johnny stick
Well Known Member
Bob, thanks for presenting your results to date in a chart form, this is good stuff.
Van's Air Force
Don't miss anything! Register now for full access to the definitive RV support community.
Forming aluminum for a cooling outlet fairing
- Thread starter Bob Axsom
- Start date
A good idea
Any credible test program would need that. I want to try to see what I can get out of this strictly for the speed of this airplane so I am motivated a little differently. Test 14 in this series was a little disappointing - I am starting to visualize a combination of forces at play and I thought the relief of the cutouts in the cover would have a greater positive effect than I saw.
With the test rsults sorted by speed I think I have a better direction for the next step:
Here are some things I think are true from the test results so far:
The bump is better than no bump
Restriction in the outlet path can slow the plane down
There are a lot more things that I suspect but I do not know so I will just stew on those as I proceed. I am going to go back to test configuration #10 and re-fly that. I can rationalize that it is the best configuration tested so far just by direct observation but even with a test method error margin of 2kts it is still well within the top group tested.
Test configuration 7 and the related long cover configuration that was not tested but appeared to perform well but was shattered are also on my mind. I am not convinced that the long tail bump is better than the symetrical bump and I am not convinced that there is nothing to be gained from some form of cover - yet.
For now I'll re-fly test configuration #10 which I hope will not disappoint me. If test #10 results are validated I can focus on optimizing that configuration for a while.
Bob Axsom
Any credible test program would need that. I want to try to see what I can get out of this strictly for the speed of this airplane so I am motivated a little differently. Test 14 in this series was a little disappointing - I am starting to visualize a combination of forces at play and I thought the relief of the cutouts in the cover would have a greater positive effect than I saw.
With the test rsults sorted by speed I think I have a better direction for the next step:
Here are some things I think are true from the test results so far:
The bump is better than no bump
Restriction in the outlet path can slow the plane down
There are a lot more things that I suspect but I do not know so I will just stew on those as I proceed. I am going to go back to test configuration #10 and re-fly that. I can rationalize that it is the best configuration tested so far just by direct observation but even with a test method error margin of 2kts it is still well within the top group tested.
Test configuration 7 and the related long cover configuration that was not tested but appeared to perform well but was shattered are also on my mind. I am not convinced that the long tail bump is better than the symetrical bump and I am not convinced that there is nothing to be gained from some form of cover - yet.
For now I'll re-fly test configuration #10 which I hope will not disappoint me. If test #10 results are validated I can focus on optimizing that configuration for a while.
Bob Axsom
Last edited:
johnny stick
Well Known Member
the difference between runs 4 and 6 are significant and probably outside of the margin of error. I think here is a clue. By looking at the data, maybe the long rectangular fences and/or long cover is bringing the cowling air too far into the high pressure area from the wings. I am also wondering if the outside edges are seeing up flow from the wings or a slow down of the air on the underside due to influence of the wings. The air in the outlet area may not be going straight back along the fuse, but rather have an outward component; maybe the outside fences are catching this and that is why the rectangular are slower. maybe maybe maybe??
johnny stick
Well Known Member
Why not try flow fences along the fuselage left and right sides just ahead of the leading edge maybe extending down the wing a little bit? Isn't this similar to the triangle fins but slightly more outboard?
It did not
I flew test case 10 again and the speed dropped off to 179.3 kts TAS. Then I reinstalled the vented cover and reflew test case #14 as suggested by Kevin Horton. It did not hold up either - 178.3 kts TAS. The hierachy is the same but the separation is reduced:
More thought required.
Bob Axsom
I flew test case 10 again and the speed dropped off to 179.3 kts TAS. Then I reinstalled the vented cover and reflew test case #14 as suggested by Kevin Horton. It did not hold up either - 178.3 kts TAS. The hierachy is the same but the separation is reduced:
More thought required.
Bob Axsom
Last edited:
May 6, 2012 AM thoughts
1 - Tests 15 and 16 show in the current form the uncovered version is still faster than the covered version even though there was a 2.8 kt drop in speed from the earlier uncovered test 10 and only a 1.5 kt drop from the repeated covered test 14.
2 - Tests 8 through 16 do not clearly demonstrate that the long tail bump is faster that the equal tail bump. A replacement for the operationally destroyed equal tail bump needs to be made and tested further.
3 - The results from test 13 are are very little different from those of test 15 (-0.2 kt) so the percieved ineffectiveness of the small triangular fins when compared with the results of test 10 may not be valid even though it is still slower.
4 - Tests 10 and 15 were flown with the aluminum angle and all the open platenuts for mounting the cover exposed to the open air. If new fins were made for this "no cover" version without cover mounting provisions would it be measurably faster?
5 - Test 13 seems to indicate that the outboard flow fences have to be taller than the bump by some amount for the bumps to be effective.
6 - Is the center fin necessary or does it add drag with no benefit?
7 - All tests have been with a 2" high bump, would a shorter bump be better?
First, I think I will answer the questions in thoughts 4 and 6. This could be done by making similar shaped fins from 0.032 aluminum, assemble, fly a test then mark and cut away the exposed portion of the center fin and fly another test. Gross differences in performance should be revealed. I may be able to do that today.
Bob Axsom
1 - Tests 15 and 16 show in the current form the uncovered version is still faster than the covered version even though there was a 2.8 kt drop in speed from the earlier uncovered test 10 and only a 1.5 kt drop from the repeated covered test 14.
2 - Tests 8 through 16 do not clearly demonstrate that the long tail bump is faster that the equal tail bump. A replacement for the operationally destroyed equal tail bump needs to be made and tested further.
3 - The results from test 13 are are very little different from those of test 15 (-0.2 kt) so the percieved ineffectiveness of the small triangular fins when compared with the results of test 10 may not be valid even though it is still slower.
4 - Tests 10 and 15 were flown with the aluminum angle and all the open platenuts for mounting the cover exposed to the open air. If new fins were made for this "no cover" version without cover mounting provisions would it be measurably faster?
5 - Test 13 seems to indicate that the outboard flow fences have to be taller than the bump by some amount for the bumps to be effective.
6 - Is the center fin necessary or does it add drag with no benefit?
7 - All tests have been with a 2" high bump, would a shorter bump be better?
First, I think I will answer the questions in thoughts 4 and 6. This could be done by making similar shaped fins from 0.032 aluminum, assemble, fly a test then mark and cut away the exposed portion of the center fin and fly another test. Gross differences in performance should be revealed. I may be able to do that today.
Bob Axsom
Last edited:
kentb
Well Known Member
Bob, a couple of thoughts.
This most likely will not help you, but....
In looking at your speed sorted chart and dividing into good speeds and not so good speeds (I chose 180 as the break point). The only thing that can be eliminated is the "extended short cover". All other parts of the configuration appear both above and below the dividing line.
I believe that what you are seeing is the changes interact with each other and it will be hard to deduce the correct parts to use.
The only only hope will be to continue to try all the combination that you can think of and see which you like the best (fastest of course).
Good luck. I love you efforts and reports.
Kent
This most likely will not help you, but....
In looking at your speed sorted chart and dividing into good speeds and not so good speeds (I chose 180 as the break point). The only thing that can be eliminated is the "extended short cover". All other parts of the configuration appear both above and below the dividing line.
I believe that what you are seeing is the changes interact with each other and it will be hard to deduce the correct parts to use.
The only only hope will be to continue to try all the combination that you can think of and see which you like the best (fastest of course).
Good luck. I love you efforts and reports.
Kent
RVbySDI
Well Known Member
Bob,
With the test rsults sorted by speed I think I have a better direction for the next step:
I am getting an error message on your images stating they have been moved. Any thing you can look into?
Thanks Kent
It has been apparent to me over the last few days that I am not dealing with a single variable even when I change only one variable. I agree with your observations and conclusion.
All I got done yesterday was pull the parts off to replace the three vertical elements. Will try to do more today. I am trying to force myself to exercise properly for my health. I have been working until 3 am many nights until I was too tired to exercise so yesterday I switched priorities and you can imagine the result.
I also deleted some of the obsolete tables from photobucket and learned not to do that again. It effects the thread here in a couple of ways. The images are linked and not really resident here so when they are deleted from there they are deleted from here also. There is usually a notice posted here so there is some awareness. However, I found one image that was not in this thread before so at least under some conditions the link can be associated with the wrong image. In the case I saw it was a different subject all together (a magazine cover) but some different image re-links could go undetected and provide incorrect information. I will not do that again. Well it's afternoon and I haven't exercised so I had better get at it or I will not get the change for retest done today either.
Bob Axsom
It has been apparent to me over the last few days that I am not dealing with a single variable even when I change only one variable. I agree with your observations and conclusion.
All I got done yesterday was pull the parts off to replace the three vertical elements. Will try to do more today. I am trying to force myself to exercise properly for my health. I have been working until 3 am many nights until I was too tired to exercise so yesterday I switched priorities and you can imagine the result.
I also deleted some of the obsolete tables from photobucket and learned not to do that again. It effects the thread here in a couple of ways. The images are linked and not really resident here so when they are deleted from there they are deleted from here also. There is usually a notice posted here so there is some awareness. However, I found one image that was not in this thread before so at least under some conditions the link can be associated with the wrong image. In the case I saw it was a different subject all together (a magazine cover) but some different image re-links could go undetected and provide incorrect information. I will not do that again. Well it's afternoon and I haven't exercised so I had better get at it or I will not get the change for retest done today either.
Bob Axsom
kentb
Well Known Member
Bob, you're on the right track.
And I don't be the flight testing. We all need to take care of our bodies, which means our minds too. A rested and health test pilot will always collect the best information.
Rest, exercise and have a good meal. Than go back to doing flight testing.
Kent
And I don't be the flight testing. We all need to take care of our bodies, which means our minds too. A rested and health test pilot will always collect the best information.
Rest, exercise and have a good meal. Than go back to doing flight testing.
Kent
I saw the error myself
I have all of the images and can supply them to you. I guess I owe it to everyone to go through the thread and put the images back in or insert the updated table. I have the excel files but in order to get them presented here properly I have to print then, scan them, move them to the desk top, import them to iphoto, edit them by cropping to the proper size (the size of the table), upload to to photobucket, copy and paste as an image here. Maybe I will just do that for each table configuration that was in the thread - it is kind of important to the continuity of it all. That will be tonight's task. Jeanine always said I have OCD.
Bob Axsom
I have all of the images and can supply them to you. I guess I owe it to everyone to go through the thread and put the images back in or insert the updated table. I have the excel files but in order to get them presented here properly I have to print then, scan them, move them to the desk top, import them to iphoto, edit them by cropping to the proper size (the size of the table), upload to to photobucket, copy and paste as an image here. Maybe I will just do that for each table configuration that was in the thread - it is kind of important to the continuity of it all. That will be tonight's task. Jeanine always said I have OCD.
Bob Axsom
Camera & tufts?
Bob,
I truly applaud all the work you are doing. We will all benefit.
While I know part of the enjoyment of this excercise is getting to fly the plane, you've burned a lot of AvGas flying straight and level very carefully.
Do you think it would help to tuft the airframe and video it in flight? There are enough interactive variables that a person more skilled than I (i.e., you) might spot trends.
During testing your speed differentials have been on the order of 1 - 2%, which are pretty difficult to attribute solely to airframe modifications since variations in temperature, pressure and humidity are wider than that.
Larry Tompkins
544WB -6A
W52 Battle Ground WA
Bob,
I truly applaud all the work you are doing. We will all benefit.
While I know part of the enjoyment of this excercise is getting to fly the plane, you've burned a lot of AvGas flying straight and level very carefully.
Do you think it would help to tuft the airframe and video it in flight? There are enough interactive variables that a person more skilled than I (i.e., you) might spot trends.
During testing your speed differentials have been on the order of 1 - 2%, which are pretty difficult to attribute solely to airframe modifications since variations in temperature, pressure and humidity are wider than that.
Larry Tompkins
544WB -6A
W52 Battle Ground WA
The fellow in the hangar next to me suggested that
We all have our ways and many work. My way is to go for the objective directly. When the outlet configuration allows the plane to go faster the plane will be faster regardless of the way it came to be. I just can't use the camera and yarn approach to in hopes that when the tufts align with the flight path that is proof that all of the physical variables of the outlet are optimized. It just can't be true that there is only one configuration that will yield straight after outlet flow and that configuration is also the fastest configuration. The speed is ALL I care about.
I guess I am committed to go back and try to fix my problem with the posted images now - that pains me.
Bob Axsom
We all have our ways and many work. My way is to go for the objective directly. When the outlet configuration allows the plane to go faster the plane will be faster regardless of the way it came to be. I just can't use the camera and yarn approach to in hopes that when the tufts align with the flight path that is proof that all of the physical variables of the outlet are optimized. It just can't be true that there is only one configuration that will yield straight after outlet flow and that configuration is also the fastest configuration. The speed is ALL I care about.
I guess I am committed to go back and try to fix my problem with the posted images now - that pains me.
Bob Axsom
rvmills
Well Known Member
Bob,
Been watching this each night as I toil away at the other end of my cowling. Haven't posted in this thread in a bit, but lots of interesting things going on.
I'm glad you are re-testing various configurations and doing multiple comparative tests. With the numbers so close, its hard to tell whether results are in the scatter and noise, or if you are seeing real delta v. That's the hardest thing to determine in our speed mod testing, eh!
I queried a while back on how committed you are to the long fences, because I have the forms for Gary Reed's original outlet mod design (this one):
The feeling I got from the last time we corresponded on it was that the large fences are a bit too permanent to swap out for something like Gary's design. However, if you can make the swap more easily than I understood it before, with Gary's permission, I can send you his templates to work with and send back. (I won't have the time to work exit shape this season (well, if I want to fly this year!). I also want to test the new Ram Air set-up first, which (as you know) I've been working to blend into my current cowl exit shape. Putting glass on the form tomorrow...here's the current look:
Between the ram-to-cowl blend and the spinner gap fill I've got a pretty complex layup ahead of me (old spinner was 12" and had a 1.5" wide backplate...new Van's spinner is 13" and has the narrow backplate...result, BIG gap to fill).
To keep this on topic, after I test and fly this configuration for this season, I want to go back and mod the exit. As you can see, my tunnel diverges laterally...my guess is that is not a good thing. Looks cool, but probably not conducive to acceleration and alignment of the exit air. I'd like to change the shape to make the tunnel sides parallel to the longitudinal axis, and decrease the cavernous volume it currently has. With the new ram air having a round shape, it's started me thinking about a 6 into 1 exhaust, and keeping that round shape moving aft to the exit, much like Dan's smallest exit. Different to a degree because I have no exit ramp like the 8, but along those lines. If I stay with 6 into 2, I'd have to figure out how to decrease the cross section and deal with the two pipes. Channeling flow aft of the firewall would be step 2 for me (my thinking at this point anyhoo.)
That last part is the other suggestion I thought I'd make to you, though its a departure from your current path. Ken and Dan have had success by reducing the volume of the exit tunnel (Dan with a variable insert, Ken with a smaller cross-section exit and a cowl flap). But both those guys seem to have pinched the flow forward of the firewall (at least to a degree).
Have you given any thought to reducing the size of the cowl tunnel? I know the NG leg slot makes it harder, but that might be where some more acceleration potential of exit air is hiding.
Just some thoughts...really admire your rigor in all of this, and the multiple tests will be interesting to watch for data. I feel like I should throw some Avgas $$ your way, since you are testing a bunch of stuff for all of us!
Good luck on the testing ahead!
Cheers,
Bob
Been watching this each night as I toil away at the other end of my cowling. Haven't posted in this thread in a bit, but lots of interesting things going on.
I'm glad you are re-testing various configurations and doing multiple comparative tests. With the numbers so close, its hard to tell whether results are in the scatter and noise, or if you are seeing real delta v. That's the hardest thing to determine in our speed mod testing, eh!
I queried a while back on how committed you are to the long fences, because I have the forms for Gary Reed's original outlet mod design (this one):
The feeling I got from the last time we corresponded on it was that the large fences are a bit too permanent to swap out for something like Gary's design. However, if you can make the swap more easily than I understood it before, with Gary's permission, I can send you his templates to work with and send back. (I won't have the time to work exit shape this season (well, if I want to fly this year!). I also want to test the new Ram Air set-up first, which (as you know) I've been working to blend into my current cowl exit shape. Putting glass on the form tomorrow...here's the current look:
Between the ram-to-cowl blend and the spinner gap fill I've got a pretty complex layup ahead of me (old spinner was 12" and had a 1.5" wide backplate...new Van's spinner is 13" and has the narrow backplate...result, BIG gap to fill).
To keep this on topic, after I test and fly this configuration for this season, I want to go back and mod the exit. As you can see, my tunnel diverges laterally...my guess is that is not a good thing. Looks cool, but probably not conducive to acceleration and alignment of the exit air. I'd like to change the shape to make the tunnel sides parallel to the longitudinal axis, and decrease the cavernous volume it currently has. With the new ram air having a round shape, it's started me thinking about a 6 into 1 exhaust, and keeping that round shape moving aft to the exit, much like Dan's smallest exit. Different to a degree because I have no exit ramp like the 8, but along those lines. If I stay with 6 into 2, I'd have to figure out how to decrease the cross section and deal with the two pipes. Channeling flow aft of the firewall would be step 2 for me (my thinking at this point anyhoo.)
That last part is the other suggestion I thought I'd make to you, though its a departure from your current path. Ken and Dan have had success by reducing the volume of the exit tunnel (Dan with a variable insert, Ken with a smaller cross-section exit and a cowl flap). But both those guys seem to have pinched the flow forward of the firewall (at least to a degree).
Have you given any thought to reducing the size of the cowl tunnel? I know the NG leg slot makes it harder, but that might be where some more acceleration potential of exit air is hiding.
Just some thoughts...really admire your rigor in all of this, and the multiple tests will be interesting to watch for data. I feel like I should throw some Avgas $$ your way, since you are testing a bunch of stuff for all of us!
Good luck on the testing ahead!
Cheers,
Bob
Welcome back Bob Mills
Yes I have thought of pinching the flow forward of the firewall and made several mods and tested only to find they slowed the plane down. the only thing that worked was the curved surface extending from the bottom rear of the engine to the bottom of the fuselage, side wedge planes that reduce the width of the outlet flow from full cowl to the ~14" wide outlet opening, and the horizontal plates closing off the the lower cowl outboard of the baffle and valve covers just below the cowl split line. This combination yielded the largest single speed gain of any mod I have made to the airplane - 4kts. Nothing else including something that looked like a rectangular venturi was anything but negative (many aluminum and rubber assemblies offered to the god of speed to no avail). I am completely against any further work inside the cowl. I think I have very close to the best possible combination in there.
I do not feel the same way about the exit itself and I am a long way from exhausting the possibilities. The problem is I cannot conceive an idea with supporting rational and immediately test it. Each variation has to be fabricated, materials have to be purchased and shipped, implementation details like mounting have to be worked out, the work has to be done, the test flown using exactly the same test method as previous tests, the raw data post processed, the results studied, compared with other test results and recorded for the evolving effort. Judgement, aversion to wasted effort and laziness do effect my progress - imagination is not a problem. If I merely copy mindlessly there would be little motivation for me to do the work - think of a prospector and a ditch digger. This little paragraph describes some realities I have to deal with in my mental, physical, time and life process.
Reducing the exit cross section in the outlet flow path - This has been accomplished by the two 2" bumps between the flow fences on the bottom of the fuselage and the cover. The evidence is provided by the increased cylinder head temperatures. Is the 2" height too much? I suspect it is. This was an arbitrary size triggered as a starting point by Mark Frederick's comment about cutting the size in half earlier in this thread. The installation of these bumps is quite complex and that is a hinderance to finding the right size. Since this may be turning into a long term effort I may have to stop and come up with a different installation method. The RTV is also a significant problem. It has been pointed out that it is not structural but you wouldn't think so if you had to clean it off. I have some ideas.
Turning the outlet flow with the shape of the bump - This may be the most important part of the whole experiment. Most of the testing has been with the one 2" high long tail shape and the most successful configuration is without a cover. I suspect that the long tail eliminates the turning force too early in the flow and the tested and raced once equal tailed bump may be better. I need to test it more.
The sun is coming up and I need to get work done today and I need more sleep so I will stop this post.
Bob Axsom
Yes I have thought of pinching the flow forward of the firewall and made several mods and tested only to find they slowed the plane down. the only thing that worked was the curved surface extending from the bottom rear of the engine to the bottom of the fuselage, side wedge planes that reduce the width of the outlet flow from full cowl to the ~14" wide outlet opening, and the horizontal plates closing off the the lower cowl outboard of the baffle and valve covers just below the cowl split line. This combination yielded the largest single speed gain of any mod I have made to the airplane - 4kts. Nothing else including something that looked like a rectangular venturi was anything but negative (many aluminum and rubber assemblies offered to the god of speed to no avail). I am completely against any further work inside the cowl. I think I have very close to the best possible combination in there.
I do not feel the same way about the exit itself and I am a long way from exhausting the possibilities. The problem is I cannot conceive an idea with supporting rational and immediately test it. Each variation has to be fabricated, materials have to be purchased and shipped, implementation details like mounting have to be worked out, the work has to be done, the test flown using exactly the same test method as previous tests, the raw data post processed, the results studied, compared with other test results and recorded for the evolving effort. Judgement, aversion to wasted effort and laziness do effect my progress - imagination is not a problem. If I merely copy mindlessly there would be little motivation for me to do the work - think of a prospector and a ditch digger. This little paragraph describes some realities I have to deal with in my mental, physical, time and life process.
Reducing the exit cross section in the outlet flow path - This has been accomplished by the two 2" bumps between the flow fences on the bottom of the fuselage and the cover. The evidence is provided by the increased cylinder head temperatures. Is the 2" height too much? I suspect it is. This was an arbitrary size triggered as a starting point by Mark Frederick's comment about cutting the size in half earlier in this thread. The installation of these bumps is quite complex and that is a hinderance to finding the right size. Since this may be turning into a long term effort I may have to stop and come up with a different installation method. The RTV is also a significant problem. It has been pointed out that it is not structural but you wouldn't think so if you had to clean it off. I have some ideas.
Turning the outlet flow with the shape of the bump - This may be the most important part of the whole experiment. Most of the testing has been with the one 2" high long tail shape and the most successful configuration is without a cover. I suspect that the long tail eliminates the turning force too early in the flow and the tested and raced once equal tailed bump may be better. I need to test it more.
The sun is coming up and I need to get work done today and I need more sleep so I will stop this post.
Bob Axsom
rvmills
Well Known Member
Hi back atcha Bob! All that you said with respect to experimenting and testing makes sense. I've considered how to work with my current exit shape to see if I can find improvements, and still may do some of the followng:
1. Because my exit bottom is large and raked forward (the aft edge of the cowl tunnel is not vertical...the lower edge is forward of the firewall, giving me about 85 sq in of exit...cavernous), Tom Martin told me to first extend the skin (with sheet AL) back to the pipes and see what the result was. Easy first step.
2. Then I could start filling in the cavern from the bottom up to see if acceleraton of the air and more speed results (foam or whatever...gotta think on that, as I don't want to create a fire hazard).
3. Then I could add fences and/or a Vetterman-like after-body to check the results.
All that sounds good, but my lower cowl tunnel is so low, that the more I think about it, the more I think I need to just bite the bullet and reduce the area by cutting the aft cowl and re-shaping it. Dan's switch-able set-up is attractive, because it allows one to go bigger if you find cooling inadequate.
Recently a smart aero guy whispered in my ear that reducing the stock -6 exit to about 1/2 the normal area, and adding a cowl flap for climb may be as effective as all the fence/afterbody work.
Not sure how one would accomplish this is a -6A (with the nose strut and slot), but what I was getting at above and am referring to here is that you might find more benefit from taking the plunge and reducing the outer cowl tunnel dimensions, instead of just filling the inside. Inside shape will remain important, but reducing the outer shape will give you convergence and reduced frontal area all at once. Its a large commitment though...and one I'm wrestling with hard right now!
Just more thoughts...looking forward to your continued work Bob! Having fun following along, and scheming with you!
Cheers,
Bob
1. Because my exit bottom is large and raked forward (the aft edge of the cowl tunnel is not vertical...the lower edge is forward of the firewall, giving me about 85 sq in of exit...cavernous), Tom Martin told me to first extend the skin (with sheet AL) back to the pipes and see what the result was. Easy first step.
2. Then I could start filling in the cavern from the bottom up to see if acceleraton of the air and more speed results (foam or whatever...gotta think on that, as I don't want to create a fire hazard).
3. Then I could add fences and/or a Vetterman-like after-body to check the results.
All that sounds good, but my lower cowl tunnel is so low, that the more I think about it, the more I think I need to just bite the bullet and reduce the area by cutting the aft cowl and re-shaping it. Dan's switch-able set-up is attractive, because it allows one to go bigger if you find cooling inadequate.
Recently a smart aero guy whispered in my ear that reducing the stock -6 exit to about 1/2 the normal area, and adding a cowl flap for climb may be as effective as all the fence/afterbody work.
Not sure how one would accomplish this is a -6A (with the nose strut and slot), but what I was getting at above and am referring to here is that you might find more benefit from taking the plunge and reducing the outer cowl tunnel dimensions, instead of just filling the inside. Inside shape will remain important, but reducing the outer shape will give you convergence and reduced frontal area all at once. Its a large commitment though...and one I'm wrestling with hard right now!
Just more thoughts...looking forward to your continued work Bob! Having fun following along, and scheming with you!
Cheers,
Bob
F1Boss
Well Known Member
Idea:
If the nose strut is in the way for proper outlet placement, or function, why not move the outlet?
Statement overheard at Reno:
"I've never seen an idea so good that I couldn't copy it!"
Look where Lancair put the outlets on the go-fast Legacy. There is science behind that placement (on the bottom, near the outer corner) - might as well give it a shot! Might take some exh system re-configuring? Won't work on the skinny tandem airframes...
Using Dan's idea of a replaceable outlet section, it would be easy to experiment. Are ya listenin', Super Bob?
Carry on!
Mark
If the nose strut is in the way for proper outlet placement, or function, why not move the outlet?
Statement overheard at Reno:
"I've never seen an idea so good that I couldn't copy it!"
Look where Lancair put the outlets on the go-fast Legacy. There is science behind that placement (on the bottom, near the outer corner) - might as well give it a shot! Might take some exh system re-configuring? Won't work on the skinny tandem airframes...
Using Dan's idea of a replaceable outlet section, it would be easy to experiment. Are ya listenin', Super Bob?
Carry on!
Mark
A quick response
Bob, I see now how your non-standard exit changes from the standard.
By all means, forward my prototype to where ever it might do the most good (or harm!?!?). I remain convinced that just streamlining the exit air has significant benefit, both for cooling efficiency and drag reduction.
Right now I am kinda freezing my design. I have some ideas about where to go, but I am wondering about the benefit at this point. It seems that I am stuck on a plateau.
Multiple 3 leg NTPS speed runs without configuration changes on different days with the same method yield results +/- 1.5 knots.
I recently closed my elevator ends with no apparent speed increase (maybe a decrease). I kind of wonder if 180 - 181 KTAS at 6000 DA corrected for temps is as fast as my stock IO 360 can make my particular -6 go?????
Waiting for my first SARL opportunity (Indy) to see what happens.
Bob, I see now how your non-standard exit changes from the standard.
By all means, forward my prototype to where ever it might do the most good (or harm!?!?). I remain convinced that just streamlining the exit air has significant benefit, both for cooling efficiency and drag reduction.
Right now I am kinda freezing my design. I have some ideas about where to go, but I am wondering about the benefit at this point. It seems that I am stuck on a plateau.
Multiple 3 leg NTPS speed runs without configuration changes on different days with the same method yield results +/- 1.5 knots.
I recently closed my elevator ends with no apparent speed increase (maybe a decrease). I kind of wonder if 180 - 181 KTAS at 6000 DA corrected for temps is as fast as my stock IO 360 can make my particular -6 go?????
Waiting for my first SARL opportunity (Indy) to see what happens.
Last edited:
NDrv8r
Well Known Member
I previously posted a video on tuft testing using a cowl flap. The testing revealed poor airflow across the flap when closed.
I just returned from a 10 hour trip, and having the flap closed makes a definite 4kt gain over open. the cyl temps are in the 325 range, the ot 195, ff 8.5gph, TAS 169, somtimes 170kt at 10000.
Makes me wonder how it will work with a smoothing after body ala Veterman in conjunction with the movable flap.
I want to focus on the magic 7 degree number blending back into the bottom fuselage. The idea is close off excess exit area, smooth the flow over the cowl flap between the stacks, increase the flow around the exhaust pipes to better match exterior flow.
On a related note, Dan, Kevin and others that have done cowl pressure tests. what kind of pressures have you noted in the upper plenum? At 180Kt I am seeing 14.7in H2O. At 180 kt the pitot pressure should be around 21.5 in h2o. I have the low side of the sensor tied to my aircraft static port and the high side to a static port (1/8 tube with .040 holes drilled on 4 sides) at the fuel distribution block. Should there be this great of a loss?
If I understand, if there were perfect pressure recovery I should get the same pressure as a pitot tube???
I also have a sensor on the engine mount tube. The upper plenum pressure stays constant and the lower plenum pressure increases from 2 to 4 when closing the cowl flap.
I just returned from a 10 hour trip, and having the flap closed makes a definite 4kt gain over open. the cyl temps are in the 325 range, the ot 195, ff 8.5gph, TAS 169, somtimes 170kt at 10000.
Makes me wonder how it will work with a smoothing after body ala Veterman in conjunction with the movable flap.
I want to focus on the magic 7 degree number blending back into the bottom fuselage. The idea is close off excess exit area, smooth the flow over the cowl flap between the stacks, increase the flow around the exhaust pipes to better match exterior flow.
On a related note, Dan, Kevin and others that have done cowl pressure tests. what kind of pressures have you noted in the upper plenum? At 180Kt I am seeing 14.7in H2O. At 180 kt the pitot pressure should be around 21.5 in h2o. I have the low side of the sensor tied to my aircraft static port and the high side to a static port (1/8 tube with .040 holes drilled on 4 sides) at the fuel distribution block. Should there be this great of a loss?
If I understand, if there were perfect pressure recovery I should get the same pressure as a pitot tube???
I also have a sensor on the engine mount tube. The upper plenum pressure stays constant and the lower plenum pressure increases from 2 to 4 when closing the cowl flap.
hydroguy2
Well Known Member
Kind of where I am. I don't have the time or gas money to run the meticulous testing that Racer Bob has been doing for us all. I would really like to make some significant modifications to my lower cowl and even try a new 4 to1 exhaust....but this ain't happening any time soon. I have built a after fairing for the cowl exit, but afraid to drill the fuselage to mount it. I've seen 192-193kts TAS at 6600' DA. but usually it's nearer to 189. I know(in comparison to others) I'm slippery for a -7, since I cruise at 175kts burning 8.5 or so. 9gph is 178-180kt TAS.
At this point, my plane is faster than me!
I should be able to run down that #71, but still trying to figure out the methodology. Get me faster.......then we'll work on fine tuning Aurora is my next step.
Ready for test #17
This was a dogged task to make these fins to duplicate the profile of the heavy ones without the cover mount hardware but they are finally done and ready to test. These are made of 0.032 2024-0 in place of the 2024T3 0.090 originals. I was ready to fly this evening but the ceiling wouldn't allow me to go to test altitude. Tomorrow is supposed to be sunny.
Once this test is done (basically to see how much the cover mounting hardware affected speed) I think I will try a couple of changes on the fin profiles.
Bob Axsom
This was a dogged task to make these fins to duplicate the profile of the heavy ones without the cover mount hardware but they are finally done and ready to test. These are made of 0.032 2024-0 in place of the 2024T3 0.090 originals. I was ready to fly this evening but the ceiling wouldn't allow me to go to test altitude. Tomorrow is supposed to be sunny.
Once this test is done (basically to see how much the cover mounting hardware affected speed) I think I will try a couple of changes on the fin profiles.
Bob Axsom
vlittle
Well Known Member
Bingo. Reducing the exit area and the frontal area of the cowl is the secret to success, however....
As pressure in the lower cowl increases due to the reduced outlet area, air will tend to leak in other areas. Around the prop spinner is one, and any gaps in the hinges/platenuts/camlocks as well. Tightly sealing the cowling and making all of the cooling air exit the bottom will be the challenge. A good way to test this is to tape the cowling joints all around and measure cht and airspeed.
Around the spinner, Tom Martin's method seems to work... but his results actually reduced airspeed due to increased mass airflow... so the reduction in outlet area is necessary to regain speed.
Bottom line... reduce cowl exit area, seal the cowl air leaks and reduce cowl frontal area. 100% money-back guarantee to increase speed.
Dynamic pressure of 21.5" H2O at 180 knots on a standard day means you would be flying just slightly below sea level. You're probably not piloting a 180 knot submarine
. What is your altitude when you're measuring 14.7"?That will work.
I think you'll find your max possible Q to be quite a lot less when you factor in altitude (multiply by density ratio)...so your 'loss" will be less than you think now. That said, it's not really a loss. Upper plenum pressure will never equal max Q. Remember, the airspeed pitot is a closed end system, while the cooling flow is open ended. The cooling system does not bring the air to a full stop, thus it does not convert all the available kinetic energy to increased static pressure.
reduced exit
I beleive you are on the right track reducing the air exit size. That is for sure if you have the cowling set up for fuel injection with a stock vans intake scoop. On my IO320 I used the intake scoop for a carburated engine which has nearly a quarter or one third less exit area. I was able to make this work with my Bexdix injection system by thinning my fiberglass air filter housing by an inch or so. All fit fine. Then added my gen two Vetterman fairing and life was good and faster also. Just remember there is a fine line between the sweet spot (size/area) and temperture control. We all try to fine tune our planes but in the end they are all just a mixture of compromises.
I beleive you are on the right track reducing the air exit size. That is for sure if you have the cowling set up for fuel injection with a stock vans intake scoop. On my IO320 I used the intake scoop for a carburated engine which has nearly a quarter or one third less exit area. I was able to make this work with my Bexdix injection system by thinning my fiberglass air filter housing by an inch or so. All fit fine. Then added my gen two Vetterman fairing and life was good and faster also. Just remember there is a fine line between the sweet spot (size/area) and temperture control. We all try to fine tune our planes but in the end they are all just a mixture of compromises.
rvmills
Well Known Member
OK, here goes the mutha of all multi-quotes!
A. Hey man, stop listening to my conversations at Reno!
B. I'll buy the beer and provide a shady spot under my cowling for you to show me your idea (so I can steal in!)...right after we crawl under Lee's Legacy at PRS...deal?
C. Kinda leanin' towards 6 into 1 and a single opening...two openings seems counterproductive...but if they are small, and the air is well channeled...hmmm...and I'm always listenin'!
Gary, I've played with the forms, and I really need to mod the cowl to make it work. If/when I go there, I may not get back to stock. Still interested to see if you have recorded speed gains from the firewall radius alone and the exit mod alone? Are those experiments reversible and testable one at a time and back to ground zero?
Larry, can you repost the link to the video? Any pics of the cowl flap set-up? Thanks!
1Og, any pics of the V gen 2 afterbody? Apologies if its already in this thread...
All good info Vern! I did not know Tom lost speed when he closed down the cowl on the prop hub. I knew he no longer had to open his cowl flap for climb, so it makes sense. Glad you mentioned it here, as that would have been disappointing to seal there and slow down...until the light bulb came on. Will plan those mods together, in sequence. I discovered that with my new Van's spinner and my extended hub, I can't use Tom's method, as the forward edge of the cowl won't close on the round part of the hub. Will be using Dan's method instead.
Hey Dan, have you measured the effect (temps and speed) of the baffle material closures you use at the hub? I know you built it that way, but given Vern's comments, I wonder if there is an optimum blend of sealed prop hub and exit area...perhaps its just that the seals allow you to go smaller and smaller, and gain speed via mass airflow reduction, until you reach your EGT CHT comfort limits. Thoughts?
Do both Brian! Keep working on the airplane, then remember, "slow is smooth, and smooth is fast". Lead the turns, don't over-g and bleed, and get on and stay on the line. Oh, and always guess right on the winds!
Just trying to fly my baby again before your deployment is up...at this rate...
An HRII, a J-3 and an RV-7 have kept me going, but things are going together versus coming apart now...I sat in the plane for the first time in 5 months to replace the 396 and made airplane noises again...getting closer!
Bob, what's this test looking at? Thinner guage fences? Hardware changes...anything else? With long .032 fences in there, I'd be thinking about vibes too. Wanna borrow my Countour camera and take some video down there? I can send it with Gary's forms if you like.
I asked Gary about the separate gains from his firewall radius and his afterbody fairing. I think the sum was 3-4 knots (Gary?). The radius is reported to be a very positive thing, so the afterbody may be incremental, versus a silver bullet. And going long on the fences may be hurting the overall effect. We should research the effect of underbody strakes and fences on speed. They usually add stability, not speed. Not sure at what length or what station your flow fences become draggy. Alignment becomes more and more critical as length increases. If they are adding a yaw input anywhere, it will be draggy.
Are you going to stop by Reno and PRS on your way to the Euphrata race? Would like to see all this in person...and you can join Mark and I under the Legacy!
Cheers,
Bob
A. Hey man, stop listening to my conversations at Reno!
B. I'll buy the beer and provide a shady spot under my cowling for you to show me your idea (so I can steal in!)...right after we crawl under Lee's Legacy at PRS...deal?
C. Kinda leanin' towards 6 into 1 and a single opening...two openings seems counterproductive...but if they are small, and the air is well channeled...hmmm...and I'm always listenin'!
Gary, I've played with the forms, and I really need to mod the cowl to make it work. If/when I go there, I may not get back to stock. Still interested to see if you have recorded speed gains from the firewall radius alone and the exit mod alone? Are those experiments reversible and testable one at a time and back to ground zero?
Larry, can you repost the link to the video? Any pics of the cowl flap set-up? Thanks!
1Og, any pics of the V gen 2 afterbody? Apologies if its already in this thread...
All good info Vern! I did not know Tom lost speed when he closed down the cowl on the prop hub. I knew he no longer had to open his cowl flap for climb, so it makes sense. Glad you mentioned it here, as that would have been disappointing to seal there and slow down...until the light bulb came on. Will plan those mods together, in sequence. I discovered that with my new Van's spinner and my extended hub, I can't use Tom's method, as the forward edge of the cowl won't close on the round part of the hub. Will be using Dan's method instead.
Hey Dan, have you measured the effect (temps and speed) of the baffle material closures you use at the hub? I know you built it that way, but given Vern's comments, I wonder if there is an optimum blend of sealed prop hub and exit area...perhaps its just that the seals allow you to go smaller and smaller, and gain speed via mass airflow reduction, until you reach your EGT CHT comfort limits. Thoughts?
Do both Brian! Keep working on the airplane, then remember, "slow is smooth, and smooth is fast". Lead the turns, don't over-g and bleed, and get on and stay on the line. Oh, and always guess right on the winds!
Ahhh! Bob, what have you done!?!
Just trying to fly my baby again before your deployment is up...at this rate...
An HRII, a J-3 and an RV-7 have kept me going, but things are going together versus coming apart now...I sat in the plane for the first time in 5 months to replace the 396 and made airplane noises again...getting closer!
Bob, what's this test looking at? Thinner guage fences? Hardware changes...anything else? With long .032 fences in there, I'd be thinking about vibes too. Wanna borrow my Countour camera and take some video down there? I can send it with Gary's forms if you like.
I asked Gary about the separate gains from his firewall radius and his afterbody fairing. I think the sum was 3-4 knots (Gary?). The radius is reported to be a very positive thing, so the afterbody may be incremental, versus a silver bullet. And going long on the fences may be hurting the overall effect. We should research the effect of underbody strakes and fences on speed. They usually add stability, not speed. Not sure at what length or what station your flow fences become draggy. Alignment becomes more and more critical as length increases. If they are adding a yaw input anywhere, it will be draggy.
Are you going to stop by Reno and PRS on your way to the Euphrata race? Would like to see all this in person...and you can join Mark and I under the Legacy!
Cheers,
Bob
The test in a few hours (sleep is a problem)
The test in a few hours is using 0.032 of place of the original 0.090 partly because that is what I had on hand. There is a little more to it than thinner material. That configuration was flown in test case number 10 and provided the fastest speed in this series of tests 182.1 kts. Then I repeated it in test case 15 and the speed was only 179.3 kts. I have always said that I thought the error margin of the test method I use is +2 so the the difference may be just that. Both flights had the 2" high bump with a long tail (you can see it in the photo) and the three vertical members (flow fences, fins, separator ...). The vertical members had a lot of stuff at the lower end that is used for mounting the cover (cover is not installed in this configuration). A ~10 long piece of 3/4"x3/4"x1/16" aluminum angle and 19 #8 floating platenuts. This test case has the same fin profile on all three parts but none of the cover mounting hardware. It should be faster but the difference may be lost in the test method error margin. It was agonizing to build the parts for such a seemingly trivial test but it is done now and it will be good to see the results in the table. Another previous test to keep your eyes on with respect to the upcoming test is test case 13 which had the same bumps, no cover and 0.032 fins of the straight triangular configuration. Its speed was 179.1. By the way these triangular fins were also used with no cover and no bump in test case 6 and yielded a speed of 181.5 kts.
I have to use my judgement in evaluating these results and call things meaningful as I see them and I may get it wrong but even then it keeps me from getting into a mental fog with no direction. "Do something even if it's wrong" applies here.
I have several ideas about what is wrong with my test configurations so far and I just received a shipment of material from Aircraft Spruce today to use in addressing the problems.
A difference to observe in test case 13 fin and bump height relationship and those same relationships in test cases 10, 15 and 17 (after sun rise) is, at the bump peak in 13 the small triangular fins extend just slightly beyond the bump and may not isolate the cooling air outlet flow allowing interaction with the surrounding air at the worst possible time. If you try to visualize it I think you will see what I am talking about. It is possible that test 17 will yield a new high in this series because the bump will be working the Coanda magic and the cover mounting hardware will be gone.
The bump shape. I have only tried two and one had a catastrophic failure observed at the fuel pump after completing the very successful race at Llano. I thought some advice in this thread was good and I took the bump profile in that direction - the long drawn out shape after the node. From a streamlining perspective it seemed like a good idea but the Coanda effect depends on the curved surface continuing to curve in the direction you want the air to go. When I flatten the surface out the air flow loses interest and goes its own way. I think I need to work back toward my original shape but the 0.016" 2024 T3 seemed too weak for the task. This may be misleading though because I observed evidence of high frequency trailing edge oscillations with excursions of 2 to 3 inches on the side of the center fin. It could have shattered from fatigue. The long bump is made from 0.032" 2024 0 aluminum and the trailing edge is secured with screws so it is much stronger, less responsive and mechanically secured against oscillation. These features could be applied to the old shape or at least something in that direction.
I am seriously considering stopping at the PRS event on the way to Washington. Is the Legacy you are talking about race #5?
Well I'm going back to bed for a little while.
Bob Axsom
The test in a few hours is using 0.032 of place of the original 0.090 partly because that is what I had on hand. There is a little more to it than thinner material. That configuration was flown in test case number 10 and provided the fastest speed in this series of tests 182.1 kts. Then I repeated it in test case 15 and the speed was only 179.3 kts. I have always said that I thought the error margin of the test method I use is +2 so the the difference may be just that. Both flights had the 2" high bump with a long tail (you can see it in the photo) and the three vertical members (flow fences, fins, separator ...). The vertical members had a lot of stuff at the lower end that is used for mounting the cover (cover is not installed in this configuration). A ~10 long piece of 3/4"x3/4"x1/16" aluminum angle and 19 #8 floating platenuts. This test case has the same fin profile on all three parts but none of the cover mounting hardware. It should be faster but the difference may be lost in the test method error margin. It was agonizing to build the parts for such a seemingly trivial test but it is done now and it will be good to see the results in the table. Another previous test to keep your eyes on with respect to the upcoming test is test case 13 which had the same bumps, no cover and 0.032 fins of the straight triangular configuration. Its speed was 179.1. By the way these triangular fins were also used with no cover and no bump in test case 6 and yielded a speed of 181.5 kts.
I have to use my judgement in evaluating these results and call things meaningful as I see them and I may get it wrong but even then it keeps me from getting into a mental fog with no direction. "Do something even if it's wrong" applies here.
I have several ideas about what is wrong with my test configurations so far and I just received a shipment of material from Aircraft Spruce today to use in addressing the problems.
A difference to observe in test case 13 fin and bump height relationship and those same relationships in test cases 10, 15 and 17 (after sun rise) is, at the bump peak in 13 the small triangular fins extend just slightly beyond the bump and may not isolate the cooling air outlet flow allowing interaction with the surrounding air at the worst possible time. If you try to visualize it I think you will see what I am talking about. It is possible that test 17 will yield a new high in this series because the bump will be working the Coanda magic and the cover mounting hardware will be gone.
The bump shape. I have only tried two and one had a catastrophic failure observed at the fuel pump after completing the very successful race at Llano. I thought some advice in this thread was good and I took the bump profile in that direction - the long drawn out shape after the node. From a streamlining perspective it seemed like a good idea but the Coanda effect depends on the curved surface continuing to curve in the direction you want the air to go. When I flatten the surface out the air flow loses interest and goes its own way. I think I need to work back toward my original shape but the 0.016" 2024 T3 seemed too weak for the task. This may be misleading though because I observed evidence of high frequency trailing edge oscillations with excursions of 2 to 3 inches on the side of the center fin. It could have shattered from fatigue. The long bump is made from 0.032" 2024 0 aluminum and the trailing edge is secured with screws so it is much stronger, less responsive and mechanically secured against oscillation. These features could be applied to the old shape or at least something in that direction.
I am seriously considering stopping at the PRS event on the way to Washington. Is the Legacy you are talking about race #5?
Well I'm going back to bed for a little while.
Bob Axsom
Last edited:
rvmills
Well Known Member
Good luck in the tests Bob! I think Lee's Legacy is Race 5, but there are several Legacy entrants...all fast! Talked to Rick Vandam (Sport Class Sec) yesterday, and there are 10 Sport Class Rookies coming to PRS (a few of them are RVs...couple 8's and a 3), and a total of 18 Sport Class Racers will be at PRS...more than any other class! Hope to see you there...come hang out with us!
Cheers,
Bob
Cheers,
Bob
Tests #18 and 19 Same configuration same flight
Test #18 achieved 177.8 kts - pretty bad and I had time left so I flew the triangle again and got higher speeds on every leg and ended up with 180.2 Using Kent's 180 good/bad split logic this one is still barely in the game. I will update the tables and think about what to do next.
Bob Axsom
Test #18 achieved 177.8 kts - pretty bad and I had time left so I flew the triangle again and got higher speeds on every leg and ended up with 180.2 Using Kent's 180 good/bad split logic this one is still barely in the game. I will update the tables and think about what to do next.
Bob Axsom
Thats the nature of experimental development
I'll drop out for a while and get back to you later when I have something on the other side. I ordered some more aluminum tonight/this morning and I am going to construct a removable base that I can use for more rapid change of experiments and avoid the use of RTV.
Bob Axsom
I'll drop out for a while and get back to you later when I have something on the other side. I ordered some more aluminum tonight/this morning and I am going to construct a removable base that I can use for more rapid change of experiments and avoid the use of RTV.
Bob Axsom
Last edited:
Kevin Horton
Well Known Member
Test #18 achieved 177.8 kts - pretty bad and I had time left so I flew the triangle again and got higher speeds on every leg and ended up with 180.2 Using Kent's 180 good/bad split logic this one is still barely in the game. I will update the tables and think about what to do next.
Bob Axsom
Bob - The fact that you can fly the same configuration twice and get speeds that are 2-3 kt different suggests your current test technique produces results that are only accurate to 3 kt or more. The changes your are making to the aircraft probably have an effect on performance that is smaller than the accuracy of your test. It is useless to try to compare results from single data points given the demonstrated test to test variability. If you wish to determine whether any given mod helps or hurts you'll need to get a test protocol that yeilds higher accuracy.
Given the level of accuracy you need to measure, you need to fly several tests on each configuration and average the results. Given that there can be small effects from rising and descending air mass, you should do these tests on several different days.
Do you have the ability to record engine and flight data for later processing? What test protocol do you use? It might be possible to tweak the test protocol somewhat to reduce the test to test variablity.
Thanks for the help Kevin
All good observations.
Aircraft Spruce got a little richer last night as I try to improve the test bed. I bought enough 2024 T3 0.063" and 0.090" alclad for four base plates that span the the outboard flow fence mounts I have riveted to the bottom of the fuselage with Cherry rivets. One thing that has been a hinderance is not wanting to deteriorate the fuselage in the process of attaching of test articles. This change will free me from that problem/worry. I will have a easily removable and replaceable false fuselage bottom a little less than 1" below the fuselage directly behind the outlet.
Bob Axsom
All good observations.
Aircraft Spruce got a little richer last night as I try to improve the test bed. I bought enough 2024 T3 0.063" and 0.090" alclad for four base plates that span the the outboard flow fence mounts I have riveted to the bottom of the fuselage with Cherry rivets. One thing that has been a hinderance is not wanting to deteriorate the fuselage in the process of attaching of test articles. This change will free me from that problem/worry. I will have a easily removable and replaceable false fuselage bottom a little less than 1" below the fuselage directly behind the outlet.
Bob Axsom
Last edited:
kentb
Well Known Member
A question.
Bob, I notice that you don't have hinge pin in the middle of your exit air like I do on mine. Do you not use hinge to attach the lower cowl? If you do, how do you secure the ends of the pins.
I have been think on how to move mine so they are not in the exit air. I am sure that they don't effect the air much, but they will be in the way when I start doing my experimenting.
Kent
Bob, I notice that you don't have hinge pin in the middle of your exit air like I do on mine. Do you not use hinge to attach the lower cowl? If you do, how do you secure the ends of the pins.
I have been think on how to move mine so they are not in the exit air. I am sure that they don't effect the air much, but they will be in the way when I start doing my experimenting.
Kent
I had them once but rivet failures forced change
There was so much stress on the lower cowl to fuselage interface that I had to remove the hinges from that location - all the others are OK. Rivet heads were popping off on the fuselage side of the hinges. At first I went up in rivet diameter then I replaced the hinges with solid plates. They have platenuts for the cowl side. When I had hinges I had a small loop in the end of the hinge pins and after fully inserting them I safety wired them to the engine mount. I made up a little hinge pin puller tool out of the soft hinge pin material (that I had replaced with hard hinge pin material) to reach in and grab the little loop (after cutting the safety wire from the inside) and pull them out. I still use that puller for the other hinge pins - it's just a finger loop and a hook.
Bob Axsom
There was so much stress on the lower cowl to fuselage interface that I had to remove the hinges from that location - all the others are OK. Rivet heads were popping off on the fuselage side of the hinges. At first I went up in rivet diameter then I replaced the hinges with solid plates. They have platenuts for the cowl side. When I had hinges I had a small loop in the end of the hinge pins and after fully inserting them I safety wired them to the engine mount. I made up a little hinge pin puller tool out of the soft hinge pin material (that I had replaced with hard hinge pin material) to reach in and grab the little loop (after cutting the safety wire from the inside) and pull them out. I still use that puller for the other hinge pins - it's just a finger loop and a hook.
Bob Axsom
Prop baffle
I added the baffle material like Dan's before the Texoma race. Prior to this mod my small oil leaks under the front of the engine migrated forward. After adding the prop seal the oil moves rearward. Makes sense. I also noticed my CHT's increased about 15 degrees which does not agree with Tom's result. Did not see an obvious speed change but I did not run the triangle course either. It appears a significant amount of air from the lower cowling was exiting between the cowl/spinner gap on my RV8.
Bob Mills - It will be interesting to see how you decide to modify your exhaust. I chose the 4-1 and extended the collector with the intention of reducing the cowl outlet. Still trying to decide if I go the bluff body route or not. Not enough time for me to try lots of options so the result is no action.
On a side note I would like to see the T-6 class at Reno go away and see a 4 cylinder RV only class added. RV's are not competitive in any Reno class (not talking about Super RV's here).
I added the baffle material like Dan's before the Texoma race. Prior to this mod my small oil leaks under the front of the engine migrated forward. After adding the prop seal the oil moves rearward. Makes sense. I also noticed my CHT's increased about 15 degrees which does not agree with Tom's result. Did not see an obvious speed change but I did not run the triangle course either. It appears a significant amount of air from the lower cowling was exiting between the cowl/spinner gap on my RV8.
Bob Mills - It will be interesting to see how you decide to modify your exhaust. I chose the 4-1 and extended the collector with the intention of reducing the cowl outlet. Still trying to decide if I go the bluff body route or not. Not enough time for me to try lots of options so the result is no action.
On a side note I would like to see the T-6 class at Reno go away and see a 4 cylinder RV only class added. RV's are not competitive in any Reno class (not talking about Super RV's here).
I pulled the previous mod
I'm a little under the weather at the moment but I pulled the last mod this evening. I have to remove a lot of Dow 736 red RTV still but then I think I will fly a test with no fins or covers or bumps before I start working with the test bed.
Bob Axsom
I'm a little under the weather at the moment but I pulled the last mod this evening. I have to remove a lot of Dow 736 red RTV still but then I think I will fly a test with no fins or covers or bumps before I start working with the test bed.
Bob Axsom
No fins, no bumps and no covers test
I flew the test twice today and the speeds were all over the place on each leg. I never came close to meeting the USAR test method requirement of five 20 second interval recordings within 1 kt. Regardless, I recorded a large number of speeds and divided by the total to get the inputs for the NTPS spread sheet then I repeated it. The input average numbers look right but the results are disappointing. 178.8 kts and 177.6 kts respectively. Test fuel cost $71.50. The RPM was 2720, the EGT on CYL #4 was 1300, the CHTs were 357, 369, 365 and 342(?) but the power didn't seem as robust as it had been in the past. For what its worth here are the updated charts:
Bob Axsom
I flew the test twice today and the speeds were all over the place on each leg. I never came close to meeting the USAR test method requirement of five 20 second interval recordings within 1 kt. Regardless, I recorded a large number of speeds and divided by the total to get the inputs for the NTPS spread sheet then I repeated it. The input average numbers look right but the results are disappointing. 178.8 kts and 177.6 kts respectively. Test fuel cost $71.50. The RPM was 2720, the EGT on CYL #4 was 1300, the CHTs were 357, 369, 365 and 342(?) but the power didn't seem as robust as it had been in the past. For what its worth here are the updated charts:
Bob Axsom
Last edited:
rvmills
Well Known Member
Jason,
On the day of the test for the prop seal baffle, was the OAT hotter, or did you do that test with a heat-soaked engine (second, third flight on a hot day)? All that could raise oil temps and CHT and give an apples to oranges comparison. Kevin is spot on on the multiple test profile. When I'm really in the test mode, I try to do them on multiple days and, if its a reversible mod, do them in opposite order (A first on one day, B first on the next...hoping for similar conditions...which can be the hard part).
As far as exhaust goes...I need to do more research, and talk to Larry. There is a 6 into 1 Rocket exhaust out there too, but I've never seen it on a Rocket (hmmm). Before I start messing with exhaust and the outlet, I want to get the current configuration flying (new baffles, plenum and ram air), get some baseline data, including instrumenting up per Dan's pressure differential set up (need to get smarter on that too...looks pretty simple, but just need to work through the set up).
Dan, if you're still reading this thread, have you measured an impact to performance from the measuring devices? (Might be a dumb question, but I'm good at those!
)Jason, as far as Reno goes...with a few more RVs coming to PRS this year (and where is Race 391, hmmmm?), if we get enough RVs there in September (any September), the Sport Bronze could almost become a de-facto RV/Rocket class. And if there are a couple mini-races within the heats (6 bangers up front, 4 bangers a bit back), its still good racing and good entertainment (for the pilot and the crowd). From what I've seen, starting a new class at Reno is very difficult...politically with RARA, scheduling-wise with the Air Boss...and logistically, you really gotta have your stuff in one sock to play there as a new class. Also, good luck getting rid of the T-6s...good group, cool airplanes, and crowd-pleasing NOISE!
All in fun!
Off to sand-fill!
Cheers,
Bob
My observation is about 15 deg CHT rise over many days at many different ambient temps. Just seems to run a little hotter which is fine with me. I was kindof on the cool side and now in cruise and 75% power the CHT spread is around 360-370. When in an aggressive climb the CHT's rise faster now than before. I know this is not good data numbers but I really don't have an interest performing back to back comparisons on the front prop seal. It seemed like a good thing to do and am leaving it on regardless. Because my CHT's are a little higher I assume mass airflow through the engine is a little lower and speed would be slightly higher but cannot confirm any of that.
Thanks for the Reno comment. Even though the T-6 guys are nice and make lots of noise at 230 mph there is very little if any interesting stuff going on in that class.
If I ever have $48,000 laying around then I will get the Superior XP-408 engine and come to Reno. It is a 230 hp parallel valve so I assume 50 more hp for little weight penalty. My stock XP-360 180 hp just isn't enough to bother.
Jason
$48,000 - that is obscene (sigh)
No wonder the engine guys are going out of business. I've looked at my data and I think for this race season I'm going to have to find the best outlet configuration I can and go with it. The last two test I ran yesterday I know are invalid (20 & 21). The wide velocity swings (>10 kts, the worst I have seen) from the up and down drafts over the Ozarks on altitude hold was unavoidable. I can go out there at dawn tomorrow and fly it before the sun heats everything up and gets the air moving around. I am leaning toward the configuration in test 7. I know how silly this sounds but it just looked right. I bought a sheet of 0.016 stainless steel to make the bumps out of instead of the 0.016 2024-T3 aluminum that had shattered by the end of the race at Llano, TX. I may make the bump lower and I may make the cutouts larger before I go the the test bed configuration. The methodical repetitive testing is all well and good but speaking from the perspective of a mature retired guy on fixed income you would like throw away some less promising options before wasting a lot of time and money on them.
Bob Axsom
No wonder the engine guys are going out of business. I've looked at my data and I think for this race season I'm going to have to find the best outlet configuration I can and go with it. The last two test I ran yesterday I know are invalid (20 & 21). The wide velocity swings (>10 kts, the worst I have seen) from the up and down drafts over the Ozarks on altitude hold was unavoidable. I can go out there at dawn tomorrow and fly it before the sun heats everything up and gets the air moving around. I am leaning toward the configuration in test 7. I know how silly this sounds but it just looked right. I bought a sheet of 0.016 stainless steel to make the bumps out of instead of the 0.016 2024-T3 aluminum that had shattered by the end of the race at Llano, TX. I may make the bump lower and I may make the cutouts larger before I go the the test bed configuration. The methodical repetitive testing is all well and good but speaking from the perspective of a mature retired guy on fixed income you would like throw away some less promising options before wasting a lot of time and money on them.
Bob Axsom
No attempt to do so.....any impact due to measurement would probably be within the data scatter.
Here's the stuff currently living in the cowl exit flow, pitot-static on the left and shielded temperature probe on the right. In theory the two devices probably slow exit velocity a little. I wouldn't expect any change due to the piccolo tubes above and below the cylinders.
Yep. Oil and cylinder head temperatures alone are poor indicators; too many other factors make them rise or fall. If you want the same CHT or oil temp with less mass flow you must improve heat transfer efficiency, i.e. make less air carry more heat. Consider using a different figure of merit for development:
change in air temperature / media delta = heat transfer efficiency
or (for example)
(exit air temperature - OAT) / (CHT - OAT)
Here is a real example, data from a flight to Jackson MS and back last summer. The outbound leg was in the early AM, and being a business trip I was in a hurry to get there. So max cruise with rich-of-peak mixture, and thus elevated CHT despite low OAT. The return leg in the afternoon was more relaxed, running LOP with lower CHT despite higher OAT.
(exit 234 - OAT 64) / (CHT 324 - OAT 64) = 0.65
(exit 232 - OAT 86) / (CHT 307 - OAT 86) = 0.66
Efficiency is similar, despite very different CHT and OAT conditions.
Let's suppose you made some clever modification which caused the air to pick up more heat......say exit temp was now 250F instead of 232F:
(exit 250 - OAT 86) / (CHT 307 - OAT 86) = .74
A few notes:
Exit air temperature in this case was measured immediately below #1 cylinder, not at the cowl exit. I wanted separate cylinder and oil cooler data. Temperature (thus efficiency) measured at the cowl exit will be lower.
Most of us have fixed cowl exit areas, so mass flow is increased as we go faster. Exit temperature will drop given the same CHT and OAT, but more air. So, you must fly the same TAS for direct comparison of cooling efficiency mods.
Last edited:
kentb
Well Known Member
Interesting....
Dan, I mounted to temp probes in my cowl exit air. One at the center and one under the exhaust pipe (about 1.5 inches from the pipe). I was surprised that the probe under the pipe was only 2 degrees higher then the one in the middle (this is an RV9A).
OAT 60
exit air 140
cht 340
(140 - 60) / (340 - 60) = 28.5
Does my plane suck or what?
By the way what is the exit air speed on your plane?
Kent
Dan, I mounted to temp probes in my cowl exit air. One at the center and one under the exhaust pipe (about 1.5 inches from the pipe). I was surprised that the probe under the pipe was only 2 degrees higher then the one in the middle (this is an RV9A).
OAT 60
exit air 140
cht 340
(140 - 60) / (340 - 60) = 28.5
Does my plane suck or what?
By the way what is the exit air speed on your plane?
Kent
OAT 60
exit air 140
cht 340
(140 - 60) / (340 - 60) = 28.5
Does my plane suck or what?
Remember, my example efficiency of .65~.66 was based on outlet air from a cylinder head only....meaning very hot surfaces compared to an oil cooler. The temperature of mixed cylinder outlet air and oil cooler outlet air will lower, so efficiency measured at the cowl outlet will be lower. Note this mixed air method does not result in a strictly accurate efficiency value; the equation input was CHT, not some average of CHT and oil cooler fin temperature. There's no handy way to do that. No big deal....we're looking for a figure of merit, something relatively easy to measure, exactly as you have done.
Back to your 0.28 value; poor, but not a lot less than values in our test database, assuming you took the temperature readings at high TAS. The test partners have agreed to not discuss actual values until we're satisfied with the work, so please be patient....all in due course.
Post sleep morning thoughts
Interestingly, you can see the pattern of the original "bump" in the RTV residue on the side of the center vane - it was 2" high. All the work last night was spent developing an effective interface with the front end of this bump and the existing lower cowl baffle I have extending from the back of the engine to the bottom of the firewall. I was able to use an existing platenut, add another one and feed the leading edge under the trailing edge of some existing sheet metal in the complex existing baffle. I didn't want to rush into the rest of the structure without some thought so I came home for a little sleep. I want to avoid RTV this time. I have an idea for the side support and anxiety is keeping me from resting so I guess it is time to get back to work.
Bob Axsom
Interestingly, you can see the pattern of the original "bump" in the RTV residue on the side of the center vane - it was 2" high. All the work last night was spent developing an effective interface with the front end of this bump and the existing lower cowl baffle I have extending from the back of the engine to the bottom of the firewall. I was able to use an existing platenut, add another one and feed the leading edge under the trailing edge of some existing sheet metal in the complex existing baffle. I didn't want to rush into the rest of the structure without some thought so I came home for a little sleep. I want to avoid RTV this time. I have an idea for the side support and anxiety is keeping me from resting so I guess it is time to get back to work.
Bob Axsom
rvmills
Well Known Member
