Sure, the intake to the airbox wasn't causing all that much of a flow restriction on the Audi. You could say that the negative boost lurking in the airbox intake snorkel was only a medium size critter - nothin' as big as the bugger discovered in the airflow meter, for example..... In fact, of the flow restriction of the standard intake, the bit before the airbox was responsible for only 28 per cent of the total. Still, if it could be got rid of...
In fact, what if it could be more than got rid of? What if I could turn that negative pressure nasty into a positive pressure delight?
Positive Intake Pressure
The trouble with 'ram-air' systems on road cars is that often they aren't. Ram-air, that is. By the time you collect enough air and shove it into the airbox, negative pressures lurking around corners in the intake duct and at its entrance have probably assassinated the pressure gained by the forward movement. Plus, how d'you know that the place you're trying to pick air up from is under positive pressure anyway?
Weeeelllll, you use your DMDPG, doncha? Aaaahh - forgotten that abbreviation again? Well your DMDPC is yer Dwyer Magnehelic Differential Pressure Gauge, which can be used to measure tiny positive pressures as well as negatives. If you want, you can use a manometer to measure these pressure build-ups as well - but not the single column one we used a few issues ago. Instead you've gotta make a U-shaped one, which will measure both negative and positive pressures. OK, but since I've got a DMDPG, I used it. Used it for what? To find areas of positive pressure close to the airbox, that's what. In fact, I only had to try one place to locate what I was looking for.
In front and below the airbox is a standard engine oil cooler. Like most of the bodywork of the Audi, this area shows intelligent aero design - the cooler's fed by a specific duct that has a wide-open mouth positioned behind a grille in the bumper. You'd expect there to be a positive pressure in front of the cooler. Why? Because as the air is pushed into the duct, it builds up in front of the cooler - creating a pressure which perhaps could be made use of to feed the airbox. But first, did reality match the theory? I placed the pressure probe in the duct in front of the oil cooler and ran the other end of the tube to the positive pressure port on the DMDPG. At 80 km/h I was reading a pressure of 2 inches of water, while at 100 km/h this had risen to 3 inches of water. Yee ha!
Hmmmm, if I could get a positive pressure of 3 inches of water in the airbox, that more than compensates for the measured pressure drop of the standard filter!
(Incidentally, if you're saying to yourself - "Hey! That's pretty good! He's measuring air pressure variations over the surface of the car!" - I'm glad you think so. I was so excited I started ringing up people to tell them about it. And if it does something for you, maybe it's worthwhile revisiting the story where I introduced the Dwyer Magnehelic Differential Pressure Gauge. For the dollars, it's a stunning instrument for finding power - and improving aerodynamics, too.)
There's no such thing as too big a duct joining the atmosphere to your airbox, ie huger = better. So when I selected the PVC pipe from which to make the new duct, I went b-i-g. Bloody big. In fact how does a 10.5cm (over 4 inches) internal diameter grab you? It should - with a cross-sectional area of 86.5 cm square it's no less than twice as big as the 7.4cm standard intake. This is about the largest PVC pipe for which local hardware stores and the like stock both the pipe and pre-formed bends. It is also quite thick-walled, which makes manipulating the pipe a bit easier.
Manipulation? What manipulation? I'll get to that in a moment.
The standard airbox opening was of course much too small to take the new monsta. So using a knife heated in a gas flame, I cut out a hole in the airbox that matched the size of the new duct. The ragged edge that resulted was sanded smooth, then the raw edge was discreetly covered in rubber edging.
The end of the new PVC pipe was cut off at a slight angle (the pipe didn't go into the airbox at exactly right-angles) then a heat gun was used to soften a small area about halfway along the pipe. Using a round metal can as a rolling pin, a depression was then formed in one wall of the pipe - this allowed the pipe to snuggle up against the bracket that holds the oil cooler. If you make these sorts of depressions with care, the internal contours stay smooth with only very gentle curves being formed - better for flow while still allowing the duct to weave around obstacles.
At the front of the pipe a bend down towards the roof of the oil cooler feed duct was needed. I used a pre-formed 90-degree bend, although I probably could have got away with a 45 degree-er. However, the 45-degree bend wasn't as well made internally - the short radius had a sharp change of direction while on the 90-degree bend, a smoothly curved inner was used.
To allow the bend to reach the roof of the oil cooler duct, the 90-degree bend was cut off at an angle, giving an oval-shaped opening. The edges of the cut-off section were then heated, one section at a time, and the food can used to roll these back, creating a large oval-shaped bell-mouth. The edges of the bellmouth were then sanded with various grades of paper until they were smooth.
A suitable opening was cut in the upper part of the plastic oil cooler feed duct, and the bellmouth could slyly peer through.
Note that in this design, the mouth of the intake duct does not face forward. I'd already proved that the area of the oil cooler duct was subjected to positive pressure, and any other duct sealed to it will sense that pressure. Facing the mouth of the airbox duct downwards reduces the chance of breathing an airborne rock flicked up by another car - something that has the potential to cause puncture problems when a paper air filter is retained in the airbox.
The duct was painted black with a spray can, then held in place by a single bolt and nut passing through a hole in the oil cooler support bracket. When using nuts and bolts on the intake snorkel, make sure that you use two nuts on the bolt, tightened against each other so that there's no possibility of the bolt coming off and floating around inside the intake... For the same reason, don't place the nut on the inside of the duct - always have the bolt facing outwards.
With the bumper back on, the new intake duct is invisible.
So howdit work? You want the short answer as to what happened to the 9 inches of water pressure drop previously recorded in the airbox before the filter? OK I'll give it to you straight: the new intake duct got rid of it all! Yep, under full power in second gear, the negative pressure nasty was banished to complete and utter oblivion. At 2000, 3000, 4000, 5000, 6000 and 7000 rpm, the pressure recorded in the airbox was just the same as atmospheric.
To make it real clear, here's the before/after pressure drop figures for the airbox, measured on the atmosphere side of the air filter:
|Standard Airbox Snorkel Pressure Drop
(inches of water)
|New Airbox Intake Snorkel Pressure Drop
(inches of water)
Wow! I just got rid of 25 per cent of the original intake system's pressure drop. Another hairy, smelly, repugnant negative boost banished!
And even better than that, at anything less than full throttle, a measurable positive pressure could be seen in the airbox. So we're talking a zero flow restriction at full throttle, and a positive pressure at anything less than full throttle. And how much positive pressure? Up to 9 inches of water at high cruising speeds like 140 km/h!
Aaaa-haaa - so what was the performance now like? Well, if this were a typical car magazine tech story sponsored by a workshop, at this point the story'd say something like. "And the performance gain was just as you'd expect - staggering. Make sure that you, too, go to ABC Auto, etc." However, this isn't that sorta story - and there was a problem. A couple of problems, actually.
I'd noticed previously that the Audi - running the very high factory compression of 9.3:1 together with a standard 1 Bar of turbo boost - would give slight jerks (little hesitations) when driven at full throttle on our so-so (96 octane PULP) fuel. I figured the ECU was pulling back timing (and/or small amounts of boost) when the twin, cylinder-specific knock sensors detected detonation. And that had been confirmed when I'd previously put 98 octane fuel in the car - the stutter had gone away completely. However, until the very week of which I am writing, 98 octane fuel hadn't been available in South Australia... but now it was.
And the relevance? With the major improvement in flow through the intake, the stutters started to affect the car badly. The measuring accelerometer could be seen fluctuating in level, and the car felt slow - well, relatively anyway. Of course, if you increase cylinder filling you'd expect a car on the edge of detonation to detonate; and that was what it appeared was happening.
The graph tells the story - the engine comes up on boost harder, but from 4000 rpm onwards the acceleration (and so power) is actually down. So, reluctantly, I filled the tank with 98-octane. 'Course, if I wasn't writing this story I'd have had no reluctance at all - but I hate changing two factors simultaneously when doing testing. But this time there appeared to be no other course of action that made more sense.
And with the better brew in the tank? Yes, it was better - but then I noticed that boost was always lower! Aaaahhh! The ECU appeared to be dropping boost to keep the power the same, ignoring the benefits that could be had from the new high-flow intake system and better fuel! I was going around in circles while the Audi ECU thumbed its nose at me.
What to do now, then? Hmmmm, before making any more intake mods, best to develop a new boost control that does what I want it to do, eh? Changing two factors? Nah, we'll change three.... And then, when boost is doing what it should, then I can go back to making the remaining intake modifications.
- The new duct to the airbox reduced the full-load intake snorkel pressure drop to zero.
- At less than full load, a positive pressure is generated in the airbox.
- Power did not respond to the better intake flow, with initially knock-triggered ignition retard then falling ECU-controlled boost the suspected culprits.
Next week: Developing a brand new boost control system. It incorporates adjustable control of wastegate creep as well as giving adjustable max boost. And it sure as hell changed the Audi's performance - without even lifting peak boost! It is a system also suitable for all turbo cars.
Eliminating Negative Boost - Part 1
Eliminating Negative Boost - Part 2
Eliminating Negative Boost - Part 3
Eliminating Negative Boost - Part 4