Peter wrote:
In comparison, the cockpit is large and has loads of leaks.
That is a very good and obvious point which I totally missed… I guess I should revisit the “bigger is better” approach…
So I bought me a couple Bosch BME280 sensors and wired them to a Raspberry Pi. Running on a breadboard they both were fairly close to each other (within the nominal tolerances). I moved one to the top of my “dehumidifier tower” (a 110mm diameter PVC pipe), where the egress port is; ingress is at bottom. When the pipe was sealed and empty, the humidity inside measured about 45%, a good 10% more than ambient. When I added fresh silica gel (kitty litter, about 2kg), the humidity went up, to a whooping 60% (outside is steady at 35%). Sitting on my work desk, at about 19 degrees Celsius, 987hpa, both inside the pipe and out.
The aquarium air pump (rated at 300 liters/h) driving the air hasn’t seen an aquarium in more than 5 years, plus it has been running since yesterday, so any residual moisture should be gone.
The air is passing through the tower properly, it will eventually blow water bubbles if the egress , and if I block the egress port, the pressure inside goes up by 10hpa (any the excess leaks out – same as it did when it was empty).
So, what is happening and why? Right now it looks like I’d be better off not attaching it to the engine.
Yes, I’ll bake the remaining 2kg of the silica gel overnight and try that tomorrow… But that will not explain the RH rise when empty.
Did compare your device against another hygrometer to make sure it is working? Or at least took a measurement in known conditions (100% RH)? What I build almost never works rightnow out of the box.
When I test-ran the two sensors on a breadboard they behaved consistently with each other and simulated changes in environmental conditions (placed over a cup of hot water, etc). And I did compare them to a “off the shelf” weather station – again, not a calibrated setup from Davis, just a household thing from Lidl that happens to measure humidity. So nothing in the sense of “calibrated for meteorological purposes” but enough to consider their readings resembling reality.
And yes, I did fry one by running I2C over a twisted pair.
When the air from the “dry” end is pumped into a cardboard box with the weather station, the RH readout goes up from the 35% “room value” – so far up to 45%, but I believe that is averaged over a longer period of time. At any rate, air leaving the “dehydrator” is more humid than air entering it. Off to the oven…
So, after workday in a ~120 degree Celsius oven the weight of the silica gel has changed to 1300g. Full disclosure – yes, I am a dumb ass and didn’t weigh the batch when I put it into the oven, so I don’t really know how much weight got baked out. I estimate it was around 1900g originally, from eyeballing the remaining volume of the 3.8kg bag of kitty litter on my desk. “The Internet” says silica will adsorb up to 40% of its own weight, so perhaps plausible, although I think Peter quoted figures closer to 10%. Probably a function of how much / fast it adsorbs. Regardless, I will remediate this when I bake out the rest of the bag and the stuff that is in the plane.
The RH of the air exiting the contraption is now reported at around 1% RH or below, so about 30% drop in RH. Win!
Two lessons, with the most important one being BAKE IT BEFORE YOU USE IT or the cure might be worse than the disease. The second[ary] lesson, based on weighing every couple of hours is that 2h at 120C are almost enough and 4h is definitely enough. Anything above that is wasted (weight did not go down any measurable amount between hours 4 and 8 of the cycle). Will hook up a electric power meter to the stove next time, if I don’t forget.
Guess I now have a totally legitimate reason for a trip to the airfield tomorrow.
If nothing breaks I’ll be back with a humidity over time report in a couple of weeks. If only I’d have thought of getting the Pi some internet access…
I have extracted the EDM700 log and here are two profiles.
Blue trace is #1 CHT and red trace is the oil temp. Both F.
The first is the first 30 mins of a ground run. This was at 1400rpm after the green arc was reached
This one is a normal flight; again the first 30 mins
So there is not a dramatic difference between how quickly the oil working temp is reached, between a flight and a ground run. This disproves the widely quoted theory that ground runs are useless.
EDIT: the x axis is 30 mins long in each case.
@ Peter and the charts above
I could always be wrong. But I think you are missing the same point which you made yourself before. For the oil to loose all water contents it has to be heated up and stay at that (minimum) temperature for a while, how long I do not know. Your cowling is built to make sure that “some” normal temperature distribution across the engine block and cylinders is maintained in flight. But on the ground the temperatures of the different engine parts are different. One can generously extend the “warm-up” flight by 10, 20, 30, or 60 minutes – things only get better.
But I doubt you want to extend your ground runs by the same policy – more is better – even when your gauges (CHT measurement spots, nothing else ?) show the same temps as in flight. Instead I think the owner’s of Lyco/s doing ground runs are trying to find a balance between “not too long” (not enough cooling, too cool in some spots) and “not too short” (still water in oil). But there is no objective measurement to strike this balance. It’s just guess work which over time can become very expensive.
What do you think ?
I think my point (see my edit above) is that it is not hard to get the oil up to working temperature. It takes some 15 mins.
Then the debate is how long it needs to stay there for. I reckon at least 30 mins. When I do a flight I never fly for less than about 50 mins airborne time. Less than that I can drive 
Regarding uneven temperature distribution around the crankcase, I don’t think this is happening, due to oil circulation. See here. The portion of the crankcase where the temperature indicating stickers were attached is showing something less than +88C which is less than +190F, which corresponds well with the oil temp of ~ +170F which is measured at the end of the oil gallery near #1 cylinder. Normally, a circulating liquid distributes the temperature fairly well.
What we can’t say that this is true for all engines and all engine (baffle configuration) installations. There might be hot spots. Also as you can see the IO540 has an oil thermostat.
I certainly would not do the engine run method if I did not have multicylinder temperature monitoring.
Of course we need to FLY for pilot currency reasons, too. Way too many pilots have poor currency and there is a pile of fatal crashes every spring.
I was allowed a ground run and some fast Taxiing down the Runway. 45 mins run, power checks at various holds, got all the CHT’s up to 390 degrees F.
Then got the engine cooler by idling, went onto Runway, held on the brakes at full power for up to 2 mins watching CHT’s all of the time, then release brakes and 60 Kts down the runway multiple times. I wouldn’t like to do what I did without CHT probes on all cylinders.
As usual most people don’t care, lots of Aircraft sitting idle which haven’t moved for weeks/months. My engine can sit all day at 1500 RPM without overheating but I don’t think the oil would get hot enough without thrashing it a bit! Oil temps ended up similar to the cruise but is this enough to re-establish the varnish type layer on the cam? It certainly doesn’t meet Lycoming’s Service Letter L180B and the oil wasn’t up to temp for 1 hour.
