created June 12, 2004, updated August 1, 2004
I finally got my Ellison EFS-3A in, after deciding to give the Weber a break. I'm not sure that there's anything wrong with it, but it bothered me not having immediate control over the mixture. Changing jets works on the ground is a pain (it takes 10 minutes, since I have to take the carb off each time), and there is no mixture control at all while airborne.
The most disconcerting thing about the Weber was the pool of fuel that it left in the bottom of the intake manifold after a few minutes of full throttle. That may have been due to the close proximity of the vertical section of the intake with relation to the carb. It was only about 2.5 inches. My Ellison manifold has 6" of run before it turns vertical, which I'm hoping will alleviate that situation. See more details on my Ellison intake at the bottom of Intake System.
The EFS-3A is now installed, and the big benefit is a mixture control. I spent the day running it, and I must say it's going to be perfect! It starts instantly (after it's been running), idles smoothly, and runs wide open almost as smoothly. Mixture changes just a tad from dead stoichiometric (14.7 parts of air to one part of fuel by weight) to slightly rich, but I only know that because I have a sensitive air/fuel mixture gauge based on an oxygen sensor.
The manual says you need a primer, and I'm beginning to believe it. It's pretty hard to start cold, because the carb doesn't flow any fuel until there's a substantial vacuum on the carb throat, and there's no accelerator pump. To get ANY fuel to flow at all, you have to stick a piece of plastic tubing over a port and blow into it, which causes the fuel rail to spray fuel. I'll order an electric primer solenoid and plumb it in next week, and that'll fix that situation. There's a lot more information on the EFS-3A at Ellison's website, applicable to ALL aircraft intake systems, so check it out.
I ran it up to about 3000 this afternoon (the 3050 shown by the EIS is a spike). As you can see, the John Deere dynamo is charging nicely at 14.4 volts, highest EGT is 1387F, oil temp is 153 with 43 psi of pressure, 2.6 psi of fuel pump pressure, and highest head temp of 413 degrees F. One thing I learned at William Wynne's Oshkosh Corvair forum is that the stock Corvair's temp overheat sender turns on a light at 575 degrees F, so I'm a long way from that! Fuel pressure is about 6.1 psi with the engine off, 5.9 at idle, and drops to 2.6 wide open throttle (WOT). You can see the mixture gauge to the right of the EIS, slightly richer than perfect. I don't have any manifold pressure readings since I don't have that hooked up yet, but that's in the plan for tomorrow. The air/fuel ratio gauge is to the right, and is indicating a good mixture, right in the middle. When I switched it off, there was a lot of condensation on the outside of the manifold right behind the carb (humidity was close to a hundred percent, since it had just rained), but it was not cold to the touch at all. This lends credence to the opinion that the Ellison is more resistant to carb ice than most, due to its lack of a venturi. That being said, "CARB HEAT REQUIRED" is cast right into the carb casting, just in case you didn't read the extensive manual.
I ran it with the cowling on and off, and with the K&N air filter removed and installed. It runs slightly better with ram air, from what I can tell. I was able to spin my Sterba 52x66 up to 3000 rpm static, and that was with the cowling installed, no air filter, and no ram air. In this configuration, the carb is breathing hot air from the heads and cylinders above. I was worried that I wouldn't be able to spin up to a high enough rpm at high speed, but Troy Pettway tells me his C-85/0-200 will only turn his 59x74 Sterba at 2650 rpm static, but it spins up to 3300 rpm at 230 mph. If I too can gain 650 rpm at top speed, I'll be turning 3650 rpm (where I'll be pushing 130 horsepower),which is certainly good enough for a starting point, and close to the ballpark that I'd like to run in.
My engine runs fine right now, from idle to full throttle, but the rpm wanders a bit, and it's not absolutely perfectly smooth like I think it should or could be. These EGT values (1-3 top row, 4-6 bottom row) show that there's something funny going on in there. I have no reason to doubt the probes, since they all read the same number when the engine is cold, and they are all located exactly the same distance from the cylinder head. I'll swap some around just to prove the point, but I don't think I'm going to find a problem there. The funny pattern is that the right side of the engine is 1167, 1187, and 1154F (going rear to front), which is all close enough for me, but the pilot's side is 1081, 1334, and 1424F (rear to front), which tells me number 6 is running awfully lean compared to the rest, and number four is lean too when compared to the other four. Not sure what's up with that, but I'll try to figure it out. I've already checked compression and they are all very close to 128 psi. I haven't bothered with a differential pressure test yet, but I may do that today. First I'm going looking for intake leaks on the pilot's side of the engine. One more thing I'll do today is weld on a manifold pressure port to the intake between the carb and the Y.
So far, I have no oil leaks and no overheating problems. The alternator belt is still tight after three hours of running, and I've seen no other problems with anything, so I'm pretty happy with the system. Maybe I'm just spoiled by fuel injected cars, and this thing does have a Rube Goldberg intake system on it, but it could be a tad smoother at wide open throttle. Since the EIS has shown me that I have some EGT issues on the pilot's side of the engine, I know that I'll have to keep an eye on it, and eventually figure out what's going on. But right now, I have no hesitation about flying it just like it is, and with the knowlege that it will even get better later on.
June 12, 2004
Last night I got everything ready to run, and thought I'd give the engine a try. I'd already set the timing statically (I'll never understand why some folks would rather goof around with the distributor while the engine's backfiring, looking for something close to correct timing), so it was within a degree or two of optimal already. Well, on about the third blade, it fired right up and idled smoothly for a few seconds until I shut it off. The only problem was I was still inside my basement, so it took about 15 minutes for the dust to settle down, and for me to clean up the mess!
Here's what the engine looked like when I finally decided I was ready to run it. Yes, there are lots of loose ends, such as lack of locknuts and that sort of thing, but this is good enough for test runs. I hoaked up an air filter out of some sheet aluminum, a K&N filter element, and an old Weber air filter housing top. It will get me by until I can fabricate a proper filter housing with carb heat. The carb is a Weber 40DCOE that I liberated from a 1967 1600cc Alfa Romeo, found in a German junkyard in 1977. It had two on it, so I also have a spare just like it. I can't resist gloating that I paid $7 each for them! So it's not that I thought the Weber would be a perfect airplane carb, but I had two in-hand, and in the automotive world, the Weber is considered to be the ultimate high performance carburetor. One thing I know for sure is that I put this pair of carbs on my fuel injected CIS Rabbit about 15 years ago and it turned it into a tire-smoking beast!
Only problem is that it doesn't have an inflight mixture adjustment, so that's a problem. The Weber has an enrichment device for cold starting, rather than the more typical "strangler choke" so I'm hoping to rig that as "normal", and the "choke" will become the leaning knob. Another possibility is controlling float bowl pressure, which I'll explore later if I have to. Carb heat and an outside ram air source will have to come later, after I've convinced myself that the Weber is worth the trouble. If not, I'll be ordering up an Ellison EFS-3 pronto.
Here you can see the oil cooler and shroud on the passenger side of the firewall.
Today I took it outside, tied it to a tree, and ran it up for the first time. Again, it fired right up, but the air/fuel ratio gauge showed it running lean at idle, and extremely lean at midrange and high rpm. If I'd run it for any amount of time that way, I'd have fried the engine from detonation.
This is a Cyberdyne air/fuel gauge that I bought years ago for cheap. I think they've departed the scene now, but lots of folks make 'em now. When properly leaned, the lit bars only extend upward from the bottom, up to the bottom green LED. Here it's showing way rich (the engine is turned off, so it's reading the open exhaust pipe). [Dec 2005...check out the LASCAR EMA 1710 mentioned at the bottom of the page. It's a $26 LCD analog volt meter that works perfectly as a mixture gauge, since O2 sensors output 500 mV when the mixture is perfect. The only problem is that the needle is very jumpy without some sort of R/C circuit to smooth out the O2 sensor signal, but it does work good enough without that to give an indication of too rich or too lean.
It uses a run-of-the-mill $30 Bosch "one wire" oxygen sensor (part number Bosch 0 258 001 027) with an 18mm sparkplug thread. Most of the gauges work directly off of a similar generic oxygen sensor. I welded a stainless boss that I got from Burns Stainless into the exhaust downtube to mount the sensor. Although I have dual exhausts, I'm not sure if I see the point in having two of these.
The jets that came in the Alfa were 110 mains, 180 air correction, and F16 emulsion tubes, using 30mm venturis. That setup was way too lean for this engine, probably because the Alfa was feeding one cylinder per barrel, and my setup feeds three cylinders per barrel. Fortunately, I have 8 different Weber carbs that share those same jets and tubes on the property. The 2110cc Karmann Ghia had the biggest ones in its 48 IDF carbs...so I tried them. 165 mains, 180 a/c, and F2 emulsion tube. This time, everything but idle was rich. So I split the difference and tried a 135 main, 180 air, and F11 emulsion tube that came out of a pair 40 IDFs from a 1600cc Fiat sedan in the same junkyard.
I adjusted the idle a bit also, after reading up to remind myself how to set those up (3/4 turns out and screw the adjustments in and then out and center them up where it runs best, then alter the idle speed adjustment screw to the right range, then repeat if necessary until the idle speed is right).
When I started it, I was amazed to see the air/fuel meter stuck to the center "perfect" mixture point almost throughout the range! I could actually fly this thing just like it was. And best of all, I didn't have to mess around with it for weeks like I thought I'd have to! What really made it simple was that air/fuel ratio gauge. That dispensed with all the guesswork, and replaced it with concrete information that was easy to act on. I highly recommend these things as one of the best hundred dollar investments you could put on your instrument panel.
The exhaust is not nearly as loud as I thought it would be, and I haven't even installed the baffles yet. Only problem was, the EIS tach was reading 1200 rpm at a reasonably smooth idle, when I knew it was really a lot lower.
Time for a little data collection, so I checked max RPM with WW's loaner Sterba 54-67 cut-down prop. I managed to turn it at 5500 rpm. No, I didn't think you'd believe me, and I don't believe it either. In my mind, that confirms the EIS is a little out of whack. My guess is that it's off by 1.5x, the difference between a 6 cylinder and a 4 cylinder. I have the "TachP/R" (tach pulses per revolution) set to 3, so I don't know what's up with that. You can bet I'll be on the phone with them first thing Monday morning to find out though.
But wait, I also have a Tiny Tach in the panel that ought to shed some light on this situation....but it's reading something like 6900 rpm, so I think there's a "little" problem with it too. There's no real correlation between the Tiny Tach and EIS either, as one reads 4400 rpm when the other reads 1600, so no magic conversion factor there. Yes, I know there are three different versions of it, and I have the four stroke version, "supposedly". It wraps around a single cylinder's plug wire, so I have no idea how it could be anything but deadly accurate. The wires for it are run such that interference from any other EMI source should not be an issue, so it looks like I need to do some more research on this one!
I later stole the VDO tach out of my Karmann Ghia to get some "realistic" numbers, although I had to multiply the readings by .66 to get the difference between 4 and 6 cyliders out of it. After to talking to the Tiny Tach folks, I found that they recommend a resistive network filter for a 6 cylinder application, which I've just finished up. I'll see how that works tomorrow, and post a photo and schematic if it works.
Needing a scale factor for RPM, I bolted on Doug Steen's 58x74 Aymar Demuth prop that I ran down at WW's Corvair 2002 College. I'd managed to turn it 2650 rpm on his test stand using a small-venturi A-65 carb, and was hoping for a slightly larger number with the Weber. Assuming that 66% (4/6) ratio, I turned it 2900 rpm today, so that's just about in the ballpark of what I'd have expected. I've always thought that was too much prop, but it might just work depending on how it unloads at high speed. I'll have to check it against a high rpm Sterba, which I suspect will do a better job in the top end department. I think my Corvair's HP curve will beat the prop diameter efficiency curve to the extent that I'll get some pretty good speeds spinning the smaller prop fast, like 3800-4000 rpm for takeoff. Only real flight tests will tell for sure, and all of this is dependent on what rpm I'm actually turning.
So right now the tach issue has left me with just about as many questions as answers, but I'm working on it. But the reason why this whole thing is such a big deal to me is that I now have an airplane, rather than a project in my basement. I can roll this thing out side, touch the key for a half a second, it fires right up, and it even LOOKS and SOUNDS like an airplane! This is a huge leap forward, as I now have no doubt that it will work as is (with a carb heat muff addition), and I can go on about the business of adding seatbelts, rudder, rigging the wings, etc. I have a new attitude now, and I suspect I'll be able to fly it to the KR Gathering in September, although with no paint, wheel pants, or interior installed...
Update, as of April, 2005 - I've discovered the secret to making the Tiny Tach work! I wrapped the spiraled red pickup wire with aluminum foil, and then grounded that mess to the engine with a short length of wire. That instantly corrected the readings. Jim Faughn later did the same thing to his at the 2004 Gathering and it fixed his too. The cure for the EIS tach turned out to be a 27k ohm resistor rather than the 1k called out in the instructions, along with a P/R setting of 3. I determined the value of the resistor by temporarily installing a potentiometer and dialing it until the EIS tach agreed with the Tiny Tach (which also matches perfectly with a $40 optical tach that I bought at the RC model shop). It worked in a range from 18k to 46k, so the commonly available 27k resistor was fine. So I now know that I can turn my Sterba prop, which is a 54x66 that's been repitched to something less, at 3000 rpm static.
For more info on this engine, see my Corvair engine website.
Return to Mark Langford's KR2S