What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

The Shrinking Exit

Good stuff Dan!

Are the pressure sensors analog or digital output? Are you planning on recording the raw outputs and converting the data to pressure/velocities on the ground?

Ken
 
Are the pressure sensors analog or digital output? Are you planning on recording the raw outputs and converting the data to pressure/velocities on the ground?

Analog voltage proportional to pressure or temperature. Cockpit tools are just my trusty old Fluke and a clipboard, with conversions later. KIS is good.
 
Dan do you have any output filtering on the MPX sensors? It will be difficult to make readings with a multimeter without some sort of analog filter in-between the sensor output and your multimeter.
 
Oil Cooling

I have the inside of the cowl divided into three compartments - #1 is the area above the cylinders, #2 is the area under the cylinders leading back to the stock cowl outlet, #3 is the area above the lower cowl outboard of the plenum baffling & valve covers continuing back to the area between the engine and firewall with a vent in the upper part of the stock outlet but separate from the exit path out of #2. In race configuration I mount closure plates over all of the blast tube and heater ports at the rear of the plenum but the oil cooler continues to pass all of its cooling air into the large #3 compartment where it is vented through the small opening below the fuselage in the stock cowl cooling air port.

The small zone #3 cooling air port is provided by continuing the curved baffle dividing compartments #2 and #3 down below the fuselage in the stock cooling air outlet and securing with screws to the flange of the fuselage skin and firewall. The size of the opening ground adjustable and is controlled with AN960 washers stacked at the mounting points. The vent is approximately 3/16" tall and 7" wide. The oil temperature is stable at ~180 degrees under all conditions. I experimented with opening the vent to 1/4" and detected a slight decrease in speed with no noticeable change in oil temperature. I see no need for a separate inlet for oil cooling in my O-360 application.

I just completed a difficult (messy) instrument panel change and have not been working on speed mods this winter but I have been going over that cooling air exit in my mind for many months (years?). I want to come up with something that I feel confident will work which can be removed cleanly if it doesn't. I have the extra complication (opportunity) of the cowl and NLG strut support structure associated with the A model RV-6. The Larry Vetterman experiments add conviction that speed can be gained in this area. I am currently thinking of a vertical "keel" structure with a bull nose or wedge inside the cowl behind the FAB and flat plates attached to the keel in the back that are flat against the keel in ground and climb operations but the front deflects outward to seal against the leader bull nose or wedge in level flight.

Still not ready to commit but it is gnawing at me.

Bob Axsom
 
Dan,

Assuming you have a regulated 5v supply on-board to power the sensors, the supply decoupling circuit looks just fine from here. Just have to have clean power, so you aren't measuring alternator ripple. :)

T.J.
 
also, because the pressure changes are very small and the output range constrained to low value (5v spread) temperature stability can be a real challenge.

I used low pressure anologue sensors (0-1.0 psi and O-2.0 psi) in another project and found the output to vary significantly with temp requireing compensation.

I'd look at the sensor data sheet and see if it is supplied with a temp stability data. Also, all the components are affected by temperature to a certain extent - the power supply in particular. If the power supply voltage decreases as a function of temperature then you can't rely upon the sensor output for conversion unless you can account for it. Make sense?

Ken
 
I should add... the other option is for you to remotely mount all the sensors keeping the components from being exposed to large changes in temp but the only accurate way to do this is to provide temp compensation - which is a course all by itself :(
 
You'll have to help me Bob, I'm not an electronics guy. What sort of difficulty would you expect?

The data sheets suggest power supply and output filters for interfacing with the A/D input of a microprocesser. A quick bench check yesterday produced a stable meter reading without them.

2v2wlja.jpg

These sensors are very sensitive and they output even the tiniest of changes. What you'll see on a multimeter is a very erratic voltage and thus will be hard to read. Typically on a microcontroller one does filtering by taking something like 10 readings over X time and average them to get the real value.

Essentially this can be done in the analog world with whats called an RC (resistor/capacitor) low pass filter. There are many filters types that one could use but this is the easiest. I would try a 47K resistor in series and a 10uF resistor to ground. This will give you a time constant of .47 seconds.
 
I purchased the same sensors for my pressure measurements. The are fully temp compensating so that should not be an issue with them.
 
I agree with RocketBob about the filtering, since these sensors have a reponse time of 1 millisecond and can have an output that varies around somewhat with small changes in pressure due to minor oscillations in a plenum. Think of the plenum of a car interior with a window slightly open where it wants to take out your eardrums! A-D converters often don't like variations during their conversion process, so they often incorporate a sample-hold circuit. These are good, but due to Shannon-Nyquist sampling theorem they can often give strangely fluctuating readings.

I would suggest a very simple but easily accomplished filter of the data as it emerges from the A-D. A simple digital model of an RC network is to take the incoming value, subtract the previous filtered value from it, divide it by 2 or 4 or 8, which can be accomplished by a shift instruction, and add the result to the filtered accumulator. Using 2 is similar to having a time-constant of 0.693, and using 4 is similar to having a time constant of 0.288 relative to the accumulation interval and if you want more filtering, the 8 gives an equivalent time-constant of 0.134.

Where this kind of filtering works really well is if you are filtering data from a data set where the peridicity changes such as when counting rpm. That way the filtering is always a certain percentage of the periodicity, so whether the rpm is low or high, you still get the same filtering.
 
Last edited:
Dan, I didn't see the image in your reply as some of them get blocked by the internet nazi's at work.

The power supply filtering that the datasheet mentions is a good start for the 5V lead, which will help you get stable readings. On the output side of the sensor, connect a 47K resistor in series, and a 10uF capacitor between the resistor and your multimeter to ground. This will help eliminate noise and give you a more stable voltage to read.

As always my friend Paul is correct but it doesn't sound like Dan is ready to jump into writing code in assembler in the near term. :)
 
Last edited:
So, RB, Are you saying that fixed-point, fractional, binary machines are no longer derigeur? People no longer know octal notation? That they can't compute exactly the square root of a number in binary but must rely on the 1/2(X/A + A) approximation?
 
The DMM that the OP indicated he will use for data measurement already has a fair bit of signal integration. It may need a bit of analog filtering, it may not. As far as digital filtering, way beyond the scope of the task. But I wouldn't want to slow down the fine mental mastur / err, the fine analysis going on here. :D
 
Here's the basic diagram for the sensor board, should some brave soul with a stock cowl want to follow along. (Wade's cowl is also non-stock.)

2qla5c1.jpg


Currently no filters at all, but it's easy to add some later if necessary. As Ted mentioned, I've had good luck in the past using the old Fluke. We'll see; first step is to try the board while connected to ships power, alternator running, but with the air ports capped. If the meter is jumpy I'll add filters.

On the air side, I expect to use aquarium bubbler rocks or similar to avoid dynamic pressure effects on the small sensors, and perhaps machined restrictors to smooth pressure fluctuations. On the flip side, I'm kinda curious about what sort of flucuations may be present, so I don't want to start with a lot of damping.

Appreciate all the comments, keep 'em coming...very little electronics experience here.
 
Last edited:
While you are at it, take a reading at your dipstick, to measure crankcase pressure. I had dumped the crankcase vent on the exhaust pipe, as per Van's recommendation, and found that that location pressurized the crankcase to the exact same amount as the lower cowling, plus 1". By moving the crankcase vent outside, aft of the cowling outlet I was able to get a slightly negative number. Oil consumption, and oil on the bottom of the plane, did not change but all the little drips in the engine compartment went away. It would be interesting to see if you get the same results that I did.
 
So, RB, Are you saying that fixed-point, fractional, binary machines are no longer derigeur? People no longer know octal notation? That they can't compute exactly the square root of a number in binary but must rely on the 1/2(X/A + A) approximation?

Paul do you still have a rotary phone? :)
 
Paul do you still have a rotary phone? :)

No, but every time I call a company and their recording says "You may dial your party's number now.", I always ask the person who answers if they still have a PBX with rotary phones. They have no idea what I'm talking about! I did work for Ma Bell of Pa for 5 years after HS installing 'phones and teletypes and stuff.
 
I toyed with some VG;s fore of the cowl outlet. I did get a .5" more drop, in cruse using them.:eek: Saw some yarn tufting photos on this site. thought I would try smothing this turbulence out. burnt a bunch of gas measuring.:eek:
One of my last tests.
w8rq4k.jpg
 
Frank, does "FWD of baffle" and "Aft of baffle" mean measurements taken above and below the cylinders? Put another way, upper plenum and lower plenum pressures?
 
Frank's KIS manometer got me thinking. I'm going to ditch the electronic pressure sensors with their electrical noise and calibration concerns.

Much of the driver for using sensors was a desire to bring the signal to the cockpit via wiring rather than tubing. However, for inlet and exit pressure measurements I really need to tap the aircraft static system, so there's one tube fitting anyway.

Easy to make some small diameter steel "bulkhead fittings" for spaghetti tubing....old AN3 bolts and a few minutes each on the lathe.

Going over to direct pressure measurement offers a lot of options. About $60 will buy new differential pressure gauges calibrated in inches of water. There are handheld electronic manometers if you want to get fancy; China brands from $50, USA from $140 or so. An old ASI will read velocity. And some tubing taped to a yardstick will do in a pinch ;)

BTW, anybody have an 0-200 knot ASI cheap?
 
Last edited:
BTW, anybody have an 0-200 knot ASI cheap?

If your interested, I have a used Falcon ASIT260MK that is yours for the asking. For accurate test results, it probably need to be calibrated. I've already benefited from your contributions WAY more that this instrument is worth! Keep up the good work.
 
I used tubing taped to a yardstick. I brought about six tubes into the cockpit through the heat box. I removed the hose and used aluminium tape to seal around the tubes. I just connected the tubes to the manometer one at a time. The last test I did was a tube that was positioned at the opening of the engine cooing air inlet. There was WAY too much pressure there and it blew the water out the manometer all over me! I got a good reading from the upper mid and upper back plenum and lower plenum, dipstick and cooling air outlet. Keep it simple, cheap and easy.
 
Last edited:
Tom (or others of course), have you ever measured air temperature at the oil cooler outlet?

Last night I moved a temperature probe into my cooler outlet duct and ran hard to push oil temperature up. At 209F oil temp and an OAT of 70F, outlet air temp was only 110F, which I found to be a huge surprise. Maybe I should pull and test my vernatherm.

Postscript: Wrong...misread the chart, see below
 
Last edited:
Dan
I have not taken any temperature or plenum pressures at the oil cooler. It would be interesting to get a pressure reading at the inlet of the oil cooler and outlet of the cooler. If there is not a duct on the outside of the cooler then you would just be measuring the pressure of the lower plenum itself at that region.
With these two pressures you could determine the theoretical temperature drop and compare that to what you are actually seeing. I am sure that the supplier of your cooler would have charts with that data.
I would be very interested in your data of the oil cooler, especially if you have some before and after pressure readings on the out flow of the cooler, with and without a duct.
I ended up ducting my cooler to the side of the cowling, and it works, but I never did any actual measurements.
 
Whoops, misread my voltage-to-temperature conversion chart. Actual exit air temperature was 160F last night (not 110F) at 70F OAT and 209 OT.

This evening I rigged a manometer with one line to the oil cooler duct inlet and the other to the exit.

1" @1450 RPM warming up on the ground

7.5" climbing @100 knots IAS and 25/2500

12.25" @168 IAS (188 true) and 24/2400 (about 75%) 196F oil temp 66F OAT
3.34V 142F exit air

After a few minutes oil temp crept up to 208F @67 OAT
3.37V 147F exit air

So the exit air temp is more like expected, a heat transfer efficiency (temp rise/(oil temp - ambient temp) over 0.5, in this case between 0.57 and 0.65.

The 12.25" pressure differential across the cooler system suggests a mass air flow rate of about 44 lb/min per the Stewart Warner chart for a 10599 cooler. Hard to fault the duct design based on that chart, but does it really require more? Tom, how about some pressure measurements for your dedicated duct system....how much mass are you flowing?
 
Last edited:
Should be able to resume productive work shortly. Decided recent high oil temperatures didn't make sense, so today I pulled and tested the vernatherm. Turned out to be dead as canned tuna.....zero response in a hot water test:

25tyc9l.jpg
 
interesting timing on your post Dan. Were you by chance having fluctuating oil temp issues? Last couple of flights my OT steadily climbed to 210 in cruise flight having never been above 190 previously. It cooled down with power/mixture management. Next flight OT stayed rock solid in the 180's where it normally is. Was thinking maybe a vernatherm issue.
 
Should be able to resume productive work shortly. Decided recent high oil temperatures didn't make sense, so today I pulled and tested the vernatherm. Turned out to be dead as canned tuna.....zero response in a hot water test:

25tyc9l.jpg

Ahhh, homebuilding at its finest - I'm sure that setup passed Osha requirements,
much like when I did a similar thing with 400F peanut oil on an electric element
under the engine to test cht probes...

Look forward to seeing post-new-vernatherm data!
 
Ken, yes, oil temperature was inconsistent. Plus I had some high OAT conditions back in October and no oil temp issues.

The first signs of high OT's were not long after switching to the smallest exit. Overall I would see normal temps on the first flight then over 200 climbing out on the second. I assumed it was heat-soaking on the ground and the system lacked capacity. What made me curious was swapping back to a larger exit size with no change.....and then it started going over 200 on the first flight.

The little bugger was definitely working early on...good marking around the cone seat.

Round up an 1-1/4" socket. I set the camp stove so the water was gaining about 10 degrees per minute and kept measuring the vernatherm length with a caliper. This one stayed at 72mm from ambient to 210F.

Alex, gimme a break....the picture includes a fire extinguisher, and I wore my best boiling-water-resistant sneakers ;)

POSTSCRIPT: The vernatherm may not be bad after all. See below. Hmmmmm.....
 
Last edited:
Dan, that photo could spawn a whole host of off-topic responses. When I first glanced at it I assumed you were working on the kitchen stove, and was formulating a snide remark about doing that sort of thing while your wife was gone... until I looked closer. Myself, I mostly use my camp stove for making beer these days, but that's a topic for yet another thread.

As for the vernatherm- nice to have that information in the back of my head, since it looks like I may well fly my RV-7 before too long. In 40+ years of messing with cars, I've encountered plenty of thermostats that were either DOA or suffered rapid infant mortality. Are vernatherms subject to similar behavior? I'm not familiar with the technology (usually a wax pellet in automotive thermostats).
 
hi dan,

any news on your cowl exit progress?
what have you found to be the perfect size/shape?
any mods to the bottom of the firewall "lip"?

we did some tuft testing recently
http://www.flyvans.com/wordpress/index.php/2011/06/tuft-testing-on-the-lower-cowl/
scroll down for a video of the tufts and a landing.

we do have some vibration in the cabin floor and lack a few knots top speed and it's all most likely due to the fact that we had to enlarge the cowl opening to accomodate the special 4into4into1-muffler exhaust system.

we also fitted louvers to the cowling precautiously to aid with temperatures and run "cool" during breakin. this definitely proved to be working and the video would suggest, that the louvers also help to maintain a smooth flow of air/are not the cause of the turbulence vibration.

we're now trying to come up with a plan to tackle that item without it becoming a never-ending project...

regards,
bernie
 
Yes, Bernie, I've made progress, mostly measurements of temperature and pressure in a variety of locations. Recall I don't have the slightest issue with CHT. I really want a little less cylinder cooling and a bit more oil cooling.

Here's the rub...my system is not like yours, so the following may not apply well for you.

I'm running low velocity inlets and a throttled exit with a more-or-less sealed lower cowl. The lower cowl volume runs almost 7" H20 higher than a measurement point just aft of the cowl exit, with the medium exit panel, not the small one. With this cowl system I cannot dump the oil cooler into the lower cowl volume per conventional methods. It must be ducted to the cowl exit.

The exit end of the oil cooler duct system is several inches inside the cowl exit (ie forward of the firewall - see post #50). Pressure measurements tell me that location is roughly 3" H20 higher than the aft-of-cowl-exit location. Thus the simple way to increase mass flow through the oil cooler is to extend the oil cooler duct exit point rearward to the firewall plane and tap the lower pressure.

Another way would be to improve the oil cooler exit duct shape. Right now I think it chokes flow because of the shape of the fiberglass transitions and because of the SCEET hose bent into a 90; the interior wall is hardly smooth. I intend to build a new all-glass duct with a far better shape, with some improvements at the point where it dumps at the exit. Initially I will not extend it rearward. If I can get enough mass flow with duct shape improvement I will still be able to throttle cooler flow with cowl exit area reduction. That will be desirable in winter. At the moment I vary exit size with swappable exit panels. Later it may become a variable exit with cockpit control.

Right now temperatures are 95F to 105F on the ramp at midday in south Alabama, with 4000 ft OAT's of 75 or more. As a temporary measure I've taken my mid-sized exit panel and cut a 4" half-moon scallop out of the trailing edge. That moves the low pressure area at the cowl exit forward, meaning lower pressure at the oil cooler duct end. Because of the aluminum exit bell it doesn't change cylinder air exit area very much. A test climb yesterday at gross from 200 MSL to 6500, at 105 knots IAS and full power, gave me a highest CHT of 355F and oil temp of 205F. Oil temp dropped of course as soon as I leveled and allowed airspeed to increase. Quick fix; I may or may not get the new duct done before OSH.

Like your tuft test video...I should do it at some point.

BTW, a probe on the oil cooler duct inlet (a 4" opening on the rear baffle wall) provided the most surprising (to me anyway) temperature measurement; 15 to 18F higher than OAT. One obvious conclusion; all our baffle mounted or ducted-from-the-baffle coolers start with an inlet air temp disadvantage. A dedicated external oil cooler inlet is like a big drop in OAT.
 
Last edited:
BTW, a probe on the oil cooler duct inlet (a 4" opening on the rear baffle wall) provided the most surprising (to me anyway) temperature measurement; 15 to 18F higher than OAT. One obvious conclusion; all our baffle mounted or ducted-from-the-baffle coolers start with an inlet air temp disadvantage. A dedicated external oil cooler inlet is like a big drop in OAT.

So far.............with summer oil temps in the 180-195 degree range, I wouldn't want a farther drop in inlet temps. But then my 6A just cools well, with it's cooler mounted on the #4 baffle. I even have a Van's slide damper. Others, sometimes, don't do as well........I see.

L.Adamson --- RV6A
 
So far.............with summer oil temps in the 180-195 degree range, I wouldn't want a farther drop in inlet temps. But then my 6A just cools well, with it's cooler mounted on the #4 baffle. I even have a Van's slide damper. Others, sometimes, don't do as well........I see.

Cooling in an absolute sense is easy.....just increase mass flow. Temps will drop and the airplane will get slower.

To get 180-195F I install the "large" exit panel, which is about 77% of a stock exit.....but I don't want to go that slow.
 
Last edited:
hi dan,

thanks for the info...
i'm thinking about blocking some of the exit area initially without actually changing the outer shape of the cowling... (major mod)
just to see what it changes.
in theory this should accelerate the outflow air and lessen the turbulence below the belly skin.

the oil cooler by the way is the default one from van's and we're even partly blocking it off to get approx 190F in 80?F oat weather

regards, bernie
 
i'm thinking about blocking some of the exit area initially without actually changing the outer shape of the cowling... (major mod) just to see what it changes. In theory this should accelerate the outflow air and lessen the turbulence below the belly skin.

I have some reservations Bernie. Simply blocking exit area may or may not accelerate exit air at the enlarged ramp (it may accelerate air from the louvers or may push more air out around the propshaft opening), and will probably increase turbulence.

Want to do an easy experiment tailor made for your video camera? A lot of us think the -7's 90 degree corner at the lower firewall / belly skin intersection is a problem. Add a simple sheet metal radius, large as practical.

the oil cooler by the way is the default one from van's and we're even partly blocking it off to get approx 190F in 80°F oat weather

With both louvers and the enlarged exit ramp you have a HUGE total exit area. Perfect example of "just increase mass flow". No surprise about blocking the cooler.
 
Last edited:
i did not remove the cam mount yet... ;-)

the curved sheet metal is definitely an interesting idea, i've already thought about that as well, kind of rv-8ish... the problem will be fitting it as the area is pretty tight if i remember correctly / close to the exhaust. and it should be only semi permanent initially.

we did not actually intend to increase mass flow a gogo, it's more just a sum of fitting the louvers as per a number of recommendations and only then finding out about / fitting the muffler system.
well, for the breakin phase this certainly didn't hurt the engine and the airplane is still plenty fast.
now it's just a matter of unlocking the potential.
cowl is standard van's and i'd say the front inlet seals are tighter than average. (including blocking off the sides of the ramps in the top cowl etc...)


only problem is i'm having too much fun flying, and not enough time to tinker much at the moment.

cu bernie
 
With the store closed for the 4th I had some time to build a new cooler exit duct. The original was two fiberglass transitions and a 4" SCEET hose, and I suspect rather restrictive:

First%20Oil%20Cooler%20Exit%20Duct.jpg


So, start by shaping foam:

OC%20Foam%201.JPG


OC%20Foam%202.JPG


Prep it and do the layups:

OC%20Duct%20Prep.JPG


OC%20Duct%20Layup.JPG


Then remove the foam...rough duct ready for some sanding and general cleanup:

OC%20Duct.JPG


Hopefully back in the air by next weekend for test. I may need something similar for the entry duct from the plenum. Even SCEET (lined hose) is pretty bumpy on the inside ;)
 
Last edited:
Cooling in an absolute sense is easy.....just increase mass flow. Temps will drop and the airplane will get slower.

To get 180-195F I install the "large" exit panel, which is about 77% of a stock exit.....but I don't want to go that slow.

I agree with this statement completely, based on personal experience - no science, no engineering data - just experience.

The Van's cowl for the RV-7 has a exit to inlet ratio of about 107.69%.

My cowl, with a 5x16 exhaust exit area and 2 Bonanza side vents has a ratio of 219.51%

I also have a remote SW oil cooler with 4" scat duct from #4 aft baffle. It does not have the nice exit flow device as per Dan, but simply dumps its air into the lower engine area just above the exit. I've flown twice this past week with the OAT in the low 90's and oil temp ran at 168F! Yes, I have checked the sensor in hot water and I can get warmer oil temperature by closing the air flow butter fly valve.

The point here being, good cooling is the result of mass air flow through the cowl.

OK, does the 219% exit ratio introduce a lot of drag? Very little if any. I've done several WOT runs at 75% at altitude and the machine moves along right at 200 mph, about the same as Vans RV-7A numbers with a 180 HP engine.

Could be though, the extra drag is offset by the extra HP from BPE. The engine did crank out 187 HP on the dyno. :)
 
I also have a remote SW oil cooler with 4" scat duct from #4 aft baffle. It does not have the nice exit flow device as per Dan, but simply dumps its air into the lower engine area just above the exit. I've flown twice this past week with the OAT in the low 90's and oil temp ran at 168F!

I believe it. Understand why I need the exit duct and you don't...my lower cowl volume is pressurized, while yours is near static due to the huge exit area. I must duct to low pressure at the cowl exit.

OK, does the 219% exit ratio introduce a lot of drag? Very little if any.

Don't kid yourself.
 
Dan beat me to it. ;)

To tweak it a little. Good cooling is efficient mass air flow thru the cylinder fins. The major driver of that is pressure differential and making the air go where you want it to.
 
Last edited:
Say, Dan....

I'd like to build an exit duct for my -10's oil cooler as well. In case you don't know, they're mounted on a firewall box that's angled about 30 deg downward, with the air coming in the top.

How do I determine the exit area? The cooler is fed by a 4" duct off the back of # 6 cylinder.

Thanks,
 
I'd like to build an exit duct for my -10's oil cooler as well. In case you don't know, they're mounted on a firewall box that's angled about 30 deg downward, with the air coming in the top.
How do I determine the exit area? The cooler is fed by a 4" duct off the back of # 6 cylinder.

I'm not so sure you can accurately calculate the exit area for a particular mass flow. Pressure differential drives flow, friction reduces flow. There are equations for a theoretical perfect duct.....but we build imperfect things. It's hard (for me anyway) to define the friction due to the duct wall, bends, etc. Maybe one of the bright guys can chime in here?

The one I built this weekend is more-or-less a constant section area equal to the cooler face area, roughly 25 sq in. The previous setup took the cooler outflow and squeezed it back into a 4" tube (12.5 sq in), ran it around a right angle bend with a bumpy minor radius, then into another fiberglass transition whose outlet was partially blocked by the exhaust pipe. Surely the new one will flow more.

Pierre, the first thing to do is measure pressures. Run one leg of a manometer to the upper plenum. Run the other to a few different locations; the cooler outlet face, a spot near the cowl exit (where your duct might end), and perhaps just outside the cowl exit. Make a few flights and record the differential pressures at your usual climb speed and at cruise. A cooler exit duct would be a waste of time if there is no significant difference between the three locations.
 
Back
Top