|
AC – Air-bucket design-build for rear fender condenser
Am resurrecting AC system following Black Death (compressor failure.) While doing so, am taking some system-related side trips along the way. This is one of them.
Idea - improve rear fender condenser’s performance. I’ve got no complaint with prior condenser with Spal fan. AC with that config delivered 33dF at vent / 50dF delta. This design-build is based on premise that air ahead of rear wheel is cooler than air behind (where my fender condenser sits.) 911 here is ’80 SC with ’85 3.2. Right next to condenser is engine’s catalytic. Fender area’s hot enough so that placing a hand on fender under operating conditions = torture. Am into this project to see if a greater delta can be achieved by supplying condenser cooler air. Build - air-bucket addition to one side of condenser. Source air from just ahead of torsion mount and duct it to bucket. Put blower somewhere along the way. PS – This build was started under “Black Death” thread. It being off-topic under that thread… decided it being more appropriate to put this build on its own. Will pick up "BD" thread when this build is done. http://forums.pelicanparts.com/uploa...1501505036.JPG Serpentine condenser that was in system was cleaned after system’s “death.” Did so under assumption serp would return to service. When considering reinstall of cleaned unit… changed my mind about doing so---wasn’t worth the risk. In its place is this new 12x14” parallel flow unit. http://forums.pelicanparts.com/uploa...1501505036.JPG 4” duct is planned. http://forums.pelicanparts.com/uploa...1501505036.JPG 4” positioned above wheel. http://forums.pelicanparts.com/uploa...1501505036.JPG Intake. http://forums.pelicanparts.com/uploa...1501505036.JPG New condenser fitted to existing hose. Not working. Hose in this config occupies same space as tire. Adding more hose via splice is option but prefer to reconfig ports on condenser. http://forums.pelicanparts.com/uploa...1501505036.JPG Sheet aluminum. Attempted 730dF then 350dF welding media on this material. Is not as easy for me as it appears on YouTube. Rather than delay progress while I learn to weld thin sheet… am getting on with bonding seams with epoxy gel. http://forums.pelicanparts.com/uploa...1501505036.JPG http://forums.pelicanparts.com/uploa...1501505036.JPG Plumber’s Putty---epoxy putty. Pushed in hard and dragged slightly, stuff adheres well to this aluminum. Cures rock hard. Whether this media is suitable in a vibrating & flexing environment is questionable. Since it’s “dressing,” doesn’t matter if it fails structurally. Am applying it here mainly to see how it installs, adheres, and finishes. Will check durability downstream. http://forums.pelicanparts.com/uploa...1501505036.JPG Right side putty in pic has been sanded. Left side is raw. http://forums.pelicanparts.com/uploa...1501505036.JPG Planned port positions. |
http://forums.pelicanparts.com/uploa...1501505036.JPG
Testing 4” bilge blower and accordion duct = almost total failure… “almost” because dribble of air does come through condenser. Is useless. Attaching a low profile sucker Spal to condenser was tabled idea. There’s space for Spal but then no space for air to go. Putting pusher Spal inside bucket was another direction suggested. Doable fit but… after checking CFM chart for smaller Spals… it’s a dead end---smaller Spals are comparable to bilge blower CFM. http://forums.pelicanparts.com/uploa...1501505036.JPG Accordion duct is replaced with smooth bore, thin gauge product. Is modular and has universal elbows. Retest of bilge blower with smooth bore… http://forums.pelicanparts.com/uploa...1501505036.JPG Config here guestimates final duct length and number of turns, idea being to approximate air resistance imposed on blower. http://forums.pelicanparts.com/uploa...1501505036.JPG Static / Blower OFF. http://forums.pelicanparts.com/uploa...1501505036.JPG Blower ON - no internal deflector. http://forums.pelicanparts.com/uploa...1501505036.JPG Blower ON with internal deflector. Mod to deflector may manage more air flow to top corners of condenser. Compared to dribble of air with accordion duct, air flow’s improved but only marginally. Search for more air continues... http://forums.pelicanparts.com/uploa...1501505036.jpg Micro-turbine. Radio controlled (RC) jets use these. For size, is big air mover. Some units produce 10+ lbs of thrust. Big air comes at a price… units are screamers---no issue for flight but for Porsche AC app… uber dysfunctional. More downside… electrical service to “EDF” (Electric Ducted Fan) is more complicated than hooking up basic 12v blower/fan. More… Credit to Glen, a.k.a GH85Carrera, for thinking through driving environment relative to applying a turbine. Point is: debris sucked in and onto EDF’s finely balanced blades is surly bad news. Fine-particle air filter to protect blade balance is called for. Those who know me know I’m willing to complicate design-build work to benefit from the journey taken but… adding a GUARANTEED maintenance process---aka cleaning a condenser’s air filter---checked myself and I won’t do it. Still, EDF for this app has distinct advantages---brushless motor, compact size, aggressive air movement---am not abandoning it yet. Instead of plastic EDF seen above… an all-metal unit maybe? http://forums.pelicanparts.com/uploa...1501505036.jpg More durable than plastic but to test idea… will cost around $400 (for EDF alone.) With no guarantee EDF concept will work, flushing $400 down toilet is possibility. Unappealing. Search continues… http://forums.pelicanparts.com/uploa...1501505036.jpg HP computer server cooling module. Inside is “industrial grade” EDF. Reconditioned unit = $10 plus shipping. “Reconditioned…” fine word with infinite possibilities. For $10, is small caliber bullet to bite if unit craps out sooner than later. |
http://forums.pelicanparts.com/uploa...1501505036.JPG
Unit arrives with case looking like NEW. Housing removed. Guts have no dust. Turbine spins smoothly. Only issue found is plastic housings on both ends are loose. Found internal stand offs being slightly oversized. Will correct later if this thing works. http://forums.pelicanparts.com/uploa...1501505036.JPG Electronic board seems overkill for a blower. Not going to use any of this electronics---is very likely converting AC to DC. Will give RC jet electronics a go (speculating that all brushless motors work on same electronic principles. I may be wrong.) http://forums.pelicanparts.com/uploa...1501505036.JPG More meat here compared to RC jet turbines. Unit’s zinc alloy. Down side might be moisture injestion. YouTube’d brushless motors and water. Found submersible maker using brushless motors---with open cases---in salt water. (He flushed them in fresh water afterwards. He also did some mods to motor internals to prevent salt ruination of parts.) Is positive info but offers no guarantee motor in hand will survive moist environment … and if so for how long. Conclusion = moisture won’t be short term killer. Am moving forward with unit.Will incorporate splash shield at intake. Given reasonable duty cycle---am thinking 4 years minimum---at $10 a pop, can stock up on a few refurb’d units. Will do so after proof of concept arrives. Two critical questions loom at this moment. 1. Will motor in hand work with RC EDF electronics? 2. Will turbine’s noise level be suppress-able enough? http://forums.pelicanparts.com/uploa...1501505036.JPG To get motor running... according to RC flight hardware specs, EDF requires Electronic Speed Controller (ESC,) and either a receiver/transmitter or a Servo Tester (ST) to make motor do what you want it to. (In simplified terms, brushless motors operate on an electric pulse system via 3 wires---can see those 3 wires running from ESC to motor, pic above. Pulse speed as distributed by ESC to motor determines motor’s speed. Speed is dictated to ESC by either a receiver/transmitter or in this case an ST.) According to EDF pilot wisdom, one gets an ESC with a bit more amp capacity than motor uses. At max output, motor in hand draws under 30amps. 40amp (Turnigy Plush unit with a Battery Elimination Circuit---BEC) is purchased… unit in pic above. BEC incorporates a 5v power supply along with the signal wire from the ESC---smal white-red-black wires seen coming out bottom of ESC. As name implies, a BEC does away with need to manually add 5v power supply to either the receiver/transmitter or Servo Tester. (Am total RC flight noob getting school’d.) http://forums.pelicanparts.com/uploa...1501505036.JPG ESCs need to be programmed before use (so ESC knows how to operate.) Can do this easily with programming card, in this case a Turnigy card to match Turnigy ESC. Programming takes just seconds and ESC’s supposedly ready. This card, ESC and ST are a few dollars in total. http://forums.pelicanparts.com/uploa...1501505036.JPG Once ESC is programmed, card is detached from ESC and Servo Tester (box with blue light) is connected to ESC. Then power is connected. (Apparently connection sequence of this flyer-gear is very important---hardware is supposedly ruined if wrong connection sequence is done.) ST has knob that turns motor ON and allows for speed/thrust adjustment. Power light ON here is first sign system lives. I turn knob and thrust too lives… albeit exiting intake’s side. To invert, swap any 2 of the 3 wires serving motor. Done… http://forums.pelicanparts.com/uploa...1501505036.JPG Thrust is now exiting correctly. Turbine’s speed is controllable via ST. And noise from unit seems manageable given estimated speed of turbine needed to deliver a significant volume of air through condenser. Challenge of significance is quickly uncovered. If 12v power feeding ESC is disconnected, idea was… when 12v power was resupplied, motor would resume working at speed it was at when previously shut OFF. That’s not happening. Am having to turn Servo Tester OFF, then dial it back up to speed each time 12v power is turned ON. This protocol makes perfect sense as a fail-safe in RC jet pilot world. For Porsche AC app… this protocol is not amusing. In Porsche AC app, ON/OFF power intended to serve ESC unit is related to compressor’s ON/OFF cycle. You see problem? One way to go is to constantly power ESC but this still requires manually turning UP power to ESC each time AC is turned ON. Accepting this is not happening. There’s a way around it. Just need to discover that way. Any ideas? (Anyone know what failsafe circuit will look like? I’ll cut it in ESC.) Back to bucket building… http://forums.pelicanparts.com/uploa...1501505036.JPG Bottom brackets formed & attached. Time to fit-test in wheel well. (Cardboard is over air-exit side of condenser to keep fins from being mauled.) http://forums.pelicanparts.com/uploa...1501505036.JPG Air-bucket’d condenser sits as far back as is reasonably possible. There’s maybe ¼” clearance at top rear. Unit tilts top-forward to follow contour of tire. To get a little more clearance between top rear of bucket and fender… will reduce height of rubber mounting doughnuts later. (Bending rear bracket is not an option.) Wing bracket that mounts to fender seam is not yet made/installed. http://forums.pelicanparts.com/uploa...1501505036.JPG Bucket’s in/out angle is set to allow just enough clearance for eventual insulation to envelop bucket. Aluminum foil stuffed between condenser and chassis is a casting. Will refer to this foil casting when making screen-guard for air-exit side of condenser. Pointing of air-inlet duct as it is… is not good---should point more towards chassis. Not sure at moment if this is major or minor problem. Need to fit ducting to make final call here. |
Making screen-guard…
http://forums.pelicanparts.com/uploa...1501505036.JPG Foil cast is referenced for space allowable for guard that will protect condenser’s exposed fins. http://forums.pelicanparts.com/uploa...1501507649.JPG Aluminum rod. Copper electrical lugs for footers. Will rivet footers to condenser’s frame work. http://forums.pelicanparts.com/uploa...1501507649.JPG Aluminum mesh is “guard.” http://forums.pelicanparts.com/uploa...1501507649.JPG Mesh edges and footers will position underneath trim-strip. Mesh is notched at footers so trim strip will fit tightly over both these parts. Believing simple solution for “ON at prior speed” is possible to achieve for automated turbine motor operation, am moving forward with turbine unit’s install. Not having to operate turbine at full-screamin-throttle solves noise puzzle in part. Still, total silence---as heard from cabin with windows closed---is goal. Isolating turbine housing from rest of system seems one requirement. http://forums.pelicanparts.com/uploa...1501507649.JPG Turbine’s housing O.D. is 70mm---measurably smaller than 4” duct. I questioned whether to narrow the duct at the turbine… or… to center turbine unit within solid 4” aluminum tube. Tube direction is chosen based on materials on hand. http://forums.pelicanparts.com/uploa...1501507649.JPG Rubber fenders will center and isolate turbine housing inside 4” tube. Using EasyMold. 2 stainless 4-40 screws with washers get embeded into each fender. These two screws will secure fender to turbine housing. Once all fenders are mounted to turbine’s house… house unit gets inserted into solid aluminum tube. Will secure fenders to aluminum tube with wood screws… plus compression-grip of fenders. http://forums.pelicanparts.com/uploa...1501507649.JPG “Mold platform” set-up with screw-pairs in place. Depth of screws in platform is slightly shy of turbine housing’s thickness. http://forums.pelicanparts.com/uploa...1501507649.JPG EasyMold (EM) cures to hard rubber. Access to embedded screws is provided by channels cast into EM---small pipes placed into cast. EM cures quickly (possibly adjustable by altering quantity of hardener---I’ve always mixed 50-50 so not sure about this.) There’s just enough time to fill and work single mold in this case before time’s up. Trick = ensure entire mold cavity is filled and media is “internally connected” before times up. Once EM reaches a certain point, it no longer bonds to itself meaning if layers are added into mold and not blended together before time’s up, casting is fractured. http://forums.pelicanparts.com/uploa...1501507649.JPG Casting’s came out good except for each mold being over-filled. Was going to “squeegee” excess off top of molds but spent time ensuring cavity being filled and good internal blending existed. http://forums.pelicanparts.com/uploa...1501507649.JPG Excess rubber is sanded away from every fender. Rough grit takes EM down easily. Each fender’s height is now set to grip-compress 1/16” when installed. |
Given less than half throttle speed required of EDF to deliver high volume of air through condenser, exhaust turbulence that was a possible noise influence to deal is no longer. Am still curious to know if chevron exhaust will scale in order to quiet an EDF run at full throttle.
http://forums.pelicanparts.com/uploa...1501507649.jpg Making micro-chevron duct… http://forums.pelicanparts.com/uploa...1501507649.JPG http://forums.pelicanparts.com/uploa...1501507649.JPG http://forums.pelicanparts.com/uploa...1501507649.JPG http://forums.pelicanparts.com/uploa...1501507649.JPG Going to run motor with and without chevron. Post result later. Tested 8.1.17 - In given config and at full speed there's no audible difference with or without chevron. Fitting duct in wheel well is on today's menu. |
Wow, you keep building and thinking. Cool project.
Jay Leno has a jet powered motorcycle that he admits is insane. He said it is funny to pull up to a red light and see everyone looking UP to look for the low flying aircraft that must be near. I feel sure yours will not be as noisy as a helicopter jet engine. It will be interesting to see if is heard from inside the car with the engine running. You must be a master of Google to find all the products you keep using. The rubber bumper material and aluminum weld sticks. Cool stuff. I will keep in mind for my next garage project. Keep up the cool engineering projects. I am enjoying watching you work! :) |
Quote:
I do indeed spend a good deal of time hunting for ideas & solutions in Google and YouTube. Looking for one thing often leads to uncovering some other interesting thing. |
Wow Great Stuff Karl!
Turning your car into a Meat Locker! Enjoy following your projects |
^^^ Hope all's well with you RN! Been a while since we talked AC :) Glad you're enjoying project!
|
Test fitting ducting in fender.
http://forums.pelicanparts.com/uploa...1501541901.JPG Was planning on one elbow duct section connected to air-bucket... then a straight section where turbine would be... then an elbow to a straight section above the wheel. Point being to mount turbine section as far away from cabin as possible. Not happening. There are bends in duct from beginning to end here. Is difficult to see because of camera angle. Were one looking through fender… constant arch of duct would be seen. Plan “B” is to mount turbine next to cabin wall. Noise suppression challenge has just been elevated some degrees. http://forums.pelicanparts.com/uploa...1501541901.JPG Straight thin-wall duct squashes easily. Have not yet quashed an elbow piece. Is possible elbow's formed seams will not handle squashing well. I did not install any test duct just ahead of duct section above wheel. Space was measured instead... primarily for turbine and Glenfilter. No problems are seen fitting these sections in. Glenfilter making… Idea - filter large, lightweight debris greater than 1/8” diameter. Block splashed water from entering system. What water does reach filter mesh, offer it ability to gravity-shed rather than going further. Material is 0.040” aluminum with a 1/8” perf. http://forums.pelicanparts.com/uploa...1501541901.JPG http://forums.pelicanparts.com/uploa...1501541901.JPG Tested 350dF Muggy weld rod on this perf'd stuff. Also checked how quickly this perf'd material looses form when overheated---it does so quickly. Been playing around with the 350dF from Muggy Weld, their Super Alloy 1. May seem like 350dF is an impossible bonding temp but results are impressive---even in hands of welding moron (like me.) Shall suggest flux (that comes with rods) has much to do with bond---is chemical reaction bond as opposed to heat bond. Would I use this on connecting rods? NOT. Marketing propaganda holds welds being harder than surrounding metal when cooled---that's true. http://forums.pelicanparts.com/uploa...1501541901.JPG http://forums.pelicanparts.com/uploa...1501541901.JPG http://forums.pelicanparts.com/uploa...1501541901.JPG http://forums.pelicanparts.com/uploa...1501541901.JPG 4” O.D. aluminum tube with 1/8” wall serves 2 purposes. It’s body of completed Glenfilter… and it’s what y-arm bracket (mounted to chassis) will strap to in order to secure entire leading section of duct (to chassis.) Both ends of tube are skinned so 4” O.D. thin-wall duct can slip onto it. Minimal amount of hard-wall (1/8” thick) tube is applied in order to minimize accelerating/decelerating system mass over bumps. Attempted to apply this thinking throughout system. http://forums.pelicanparts.com/uploa...1501541901.JPG Cone and tube secure with gel epoxy where they meet. http://forums.pelicanparts.com/uploa...1501541901.JPG Glenfilter = done. Flip this upside down and it’s in operating orientation. Add thin-wall duct section over cone area that's extends slightly beyond nose = splash-guard. No maintenance is planned. Whether this is functional in real world---or not---has yet to be seen. Am back to ESC to sort out electronic challenge. |
Karl,
I will be very interested to see which inline fan you chose to use. Obviously the more CFM the better. The best inline 4" 12v fan I found can move 270 CFM (simular to the one in the first photo in post #2), wondering which of your fans can beat that and what is their CFM. |
Karl,
I was thinking about the tubing . . . Would Aviation HVAC tubes work? Lightweight flexible and durable. http://forums.pelicanparts.com/uploa...1501556250.png http://forums.pelicanparts.com/uploa...1501556292.png http://forums.pelicanparts.com/uploa...1501556408.png |
Quote:
Quote:
|
No sorry, no source. I see the guys at my base using the stuff.
I bet eBay has some ... Probably venders directly online |
Yo' Kinky Karl, Aircraft Spruce is one good source for the tubing (called SCAT tubing).
Ducting from Aircraft Spruce |
^^^
Thanks for idea RN... damm fine upgrade! RonO... you the wingman :cool:! Just checked Spruce out---am on it. TY! |
Quote:
Give me a shout out when you apply for a patent. :) . . . and shouldn't Rawkness be trashing Targas by now? |
Before going further with brushless motor program, been researching to see if & how unit in hand can be controlled “correctly” as this AC app requires. Appears it can but not without experiencing a mucho greater learning curve than anticipated. Problem that needs getting around is... standard RC speed controllers need to “arm” themselves every time before they begin to operate a brushless motor. Attending this manually is certainly doable but also certainly undesirable. There are any numbers of ways to solve this problem---challenge is to find simplest way to automate it. At this moment, using an Arduino is most appealing path.
...Point in this research is to determine---right now---whether brushless motor direction is viable or not. If not, time to head in new direction. Have chosen to stay with brushless and sort it. http://forums.pelicanparts.com/uploa...1501685914.JPG Tap = 4-40 (for screws embeded in fenders.) http://forums.pelicanparts.com/uploa...1501685914.JPG Fenders attached for compression check. 1/16” compression per fender is too much---makes for difficult fit of unit into aluminum tube. (Housing's footers will remain in place until there’s no need to bench test motor further. Will cut footers off eventually.) http://forums.pelicanparts.com/uploa...1501685914.JPG Sanded fenders down to 1/32” compression per. Should still be good grip (when inside tube) while allowing unit to get into tube without destroying fenders along the way. Finishing intake stack. http://forums.pelicanparts.com/uploa...1501685914.JPG http://forums.pelicanparts.com/uploa...1501685914.JPG Splash shield and top sections added. http://forums.pelicanparts.com/uploa...1501685914.JPG Top-to-bottom measure = 18.5” Will fine tune this measure based on next fitting into wheel well. Right now plan is to position turbine blower just after bend at stack’s top. Aviation duct RN suggested may change this positioning---may be possible to move blower as far from cabin as possible… close to condenser. Idea being to manage turbine noise---as heard from inside cabin, windows closed---with every incremental move possible. Aviation duct is ordered. Will do another fit-in-car as soon as it arrives. Will be interesting to see how that stuff compares to turns metal elbows make. Main challenge on hand now is designing "elegant" electronic solution for motor's operation. If anyone happens to know Arduino sketch that will automatically arm and operate brushless 3-phase motor at prior shut OFF speed... say so please. |
Quote:
RonO is just around next bend with his pizzing on Targas. ...Is OK. We luvs the Norflander all the same. :D |
Wiring plan. All this is necessary to deal with Electronic Speed Controller (ESC) needing to be armed every time it operates. Manually arming ESC... possible but not happening here. Why bother with ESC? Because there's no other way (I've found) to move big air through this "ducted bucket system" than with a brushless motor spinning a robust turbofan.
http://forums.pelicanparts.com/uploa...1501847375.jpg 4-wire cable goes through engine-bay-to-cabin bulkhead---plan is to run this cable into cab next to DMI harness. Arduino and ESC are headed below-dash somewhere. ON/OFF for turbine is activated by circuit from compressor---when compressor is ON... turbine is signaled ON. 12v power feeds into ESC. When ESC gets 12v, BEC cable off ESC supplies 5v power to Arduino (no step down needed.) Arduino is programed to first arm (initialize) ESC... Arduino program then instructs ESC to accelerate turbine to a defined speed. (Defining what speed will require a few test runs to sort out.) Program/sketch is possibly ready. Am waiting on Arduino to arrive to get on with this. Am collecting cable & bits today. Also waiting on smooth bore flex ducting. When it arrives, one more test-fit is called for to get a final install measure for all parts. |
| All times are GMT -8. The time now is 08:31 PM. |
Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2025, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.0
Copyright 2025 Pelican Parts, LLC - Posts may be archived for display on the Pelican Parts Website