Ascaris and Complex Impedances.

[Ascari32]

Hope this makes sense:-


Much of my time spent on the exhaust system was directed at making the Ascaris work.

Now I accept, they won't! But why?

Well, since the discs were renewed and the Alfa boxes went back on, time has been wholly devoted to making them work.

It seems unfortunate that I chose not to devote more time to them before embarking on fitting sports boxes. Looking back, it's now clear, I too was sucked in by all the hype surrounding cat - backs, etc, etc. Certainly, I wanted the exhaust to be as free flowing as possible and came to believe "What was said on the Tin", was factually correct. Not so!

Where the problems lie with the Ascaris are, they have a "Complex Impedance", an absorption section, followed by a central resonator cavity and then a second absorption section before the tail pipe. To complicate matters further, the perforated pipe through the input section, leaves it and directs the gases across the central cavity, towards the perforated pipe through the output section. But they are not on the same axis, being "Off-set" by about 1/3rd to 1/2 their diameter.

It is difficult to explain the consequences of this aspect, but in essence; both acoustically, and in gas flow terms, this introduces an element of non-linearity.

Above ~ 3000 rpm, the gases are flowing too quickly across the central cavity to "Excite" it. So characteristically it is as if it were not there and only the inlet and outlet absorption elements are having an effect upon the nature of the gases. Importantly, they have a "Linear" effect upon them. And as such, the exhaust sound is great, given its character becomes very much related to the engines rpm.

However, below ~3000 rpm, the nonlinearity of the Ascari's central resonator, has a nonlinear effect upon both the gas flow and the acoustic wave.

Pause for a moment.

Neglecting the restrictive nature of the Alfa Manifold Cats, their central Cat and their Post - Cat Resonator, the standard valve timing of the Alfa camshafts, at low revs, virtually isolate the inlet tract from the exhaust system, by virtue of the exhaust valve closing at 9deg. atdc and the inlet opening at 11.5deg. atdc.

This stymieing of the air flow; the "Alfa Romeo Staccato Effect", means the MAF has a lot of work to do, in a short space of time to generate an output voltage for the ECU to calculate the required fuel, particularly given the piston in now 11.5deg. into  its induction stroke.

So the MAF element must have a sharp rise in voltage/Kg. flow rate. And because the influence of exhaust gas flow has been checked by the NVO, the MAF output voltage can be; with the fast response time of the standard MAF, pretty accurate.

Pause for a second moment.

However, with the changes made to my engine, three more important factors come into play.

1) The high flow exhaust manifolds.

2) The high flow Sports Cat and their Siamese inputs from the front and rear banks: hitherto from the output of the Alfa Twin Cat. Plus the Stainless Steel high flow Centre Section.

3) The 23.5degree PVO of the C.B. Cams. Only their duration has increased, the lift remains the same.

Pause for a third moment.

The Autodelta Manifolds, the Supersprint Sports Cat and the Centre Section are "Linear" - they have a substantially neutral effect on the exhaust gases, regardless of whether the engine is turning at 750 rpm or 7000 rpm, they appear "Dynamically Constant".

The combination of the manifolds, the cat, the central section and the Colombo Bariani PVO, essentially creates a transparency between the inlet tract, where the MAF sits, and the tail boxes.

So attempting to combine a "Nonlinear Element" with a device which has a very fast changing rate; ie the Ascaris back boxes with a MAF, with a very fast changing rate of output voltage for a small change in flow rate, is extremely difficult and virtually impossible to ensure constant results.

These are the principal reasons why it was essential to replace the Alfa MAF with a device which has a less acute Voltage out/Airflow change: (dy/dt) characteristic. Even with a Porsche 055 element it is still too acute for the exhaust system, a principle effect the CB cams create. However, this element can be "Desensitised - the dy/dt slope cannot be reduced except by internal re-alignment, but it's "Turn - on" threshold moved.

However, despite the Transfer Characteristic of the 055 insert, and four other brand new devices - all of which have their responses published in the Bosch Data Sheets, I have found it impossible to align the exhaust system/engine characteristics with the Ascari back boxes. It seems the absorption elements of the Ascaris are too small to counter the reactive nature of the nonlinear cavity at their centre.

They may well work with valve timing such as Alfa's standard 3.2 JTS, but they cannot work under the conditions my modifications have created.

I referred to this in some of the very first discussions I had with numerous exhaust specialists - which to a man, they failed to acknowledge

The Alfa boxes however, can and do! Being absorptive and certainly more restrictive, the reduce any reactive components from the system, creating a predictable response across the RPM Range.

It was an unusual conundrum! It was only by studying the apparently lack of performance of the Alfa's, that the failings of the Ascaris came into focus.


Modify message

[Colin Edwards]

Likening the exhaust to an RLC circuit is curious.  I guess the exhaust pressure pulses and resultant sound waves will act similarly to AC. 
Given the range of frequencies and odd harmonics in the gas flow, the exhaust system will behave somewhat like a tuned circuit. 
One thing most aftermarket exhausts appear to lack is an infinite baffle.  The standard 159 V6 exhaust features these.  Given the relatively flat exhaust frequency response, these baffles seem to be doing their job. 
Always impresses me the work a manufacturer (like Alfa and Fiat/Abarth) put into the sonic performance of the system.  Possibly something beyond the capability of the aftermarket brigade?
Maybe the Ascaris need the equivalent of a sonic inerter?

[Ascari32]

At resonance, voltage and current are in - phase and any tuned device looks resistive. This principal can be applied to a multitude of problems. When "Matched", and waves are in phase the only element to consider is the resistive element. At such a point, the return loss figure is high, quoted in terms of "db's" and ideally no power is reflected back to the source.

Away from resonance, reflections increase and the return loss figure reduces. These can take the form of "Standing - Waves" and invariably cause increased power to be absorbed by the source. It was at this point my exhaust system was vibrating badly all the way back to the Sports Cat.

To maintain a good match across a wide spectrum, a device needs to have a bandwidth to match that spectrum. A single cavity/tuned circuit has a high "Q" - magnification factor, and a narrow bandwidth.

So exhaust systems, such as those for performance engines need to be complex, if they are going to be "Tuned". Alternatively, with simple devices such as the Ascaris, a resistive (absorbtion) element needs to be added to reduce the "Q", such that it's impedance becomes more linear across the range.

Clearly the Magnification factor of the Ascaris is to high, hence the resonating at low rpm. To combat this, the absorbtion (Resistive) elements at the input and output need to be much greater. So yes, adding an element, such as Colin suggests would work. At resonance, the "Q" of a tuned circuit is expressed as L/CR. With the Ascaris, the L & C components are fixed; by the physical volume and shape of the cavity/resonator, so (R) would need to be substantial, to lower the "Q". But the losses would increase, far greater than if a properly designed absorption muffler were used. And it would still suffer due to its resonant frequency, albeit at a reduced level.

Before the Ascaris were fitted, they lay on the rug in front of our hearth and I spent some days inspecting them. Internally, the perforated pipes through the input and output muffler section, abruptly end at the dividing  walls of the central resonator - that is why it is a "Resonator", and not a "Helmholtz". Had they protruded into the resonator box, unperforated and by a calculated length, the cavity would have worked as a "Harmonic - Trap"; a Helmholtz, and the Ascaris would have been characteristically very different.

[Ascari32]

Further to my original post, more information has come to light.

Having concluded there is an inextricable link between the MAF element and the complex nature of the Ascaris, due to the transparency the Colombo Bariani camshafts have created, it seemed I must look else where for a solution - if indeed one could be found.

Returning for a second to the nature of "Tuned Circuits/Devices", although they can be tuned for linearity across a certain spectrum (Bandwidth), they can; and do, go nonlinear when overdriven. In doing so, they go into "Saturation". This is where the L, C and R elements are no longer able to cope with the signal level: in classic terms the output signal no longer represents that of the input x the "Q" - magnification factor).

However, during the period where saturation occurs, the excess energy manifests itself as distortion - harmonic distortion. Turn the input level down and the device resumes linearity and the distortion disappears.

It is this distortion, due to excess input level which is one of the causes of the Ascari "Break - up". Thus, the Ascaris are not wholly to blame for my problems.

I accept, there is a need to have the fuel maps adjusted but for the time being that must wait. Any issue with the ECU maps must be reduced to being in specific and defined areas. In not doing so would result in a real dogs breakfast as softies chase ghosts and consume hours and £££££....... in development costs.

Further research into the 3.2 JTS engine confirms the ECU does not, in a conventional sense enrich the AFR on startup, when the engine is cold. However, advanced exhaust timing and delayed inlet timing effectively starves the engine of air whilst maintaining the quantity of fuel injected.

A combination of factors govern this and one in particular is the NTC Sensor, located within the head, close to the exhaust port of cylinder two. It is this element, when quickly warming up after start up, which the ECU "Reads"', and begins to "RETARD" the exhaust camshafts.

In retarding the exhaust camshafts, the piston travels further down the bore, the combustion process is more complete, the exhaust gas temperature is reduced and so to is its pressure.

It is therefore obvious the response time and the accuracy of the NTC Sensors resistive Characteristic is imperative for ECU to set exhaust valve timing for - in particular startup and low speed operation.

My NTC Sensor was found to be faulty. The additional consequences of this was the unburnt fuel being discharged into the exhaust system resulted in excessive "Burning" within the exhaust system proper, forcing the Ascaris into "Nonlinearity", ie, they were exceeding their power handling characteristics and therefore producing both excess noise and extraneous harmonics.

This I have discovered to be true of the Alfa Boxes. However, being absorption types, their characteristic is "Attenuative", and thus muffle this effect in a much more linear manner - until of course the power level becomes so excessive they too go non linear.

Running with the Alfa Boxes, the exhaust sounds incredibly sweet and pure, provoking me to drive more often with the windows down, just to listen to it. It is in every sense of the word, an "Alfa" sound beguiling, even at 74 years of age.

I am still not totally convinced the Ascaris will now work. However, I have had stainless steel plates, puddle welded along the top and bottom, from input to output, to strengthen them; the central cavity in particular, and will be refitting them in the next week or two.

Aside from the effect the plates have; and I do hope they don't end up with a drone, replacing the NTC Sensor has made a vast difference to the engine and she is now happily running with a 3.8/4.0L Porsche 911 Turbo MAF element. It however has a 270 ohm resistor in series with its output to the ECU. This just tempers the startup noise a little further. However, it does also take a little off the acceleration at the bottom end - only a little because she is still quick.

Take the resistor out however, and she becomes "R"eally "S"avage!!!!!

Ultimately however, although I miss the sound of the Ascaris over 3,000 rpm: I hope the plates haven't changed that, I believe the only solution to be Supersprint or Ragazzon back boxes.