Hi
Guys;
I
installed RTD temp sensors at various locations under my cowl and recorded their
trends using trendreader. They are converted to jpegs and attached. The colour
of the graph lines was somewhat diminished during the conversion, but should
still be clear enough. These are airflow temps, not coolant or oil. The
probes were positioned as follows; behind left & right rads, behind oil rad,
in the intake after intercooler (MAT), on the firewall tightly between 2 fuel
injection pumps, on the cabin side of my muffler housing under the insulation.
OAT-not in place.
These are results from an hour long ground run at 3000 rpm. I was recording
an 8 min window for better recording resolution. Due to an oil press concern
(more on that in another e-mail), I never flew it today.
Don't pat attention to the OAT temp. I forgot to reposition it in it's spot
outside the cowl when I replaced the cowl, so the temp it is reading is in a
space to the side and behind, but not directly in the airflow from the rad.
Outside air temp at the time of this test was -5C.
The 2 rad temps are somewhat interesting. First it is obvious from the graph
that the pump controller cycles to maintain temperatures, and easy to see where
I switched over to bypass the controller and run at full 12V. When flow is
controlled by a controller the temps are very close, but when bypassed, the
right rad is 7C hotter. The right rad has an unobstructed airflow path behind
it, while the turbo exhaust pipe is close behind the left rad, partially
obstructing airflow exit path. I'd expected some radiant heating of the probe
due to proximity to the exhaust. However the probe is not in precisely the same
spot due to the exhaust proximity. I have an idea to relocate the probes to
ensure precise positioning on both sides. I'll try to do it tomorrow. Coolant
temps at test were 80C into the rads and 60C out.
Also evident is how the oil cooler airflow temp is directly affected by the
rad airflow temp (they have separate ducts). It is also evident that oil
certainly has less thermal transfer than water as the oil temps into the cooler
were 79C and 60C out and yet the air temps are half as hot as the rad
airflow.
Somewhat disturbing to me was the temps at the firewall between the fuel pumps.
I'm unsure how much of this temp is from the circulating fuel through the rail,
back to the header tank, and into the pumps again, or directly from the cowl
air. As it does seem to fluctuate with the rad temps I would guess that it is
directly related to that, however I had originally had a plate in front of the
fuel pumps for mounting the filters, but after several vapour lock incidents I
removed the plate and problems were greatly reduced, but not eliminated. However
OAT's have also greatly reduced, so this is inconclusive as to whether a
deflector plate is a help or hindrance to keeping heat away from these pumps.
I've considered a blast tube, but I hesitate to add more air under the cowl.
I've also given serious thought to moving the pumps and filters to the cockpit
side of the FW. While not a problem now, it may be when I fly it down to the
round-up in Florida next year :-).
The other line is from my muffler housing. This is also higher than expected,
however it is nice to keep my toes toasty warm. I don't believe it gets this
warm in flight. As I have a great heater core already, I'd prefer to reduce
this. There is a 1/2 inch of insulation + carpet, so I really don't feel this.
It is definitely radiant heat from the muffler, however a large part of the cowl
exit air can travel through this duct and is augmented by the exhaust outlet (I
think), so my hope was to measure cowl outlet temp. I'll reposition this sensor
after a test flight.
Actually I plan to reposition all the sensors as I collect data. Unfortunately
I'm only home for the weekend as I fly back to school on Sunday night
(commercial). I'll be home again next weekend to resume
testing.
I
hate to ask for more, but it sure would be sweet if the EM2 has a serial data
output to input into this (or any other) trendreader program. Imagine being able
to lay all 36 trends down on one graph. Then dismissing unrelated trends to
isolate related data. This would be a powerful tool for those that are actively
experimenting with designs.
S. Todd Bartrim
Turbo 13B RV-9Endurance
C-FSTB
http://www3.telus.net/haywire/RV-9/C-FSTB.htm
"Whatever you vividly imagine, Ardently desire, Sincerely believe in,
Enthusiastically act upon, Must inevitably come to pass".