WOT Climb at altitude, CHT3 drops while others rise ?

I would like any expert advice on the following ...

When in a climb at full throttle and when climbing from altitude at full throttle, my CHT #3 drops. In the following chart, I start at 140ft MSL and climb to 8,500ft. Later in the flight I descend to 6,500ft and shortly there after climb back up to 8,500ft. The entire flight is a little more than 3 hours. This is a carbureted O320 with a fixed propeller.

I've notated the chart with arrows for the two periods.

Glen,
Hard to tell from the graph. Is is just always lower? Looks like it tracks with the others. Need more granularity on the graph. How do the egt's track on #3?

At cruise power which is about 65% (?) they track. In the graph I start with a climb to 8500'. At which time I throttle and lean. My MP was around 20 and my RPM was 2520. At which point the CHTs settle and track. The blip latter is a descent to 6500' followed by a WOT climb back to 8500'. Here again, CHT3 doesn't rise with the others.

I'll throw this out for comment: As a carbureted engine, the mixture distribution among the 4 cylinders will likely be uneven. You note that the issue occurs when you are WOT.

Some pilots have noticed that if they try to run LOP in a carbureted engine, they get better results by either closing the throttle slightly, or adding a bit of carb heat.

These measures probably alter the mixture distribution among the cylinders a little bit.

Might be worth a try to slightly close the throttle from full and see if anything changes. Apparently the throttle plate, which is no longer parallel to the airflow, introduces some mild turbulence in the airflow.

Standing by for "incoming"

Pete, I've tried most of the tricks for getting LOP

You comment does remind me of another piece of information. For the anomaly around 11:40, I was lean at cruise. I did not enrich for the climb from 6500' to 8500'. Only added throttle to max and pitch for 100kts.

Airflow during a climb changes and does random things depending on your installation. For example my C1 gets hotter, yet its at the front!

In theory the F/A ratio should stay the same when you advance the throttle, but the distribution could be changing and you might be climbing with C3 LOP and the others slightly richer. You need a full data set again to tell. Preferably EGT/CHT/FF/MP/RPM and then if available IAS and ALT etc.

There is a lot of noise on the graph, are you sure it is not a C3 CHT probe wire pulling the probe off contact a bit with the climb?

To study better you need a better data set display. One reason I suggest people get a good education in these things is so they can diagnose these things easily themselves with a full display on their PC screen.

humptybump said:

Pete, I've tried most of the tricks for getting LOP

You comment does remind me of another piece of information. For the anomaly around 11:40, I was lean at cruise. I did not enrich for the climb from 6500' to 8500'. Only added throttle to max and pitch for 100kts.

Click to expand...

So in essence you were enriching in the climb. So if you expand the graph and look at this as an enriching event, how does it compare to the reverse, the leaning event? Same cylinder get hot? (or cool last depending on the time direction)

Hi David, I have a full data set - about 50 columns in total and recordings each second for 3.4hrs (actually too much data for Excel to handle as it gives me an error that I'm trying to graph more than 32,000 data points).

I'll look at all data for the 5 minute anomaly window. Since everything was stable before and after that point, it should help find any tell tails.

The plane is running better than great. This is an education exercise. If I didn't have all the data, I'd never know to even ask the question.

I'll see what I can learn when I add back in EGTs, MP, RPM, FF, FP, OAT, ALT, and, IAS.

Bill, with a little closer scrutiny, I should be able to answer your question.

Hi David, I have a full data set - about 50 columns in total and recordings each second for 3.4hrs (actually too much data for Excel to handle as it gives me an error that I'm trying to graph more than 32,000 data points).

Click to expand...

Weed out the stuff you dont need and study what you do. This is my prefereed method. Deakin loves graphs, I love data. One of the few things we are at odds with, but then there are times when he wins