Yesterday my collection of exhaust parts grew with my latest shipment from Summit. I gave up on the local exhaust "specialists" when they wouldn't give me a quote for the complete system (only parts) and seemed a little under-informed about fabricating what I wanted (they didn't seem interested in the job at all). So I'm going to build the system (the parts were much cheaper from Summit, for example the magnaflow was 1/3 the price $54 vs $140) do a little learning and have some fun - I've always enjoyed welding stainless.

Here's what I have in the parts bin so far (all sizes 2.25" and stainless steel) - 4 - 180 degree J-bends, one large oval Magnaflow and one round (4") Magnaflow, 4 exhaust hangers (weld on) and some wide band style clamps. I'll have to get some straight pipe from the local shop, probably 6 feet total.

My plan is to mount the oval muffler before the axel (ala Dale's car) and have the round at the rear as a resonator. What I'm looking for is advice on where to break the system apart for install/removal and any other exhaust tips/tricks. My thought was to have 2 sections, one from the dual down to the rear of the oval muffler, the other going out the rear.

Never done a complete exhaust system before - help.

Craig

I think you're on the right track, Craig. For Dale's car, we welded the whole thing together into one piece. I can't see why he'd ever need to pull the exhaust, but if we have to we'll cut it apart and add a V-band clamp. I don't really trust the band type clamp yet, especially as a butt-joint clamp. To join a flared type lap joint, fine, but I'm still skeptical about the amount of torque it can take. The V-band clamp GRABS and is dead simple to break apart. Takes a bit more clearance and costs more, though.

Dale didn't want the tailpipe hanging below the body, and a round muffler definitely would hang below the body. Sure would quiet things up, I'm guessing.

IIRC, we used 4 u-bends and about 6-7 feet of pipe. Actually, the amount of pipe was probably closer to 4 feet: I think Dale bought two 7 foot lengths of pipe because he interpereted my instruction to get 10 feet as a necessity, and we still had about 3 feet left over from one length after we were done. I think almost all of the pipe went between the header and the first bend, just in front of the muffler.

Plan on installing an Oxygen sensor bung as well. Borrowing a wideband from someone will certainly aid in tuning your carbs, though it won't help with the carburator balance. Try to put the sensor bung pointing upwards at at least a 10 degree angle, and make sure it's a good 4-5 feet away from the exhaust valves. We've got Dale's pointing inboard and up, just behind the collector. Don't aim it TOO high up, because the O2 sensor is kind of long and the cable isn't supposed to be bent too sharply. You can plug it with a nut when you're not using the O2 sensor.

That's about it.

O2 sensor bung...never thought of that. I'll have to source one of them unless they are generic and I can find one locally. I looked into the V-bands and don't know why I didn't do them as well, but I'm sure it was cost. I'll give the bands a try and see how they work. My theory is that they have no, or at least not much, longitudinal stress and will work for holding pipe together that is properly supported.

Next weekend is the drop dead date for the exhaust (Deirdre is away - garage may even be cleared of house crap).

Many spark plugs use the same thread as an O2 sensor, so the trick is to use a spark plug "anti-foul" fitting, chopped down to suit the length required. You can get these "anti foul" fittings at NAPA or other places like that. It's basically a tube that threads into the head and spaces the spark plug away from the chamber, preventing spraying/splashing oil from fouling the plug. This is the method we used on Dale's exhaust, but I doubt that this piece is stainless. Dale's even got a spare, IIRC, so if you guys think of it I'm sure he wouldn't mind bringing it down to you this weekend.

Or you can try and find a stainless 18mm nut of the appropriate thread. Good luck doing that in a week - bolt supply houses almost certainly don't stock them but might be able to order one in for you, along with a bolt you can use as a plug later.

You could also pay way too much for an O2 sensor bung and plug from a hot rod shop.

Thanks Matt, I'll see if I can get an anti foul plug fitting thing. It may be a little ambitious to get it done in a weekend, but I'm going to give it a try. I've got some time before hand to get the bung fitted (I could even weld up a boss on the down and drill and tap a 18mm hole if I get stuck).

I was reading that you can use a multimeter and a one wire O2 sensor to set the mixture, instead of having to go to the expense of getting a separate tuning system. Apparently the correct mixture is around 400mV...I don't even want to think what I may be at now!

Sounds like you've been listening to a carb guy... One wire O2 sensors are only good for EFI systems, where they are constantly looking for that crossover voltage, and I'll tell you why:

A narrow band O2 sensor (one wire, three wire, four wire) is only accurate at saying whether there is an excess of oxygen in the exhaust or not. It has no way of telling you how much too much oxygen is in the exhaust, and it's incredibly unstable at the point of stoichiometry: where all the fuel is theoretically reacting to the exact amount of oxygen needed. It's very difficult to get a narrowband O2 sensor to read 0.400 volts, for example, because that's right on the lean edge of where it's very unstable. Some sensors may show 0.400V, some may go crazy at around 0.350V, or even lower. ALL narrowband O2 sensors are unstable for a range around 0.500 volts, and what the EFI systems are looking for are those swings over that point, and it adjusts richer and leaner in small increments to stay right around that mark.

It's cheap, and it's easy, but it really doesn't work. It'll tell you if you're rich or lean, but not much more than that.

Got it...If I look at the plot of A/F ratio and mV then this makes sense, it is essentially oscillating between two states. The mV change while measuring rich/lean is ok, but then essentially jumps between states...for example at 14.7:1 the slope is almost infinite! The Narrow band sensor is good for two states - rich or lean, there is no resolution at "just right". Ergo - wideband.

Damn this stuff is cool.