaking wildly different than the rest.  #6 is also a bit different, peaking early and dropping off... different than a TC failure...or partial failure...

5) jagged shaped EGTs are often known to be caused by sticking valves.  JPI graphs with this phenomena show relatively uniform in their repetition...

Start with a good slow run-up.  Allow enough time on each mag for the EGTs to Record their peaks...  

Two mags, two peaks...

It seams in the graph above, EGTs are 75-100°F Different from each other, From the beginning...

Long way of saying there is a lot of mis behavior going on.  #4 and #6 are most unique... #4 is just plain ugly...

 

Has your mechanic suggested pulling and cleaning the injectors and testing their flow balance?  6 glass jars, one for each injector, and a timer...  aka baby food jar test....

It looks like something is obstructing the fuel flow in one or two of the injectors and the extra flow is showing up somewhere else...(if that is possible)

It might be helpful to go back through the engine monitor data files and see when this behavior started, or if it has been going on for a long time...

Might be good to talk with the people at Gami... Solve this challenge and get to a really tight Gami spread at the same time?

PP thoughts only, not a mechanic...

Best regards,

-a-

[ilovecornfields]

6 hours ago, Robert C. said:

Hi Jim, I was wondering about injector cleaning so good to hear from someone else that it's a possibility. I'm using their "free" service at the moment, so no recommendations from them. Oddly enough I got an e-mail from them just a few days ago trying to persuade me to become a paying customer by flagging that my oil pressure was higher than 75% of the peer group and the oil temps below average. Both are solidly in the green range though and I'm not ready to sign up yep.

I paid for the Savvy service and thought it was well worth it. Engines aren’t cheap so  getting professional help will likely pay off in the long run.

[kortopates]

5 hours ago, Robert C. said:

Thanks Paul. I don't see any anomalies in the behavior of CHT #4 (over multiple flights) so I'll ask my shop to start with checking the probe connections. Any thoughts on the GAMI spreads? Number #4 is never the 1st nor the last to peak. When seeing a .8-1gph spread can adjustments narrow that or should I lean towards buying GAMI injectors?

You should be able to get much better gami spreads. 0.8-1.0 isn't even believable. Truth is getting good gami spreads is not easy till you get some practice and set up your engine monitor. I suggest you read our Savvy Test profile instructions and do some testing http://content.savvyanalysis.com/static/pdf/SavvyAnalysisFlightTestProfiles.pdf But what @Txbyker Russ said about data sampling rate is your first priority. Before you can collect good data you'll need to set your monitors data sampling rate to its fastest rate, preferably in the 1-2 sec range, or the fastest it will support upto 1 sec rate. If you are truly not getting better than about 1 GPH there is something else wrong. The IO-550's do pretty well, some need gami's but its not the norm yet many will benefit by them. 

Looking at your EGT spread though, if you haven't has the injectors cleaned in awhile I'd suggest cleaning them to see if that can bring in the lean outliers 6, 2 & 1 in closer. I see what you mean about 4, its not the cause of your large spread, but we don't yet know where it will fall when the probe issue is fixed, but I'd guess on the rich side.

Edited December 15, 2017 by kortopates

[sgrooves]

I have an IO-360 but I was having similar issues with my #4 EGT probe. Totally erratic with no pattern, CHT's fine. At times it would max out on the gauge visual.  There were occasional fluctuations on #2 also. This was going on for a few hours despite checking injectors etc. Turns out my #4 EGT probe was on the verge of dying and 2 flights before the annual it died completely. The erratic signals from the probe caused cross-talk on the #2 probe. Sounds like you could be having something similar. The lower reading at full rich also makes me wonder if its an old probe. I ended up replacing #4 and everything was smooth for ~20 hours until #2 started acting out. This time I knew it was likely a probe and swapped it out, no issues after.  The new probes in #2 and #4 are much more sensitive at lower temps than #1 and #3.

[kortopates]

41 minutes ago, sgrooves said:

I have an IO-360 but I was having similar issues with my #4 EGT probe. Totally erratic with no pattern, CHT's fine. At times it would max out on the gauge visual.  There were occasional fluctuations on #2 also. This was going on for a few hours despite checking injectors etc. Turns out my #4 EGT probe was on the verge of dying and 2 flights before the annual it died completely. The erratic signals from the probe caused cross-talk on the #2 probe. Sounds like you could be having something similar. The lower reading at full rich also makes me wonder if its an old probe. I ended up replacing #4 and everything was smooth for ~20 hours until #2 started acting out. This time I knew it was likely a probe and swapped it out, no issues after.  The new probes in #2 and #4 are much more sensitive at lower temps than #1 and #3.

Unlike CHT probes that would last forever, (except the wires at the base of CHT probes eventually begin to break and cause fluctuations), the EGT probes suffer from tip erosion from being directly in the exhaust stream. In other words the tips of the probe are slowly burned off. When they do they need to be replaced as well. The erratic fluctuations though are usually the probe failing electrically (wires at the base of the probe or loose electrical connection). We've also had a lot of issues with the barrel connectors used on the EI probes.  You'll also notice in some planes data you will see a constant low level noise on the EGTs and even CHT probes - this is almost always from poor installation by routing the EGT and CHT harness ty-wrapped directly to ignition leads. For short sections it won't cause problems but when ty-wrapped along ignitions wires from the front to the rear especially on 6 cyl engine it cause a real problem. This is a big no-no spelled out in the installation manual but installers still do it all the time. Why? I think its simply because the average pilot isn't yet very familiar enough with the workings of their engine analyzer and don't realize the issue and don't complain. So the installers continue to do it because they don't get any complaints and therefore think the manufacturers are making a lot a noise about nothing - but your folks data quality suffers. Other non-thermocouple probes are not so sensitive (like oil Pres, Oil Temp, MAP, RPM etc) but thermocouple probe wires need to be routed away from ignition leads and high current cables (e.g., alternator output) to provide the least background noise.

[carusoam]

Photos of the installation of the probes and wires may tell a lot...

pulling the probe to see what is left of it would be telling...

Take a look at the CHTs on the chart... two of them are different than the other four... none of them are marked with a number to know if the two bad EGT lines is related to the bad CHT (blue and red) lines....

 

it may be a smart idea to verify the EGT numbers match the CHT numbers and the cylinder numbers on the monitor...

Best regards,

-a-

[Robert C.]

Hi Anthony, thanks for the suggestions. The color coding for the CHTs are at the bottom of the lower chart (which has the CHTs). It is the same as the EGTs so cyl #4 is purple on both charts.

[Robert C.]

11 hours ago, kortopates said:

Before you can collect good data you'll need to set your monitors data sampling rate to its fastest rate, preferably in the 1-2 sec range, or the fastest it will support upto 1 sec rate.

Looking at the raw data file that the G1000 puts on the SD card it looks like the sampling rate is 1 sec.

[kortopates]

Yes, all G1000's sample once a sec. This is shown on the bottom of your Savvy data page. Sorry I have not looked at your data online just trying to give you some tips to help. Next steps would be to fix EGT4 and then fly the profile as shown on Savvy WOT at altitude. But as recommended, I'd consider cleaning the injectors soonest if it's been a long time.


Sent from my iPhone using Tapatalk

[carusoam]

Thanks Robert!  So much detail on the page, I lost sight of the data tags at the bottom...

 

1) nice that they matched the EGT/CHT colors with the cylinder numbers... (for the memory challenged )

 

2) Another good sign of the CHT portion of the graph...  CHT #5 is usually the one that is the hottest on the O.  

#5 resides behind the alternator...  the hot spot for hangar elves to cut a quarter sized cooling hole(?).

Proof that Even with the hangar elf help, the cooling does improve some, but that cylinder stays the hottest...

 

3) G1000 (1sec) sample rate is faster than JPI (2 sec, when set to the fastest)...  Nice when watching the engine monitor during a run-up...  any faster, the mass the TCs takes too long to react anyways...  earlier JPIs had a tendency to run out of memory quicker when gathering so many data points...

 

4) Other detail....   About your OilT and OilP...  I might be seeing the same issue with my O...  Looks like my vernatherm may be stuck, sending all/most/too much? Of the oil through the cooler...  resulting in chilly oil and higher pressure...  going in to annual soon will look further into this...

 

5) PP WAG... FI6 looks to have lower FF than the others resulting in a higher and sooner peak, and lower EGT and CHT when LOP... going further LOP, #6 will probably stop producing power first...and go cold...

 

6) A second PP WAG...  EGT#4 is not coinciding with CHT#4 very well...  more of an instrument challenge hiding anything that might be going on there...

 

7) Summary...

• Something real is happening causing the Gami spread to be wide.
• Leaning slowly and gathering a real good Gami spread will confirm this observation... (grab  another flight's data from before)
• Inspecting and swapping a pair of EGT probes will probably find the reason #4 is so erratic...
• While in there look at all the baffling.  I’m baffled by the really cool CHT #s 5 & 6...
• If cleaning the FIs, try to catch whatever comes out of them... Expect #6 to might be hiding a piece of black speck(?) or something...
• post another flight including start-up, run-up all the way through shut down... see what else we can see....

 

 

Thanks to Paul, for sharing his Savvy expertise....  

 

Best regards,

-a-

 

[PaulM]

the jagged line on that one purple EGT probe is a classic dying probe.. if the TC shorts out anywhere along the wiring (or in the probe) you get lower readings.   All of those excursions are down. 

[Robert C.]

Update for those of you who still might be interested

We swapped the probes for #2 and #4; Also, did a flow test on all 6 injectors. All injectors tested as clean and undamaged.

Flew over 2 hrs today and the EGTs now no longer bounce all over the place, so it looks like moving the probes around solved that problem.

While troubleshooting all this I did look harder at CHTs (during cruise). I noticed that cyl #2 & #3 sometimes have a wave pattern with CHTs rising and declining by 10F. Frequency can be anywhere from 1-6 minutes. The others are more stable. Any thoughts on whether this is worth worrying about?

Thanks for all the input and suggestions!

Robert

[Txbyker]

1 hour ago, Robert C. said:

Update for those of you who still might be interested

We swapped the probes for #2 and #4; Also, did a flow test on all 6 injectors. All injectors tested as clean and undamaged.

Flew over 2 hrs today and the EGTs now no longer bounce all over the place, so it looks like moving the probes around solved that problem.

While troubleshooting all this I did look harder at CHTs (during cruise). I noticed that cyl #2 & #3 sometimes have a wave pattern with CHTs rising and declining by 10F. Frequency can be anywhere from 1-6 minutes. The others are more stable. Any thoughts on whether this is worth worrying about?

Thanks for all the input and suggestions!

Robert

Paul suggested to me that could be due to the fact that the factory tie wraps the probe leads to the plug wires.  I see the same thing although sometimes sporadic.

Russ