Experience with Subaru at 358 Hours

DanH said:

More flywheel inertia is an absolute good for torsional behavior, as you've now seen for yourself. It will, however, add one oddity in the case of your Marcotte gearbox, which rotates the prop in the same direction as engine rotation. The additional inertia will add to gyroscopic precession. I think you'll notice the difference, although it should not be a big deal with the 6A. My scaled JN-4C (or an original Camel) was more interesting because engine/prop inertia was much higher in proportion to airframe mass.

An ideal PSRU would rotate the prop opposite the engine so the prop and flywheel gyroscopes offset each other. In theory you could have less precession than a Lycoming.

Click to expand...

I'd read about this fact on EPI's site but forgotten about until you mentioned it here again. I'll watch out for it and see what I notice. For the time being, I am very pleased with the change and this mirror's the experience of two others who switched to heavier flywheels on the Marcotte. Once I get further along in testing and have some time, I want to wade into the MOI and stiffness models. Still very interested in what could be happening further up the rpm range.

I will do some more testing on the cooling system today as it should be warmer today.

Snowflake said:

Wow, Ross, that looks pretty darned good to me. I can't wait to hear how it performs!

Click to expand...

Glad you like it. Not as elegant as a P51 but I didn't have a clean sheet of paper to start with.

I'll be adding some more instrumentation today to gather more accurate data.

Ross, even though I do not have plans to use a Subaru engine on my RV-7, I do appreciate your excellent documentation and attention to engineering details as shown in this thread. There is certainly a place in Experimental aviation for auto conversions, and that part of our sport is better served by folks like you who openly discuss the many considerations of such an endeavor.
Thanks for sharing your continuing experience.

rv7boy said:

Ross, even though I do not have plans to use a Subaru engine on my RV-7, I do appreciate your excellent documentation and attention to engineering details as shown in this thread. There is certainly a place in Experimental aviation for auto conversions, and that part of our sport is better served by folks like you who openly discuss the many considerations of such an endeavor.
Thanks for sharing your continuing experience.

Click to expand...

I appreciate the positive post. I am very interested to gather the pressure and velocity information with this rad setup. Other than some serious studies done by NASA with regards to very high altitude radiator design (with the math way over my head), there does not seem to be any current information on aircraft rad design or performance.

I have done some historical study from WW2 and there is some new info coming to light from others doing archival searches now in British and German records. From this, it is clear that most peoples attempts these days don't even approach what was being done 70 years ago. They are ignoring even basic design considerations with regards to the duct and exit geometries so they have gigantic rads with sometimes marginal cooling and a bunch of drag and weight.

The speculation about the P51 radiator performance is as amusing as it is prolific on the web. People have no numbers to go by and throw all sorts of nonsense speculation out there.

If we look at inlet areas, rad face areas and volumes compared to hp ratios, aircraft like the Spitfire, P51 and ME109 are very impressive. The Spitfire has less face area than many homebuilts yet cools a 1600hp engine. Even then, I think the Spitfire setup could be improved on considerably.

In civil aviation, we must also have good ground cooling, unlike most of the warbirds, so we'll have to make some design changes to the rad shape but all the duct geometry work is directly applicable to increase cooling while reducing drag. Very few liquid cooled homebuilts even have a variable geometry exit door despite huge evidence showing how important this is to reducing cruise drag. Historical testing shows that aircraft would gain anywhere from 10 to 30 mph by closing the exit up at high speeds.

Others ignore the historical NACA studies showing how poorly NACA ducts work for feeding radiators and use them anyway, even on clean sheet designs. I've posted one modern homebuilt study again supporting that conclusion from Keith Spreuer.

It should be interesting anyway.

WW2 Radiator and ductwork designs

I've never studied the WW2 designs, but have the opinion that under the tremendous schedule pressure of wartime, much of the airflow and cooling design and testing (as on the RR Merlin and Griffon engines) were not totally documented, although they had to have some sense of optimum duct and heat transfer design. Surely, they did wind tunnel studies, but perhaps the "tweaking" details were stored in human minds and not fully stored on paper. Or perhaps, the paper records have been lost.

My hat is off to you, and hope you have tremendous success with optimizing your design. It is interesting and should be rewarding.

rv7boy said:

I've never studied the WW2 designs, but have the opinion that under the tremendous schedule pressure of wartime, much of the airflow and cooling design and testing (as on the RR Merlin and Griffon engines) were not totally documented, although they had to have some sense of optimum duct and heat transfer design. Surely, they did wind tunnel studies, but perhaps the "tweaking" details were stored in human minds and not fully stored on paper. Or perhaps, the paper records have been lost.

My hat is off to you, and hope you have tremendous success with optimizing your design. It is interesting and should be rewarding.

Click to expand...

There was a lot of paper documentation on both sides and now people are digging this information up and publishing. Mainly flight test data rather than engineering data on the ducts and rads. Fascinating reading.

Hot Idle Tests

Today, we had unusually warm weather for early May so I took the opportunity to validate the ground cooling with the new rad setup. I installed the cowling and splitter for these tests.

OAT was 27C. I was facing into a 10 knot wind component, about 40 degrees off the actual mean wind direction. Idle speed 1000 engine rpm. Rad door fully closed. I ran the engine for 35 minutes and the engine temp never exceeded 90C. I opened the rad door fully and saw a decline in T2 after a couple minutes.

I then opened the door fully and increased idle speed to 1600 rpm. ET fell to 80C within 1 minute and T2 decreased 10C. I then increased rpm to 2000 and T2 dropped further, ET stayed at 80C. Total run time was 45 minutes. At 1600 rpm and above, the engine is essentially operating right at the thermostat opening point.

Pretty happy with this. No electric cooling fan, 130 square inches of rad area and 64 square inches of heater core/supplementary cooler area in the cowling fed by a 3 inch hose about 12 inches below the spinner centerline. A 78C thermostat is fitted. Main rad hoses are 3/4 inch, the heater core is plumbed with 5/8 hose. The cooling is substantially better than before on the ground.

Something to note yesterday, when I was running without the cowling, the cockpit air vents had no air coming out even at 2000 rpm. This demonstrates how sensitive NACA ducts are to upstream influences. Today with the cowling on, flow was pretty substantial at 2000 rpm.

Let the pain begin...

Dan, I am starting to crunch numbers on the MMOI for the gearbox parts. You warned me it would be easier to take the box apart and swing the parts... Right again. This is painfully slow to calculate. One hour to wrap my head around the formulas, get densities, convert the measurements just on one part. 7 more parts to go. I have to break most parts into 2-5 other parts and calculate each of those individually, so there are really about 15 "pieces". Shoulda listened.

Question: how do we estimate MMOI for the ball/ roller bearings on the shafts- inner race plus half the mass of the balls/ rollers?

This certainly clears out the cobwebs...

After that, on to the stiffnesses which will involve similar calcs.

rv6ejguy said:

I have to break most parts into 2-5 other parts and calculate each of those individually, so there are really about 15 "pieces".

Click to expand...

There were 37 to do a simple cog belt drive.

Question: how do we estimate MMOI for the ball/ roller bearings on the shafts- inner race plus half the mass of the balls/ rollers?

Click to expand...

Close enough.

More Ground Running

I got the instrumentation installed on the rad duct yesterday and did another 40 minutes of ground running today. The ground cooling is amazing, coolant temp never got over 70C, OAT was about 17C. I believe this thing would idle all day at 80C on a +22-24C day.

The second, very interesting thing was playing with the rad exit door at 3000 engine rpm (1364 prop rpm). With the door fully open, exit velocity was 25 knots, close it and I got almost 40 knots. Looks like the WW2 engineers had this stuff figured out as far as recovering cooling air momentum. We can certainly regain a lot of velocity by closing up the exit, will be very interesting to see what happens at 150-160+ knots.

So far I am seeing great things from this very time consuming mod. Now if it works as well in flight as it does on the ground...

Back in the Air

After waiting over a week for suitable test flight weather, it finally came tonight. Wind favored the long runway and traffic was light. Good to get up after 6 months. I was task saturated as they say despite having a nice checklist and a video camera recording a bunch of engine and radiator parameters.

Anyway, nothing fell off. #1 radio needs attention and I will pull the cowls and spinner tomorrow to check things out.

Short vid taxiing out here: http://s1105.photobucket.com/albums/h341/rv6ejguy/?action=view&current=MAH00012_zps0b3fb892.mp4

I had a quick glance at the radiator pressure delta at 100 knots IAS, had 5 inches H2O inlet and 1 on the outlet, exit V was 50 knots, all with the door fully open. Coolant temp never budged from 70C, Oat was 17C.

I'll have to review the cockpit video to see some of the other readings.

Earlier today, I did a test of the redrive bushing stiffness. These were softer than expected at 25 lb./ft/ deg. I have all the other stiffness and inertia calcs done so I can start plugging things into the spreadsheets soon.

rv6ejguy said:

After waiting over a week for suitable test flight weather, it finally came tonight. Wind favored the long runway and traffic was light. Good to get up after 6 months. I was task saturated as they say despite having a nice checklist and a video camera recording a bunch of engine and radiator parameters.

Anyway, nothing fell off. #1 radio needs attention and I will pull the cowls and spinner tomorrow to check things out.

Short vid taxiing out here: http://s1105.photobucket.com/albums/h341/rv6ejguy/?action=view&current=MAH00012_zps0b3fb892.mp4

I had a quick glance at the radiator pressure delta at 100 knots IAS, had 5 inches H2O inlet and 1 on the outlet, exit V was 50 knots, all with the door fully open. Coolant temp never budged from 70C, Oat was 17C.

I'll have to review the cockpit video to see some of the other readings.

Earlier today, I did a test of the redrive bushing stiffness. These were softer than expected at 25 lb./ft/ deg. I have all the other stiffness and inertia calcs done so I can start plugging things into the spreadsheets soon.

Click to expand...

Pleased to read the good report, Ross.

Your diligence in this matter is impressive. The darned things can be made to work.

Up early this morning to beat the rush and take advantage of a dead calm, blue sky day for the 2nd test flight.

Overhead the airport at 6000, I was able to video the engine parameters, observe and record pressures, temperatures and exit velocity on the rad plus evaluate some of the other mods. OAT was around 11C at altitude.

Cooling is way better with the new ventral rad. Temps never budged off 70C.

Rad delta P with the exit door open at 100 KIAS was 3 inches H2O, exit V 50 knots.

Rad delta with the door about half closed at 120 KIAS, 2.5 inches, Exit V 70 knots.

Rad exit temp stayed around 35C.

The new front intercooler location is worlds better than the old firewall location. Average delta T inlet to outlet was 40C at low power settings and airspeed, max observed was 46C. Down to around 10-12C over ambient, which I am very happy with.

Oil temp didn't get over 81C even in the climb at 85 knots. Now use one of the old cowling rad feed ducts to blow directly on the pan.

Redrive temp never got over 45C, roughly 15C cooler than before under similar conditions. Now have an old intercooler feed duct in the right cheek blowing on the gearbox case.

The airplane appears to be at least 5 knots faster than before at 25 inches and 4500 rpm, whether this is from the fresh engine or reduced drag or both is hard to say.

After I landed, the crankcase breather showed not a speck of oil behind it. Before at the last, the whole belly would be black in an hour.

So, all the mods seem to be working well. Pretty happy so far.

I need to fix one radio and modify the rad door actuation. It is very hard to move closed in flight due to air loads.

Need to experiment with the rad door more to really quantify performance gains.

Congratulations!!

rv6ejguy said:

So, all the mods seem to be working well. Pretty happy so far.

Click to expand...

Ross, great news

I like the way you document the changes/effects------instead of just "it feels like......"

Hope all the numbers just keep getting better as you continue the testing.