27 replies to this topic

[V8 Fireworks]

As above, just wondering what are the advantages of dual or triple exhausts (.e., a set of  3 or 2 cylinders having their own header which does not merge with the remaining cylinders) on a bank of six cylinders.  I presume the same would apply for a standalone inline six.

 

I can't imagine they would be willingly giving away scavenging performance, so I take it this arrangement performs better in these applications. Is it something to do with the high rpm and small displacement per cylinder of these applications?  

 

Matra, Triple exhaust per bank

 

Ferrari, Dual exhaust per bank

Edited by V8 Fireworks, 10 February 2018 - 11:07.

[MatsNorway]

Great question, i have been thinking about that before. Could it be because the pressure difference is higher with bigger periods of time pr. pulse?

 

Combine that with lack of serious flowbenches, too big pipes, lack of knowledge an lack of data simulations and you struggle to do it at higher rpms with multiple pulses. You end up with a finer/smaller tuning window.

 

i state big pipes because the speed of the pulse contains energy, so by maintaining that energy you can use the momentum to draw more exhaust gases out of the other pipes. At least that was what i got from one engine builder that was very focused on the flow over the entire pipe, he would go down in size further downstream to keep the exhaust gases flowing at the same speed.  That part i do not understand.

Edited by MatsNorway, 11 February 2018 - 17:01.

[Bloggsworth]

In Matra's case, it was because they thought their engines too quiet... Seriously, in over 100 years, I don't think there is any agreement on the benefits of either, as a layman, I would think it about the difference between peak power and the spread of torque/power. BRM won the World Championship with 8 individual stacks - Cause or coincidence?

[MatsNorway]

In BRMs case it probably was a blown engine or coincidence

Edited by MatsNorway, 10 February 2018 - 20:12.

[Charlieman]

This was a question for which bikers once had an answer -- about twins, threes and fours. And for the V8 1.5 litre F1 cars of the 1960s, there was the question of crankshaft angle and ignition sequence.

 

For a 2018 V6 in F1, packaging of the engine and exhaust system is probably more important than 0.5% max power. 

[Charlieman]

At least that was what i got from one engine builder that was very focused on the flow over the entire pipe, he would go down in size further downstream to keep a constant flow.  That part i do not understand.

Is it possible that something has been lost in comprehension and translation? Constant flow? Steady flow? Laminar?

 

Owing to heat loss through the walls of a pipe, you might reduce the internal diameter -- in an attempt to maintain energy flow. I'm just trying to apply Bernoulli principles.

[MatsNorway]

Edited to: keep the exhaust gases flowing at the same speed.

[Greg Locock]

As the exhaust gas cools the velocity drops, if the pipe area is constant.

[MatsNorway]

Sure, but why is that a problem? high flow also means more resistance surely. Venting into a big pipe would be more like venting into the free air.

[Kelpiecross]

In Matra's case, it was because they thought their engines too quiet... Seriously, in over 100 years, I don't think there is any agreement on the benefits of either, as a layman, I would think it about the difference between peak power and the spread of torque/power. BRM won the World Championship with 8 individual stacks - Cause or coincidence?

From memory - I don't think they ran the 8-stack exhaust in many races. Also from memory - the car was said to sound like lions roaring in the distance.

[Tim Murray]

Yes indeed - BRM used the individual ‘stackpipe’ exhausts only in the early-season non-championship races and the first two championship GPs, but the stackpipes had a habit of breaking and dropping off. From the Belgian GP onward they used a low-level twin tailpipe system.

[john winfield]

I'm not sure it answers the original question but there's an interesting piece by DSJ in the August 1979 Motor Sport. I used their archive search feature and there are some others over the years too. Here's the 1979 one:

 

https://www.motorspo...ne-trend-design

[Fat Boy]

Considering the car the OP pictured had very little scientific work done with respect to any aspect, you can almost certainly conclude the exhaust configuration was chosen for aesthetic reasons. This could be visual or aural.

[fredeuce]

Considering the car the OP pictured had very little scientific work done with respect to any aspect, you can almost certainly conclude the exhaust configuration was chosen for aesthetic reasons. This could be visual or aural.

It did read somewhere that the sound produced by this engine was a goal in the design brief. Little doubt they achieved this.

 

[Fat Boy]

It did read somewhere that the sound produced by this engine was a goal in the design brief. Little doubt they achieved this.

 

It's a shame this is neglected presently. Who gives a **** if the Matra was fast. Here we are talking about it 50 years later!

[blueprint2002]

I am new to this forum, which is why I have only just seen this thread, more than a year late. But assuming that most, if not all, of those who participated, are still active members, I would like to add something more, which they might find interesting.

What I find really different is the exhaust system of that V12 Matra; I don’t know of any other like it.

The others, whether straight-6, V12 or Flat-12, all follow much the same system: Headers from cylinders 1, 2 and 3 blend into one pipe, and the same with cylinders 4, 5 and 6. (If you regard each bank of the 12s as an independent straight-6). This may be simply explained as follows:

While the theoretical four-stroke cycle opens the exhaust valve for 180 degrees only, (during the exhaust stroke), in actual practice the valve remains open for anything between 210 and 240 degrees; opening before the power stroke ends, and closing after the induction stroke starts. Clearly, the entire pipe, from exhaust valve right up to the end, is pressurised to some extent during this period (the highest pressure being when the valve opens, falling rapidly thereafter, down to atmospheric at some point in time).

For reasons of good balance (for which the straight-6 is justly renowned), the firing order is always 1-5-3-6-2-4, with the crank throws arranged successively at 120 degrees, (to smooth out the torque as much as possible). This ensures that the firing interval between any two successive cylinders is 120 degrees of crankshaft rotation.

From the firing order above, it is clear that cylinders 1-3-2, in that order, are separated by 240 degrees each, so too cylinders 5-6-4. And now it should also be clear that that the usual arrangement of the headers (as described above) ensures that each tailpipe carries the exhaust from one cylinder only, at any time. No interference of one exhaust pulse with another, and hence free flow at all times, with only the resistance of the piping to overcome. Hence the smallest possible loss of power from this cause.

What about the Matra? They were deadly serious competitors, just like everyone else, so it is probably only legend that their six-times-two-into-one system was intended to produce the best possible sound. There must have been sound technical logic at work, (maybe known only to the designer), even if it proved fruitless in the end. Nothing unusual about that outcome in engineering, for countless reasons.

It might help if anyone knows what the exact arrangement was. Which is to say, which two cylinders were connected to each of those three tail pipes, on either bank?    

[Joe Bosworth]

bp2002 has provided the best answer in one hit. 

 

I proved it my purposes back in the 1950s when I developed a very potent for the day straight six engine capable of winning a class drag racing championship before such were formally proclaimed.  I had worked hard at getting a camshaft/header setup that seriously complemented one another.

 

The Bible of exhaust system design was written in 1970, "Scientific design of exhaust and intake systems" by Philip H Smith. His verbage commencing for six cylinder systems starts on page 146.

 

Interestingly the modern day practice has not identified that the science of gas flow has not altered jn the meantime.

 

Regards  

[Joe Bosworth]

Carrying on from my post of yesterday I might add:

 

Providing balance to a straight 8 engine requires that the firing order requires alternating power strokes between the front and rear sets of cylinders. Pairing the front three cylinders together to one primary junction box and then ditto for the rear three results in each set having just one cylinder firing every 240 degrees.  One keeps each set of three going into one exhaust pipe which is kept separate from the other three and exiting the body/chassis at a very convenient design length.  The length of primary pipe from top of each piston at exhaust opening to junction box likewise comes at a very convenient design length.

 

I might add that depending on one's choice/requirement of muffler you end up with a very pleasant even pulsed exhaust note.

 

One can go back and look at every straight six racing engine starting with C/D Jags and you can see that everyone ends up with the same solution and measuring lengths you can calculate the engine revs that each application is tuned for.

 

Regards again

[blueprint2002]

bp2002 has provided the best answer in one hit. 

 

I proved it my purposes back in the 1950s when I developed a very potent for the day straight six engine capable of winning a class drag racing championship before such were formally proclaimed.  I had worked hard at getting a camshaft/header setup that seriously complemented one another.

 

The Bible of exhaust system design was written in 1970, "Scientific design of exhaust and intake systems" by Philip H Smith. His verbage commencing for six cylinder systems starts on page 146.

 

Interestingly the modern day practice has not identified that the science of gas flow has not altered jn the meantime.

 

Regards  

 

Carrying on from my post of yesterday I might add:

 

Providing balance to a straight 8 engine requires that the firing order requires alternating power strokes between the front and rear sets of cylinders. Pairing the front three cylinders together to one primary junction box and then ditto for the rear three results in each set having just one cylinder firing every 240 degrees.  One keeps each set of three going into one exhaust pipe which is kept separate from the other three and exiting the body/chassis at a very convenient design length.  The length of primary pipe from top of each piston at exhaust opening to junction box likewise comes at a very convenient design length.

 

I might add that depending on one's choice/requirement of muffler you end up with a very pleasant even pulsed exhaust note.

 

One can go back and look at every straight six racing engine starting with C/D Jags and you can see that everyone ends up with the same solution and measuring lengths you can calculate the engine revs that each application is tuned for.

 

Regards again

 

 

Thanks JB! I hadn’t realised that the C-Type Jaguar was perhaps the first six to be fitted with the sort of exhaust system we are discussing, but it’s not surprising as methodical development seems only to have started post WW2. And widespread use of this principle had to wait until the early sixties, as far as I can tell. See, for example, Motor Sport of February 1957, with a photo feature of GP exhaust systems; the Maserati 250F is nearly there, and the BRM might be, but the picture doesn’t show it too clearly. None of the others is even close.

This discussion prompted me to look again at the references I have, and it seems that the time the Matra V12 appeared was a kind of “golden age” of the V12. Between 1966 and 1971, no less than six different engines competed in GP racing: BRM, Cooper, Eagle, Ferrari, Honda and Matra. Before they were all swept away by the relentless logic of the DFV, leaving only Ferrari.

And looking again has reminded me that the Honda RA273/300/301 engine was another exception to the usual V12 exhaust system. While the convolutions are tight and closely packed, so rather difficult to tell apart, it does appear that cylinders 1, 4 and 5 blend into one pipe, and cylinders 2, 3 and 6 into the other.

This probably has something to do with the fact that this engine, with a 90-degree angle between the cylinder banks, differs radically from the usual design. If the crankshaft were of the normal 120-degree six-cylinder type, firing intervals would be 90-30-90-30 instead of the usual 60-60-60-60. This would change the torsional vibration characteristics, which might necessitate a different crank arrangement, hence changing the firing order. And so leading to a different exhaust system. A drawing or picture of the crankshaft would be useful, but I’ve never come across one.

[Wuzak]

 

And looking again has reminded me that the Honda RA273/300/301 engine was another exception to the usual V12 exhaust system. While the convolutions are tight and closely packed, so rather difficult to tell apart, it does appear that cylinders 1, 4 and 5 blend into one pipe, and cylinders 2, 3 and 6 into the other.

This probably has something to do with the fact that this engine, with a 90-degree angle between the cylinder banks, differs radically from the usual design. If the crankshaft were of the normal 120-degree six-cylinder type, firing intervals would be 90-30-90-30 instead of the usual 60-60-60-60. This would change the torsional vibration characteristics, which might necessitate a different crank arrangement, hence changing the firing order. And so leading to a different exhaust system. A drawing or picture of the crankshaft would be useful, but I’ve never come across one.

 

 

If the Honda was to have even firing intervals with a 90° bank angle then it would, surely, require a 12 throw crank, or at least split big end journals?

 

Not sure if that is desirable in a racing engine.

 

Though i believe the Alfa flat 12 had a 12 throw crankshaft at one stage.

[blueprint2002]

If the Honda was to have even firing intervals with a 90° bank angle then it would, surely, require a 12 throw crank, or at least split big end journals?

 

Not sure if that is desirable in a racing engine.

 

Though i believe the Alfa flat 12 had a 12 throw crankshaft at one stage.

I think you are right about that; something similar has been done with production 90-degree V6s by some manufacturers.

Don't know if that is what Honda did with these engines, but they were considered at the time to be excessively large and heavy, so perhaps that is what was done.

I understand that the Flat-8 Porsche 908 engine had an 8-throw crankshaft, and it was fairly successful, so it's not impossible to get away with such a solution to the design problem, even in a racing car. Not sure about the Alfa, will check such references as I have.

[blueprint2002]

On second thoughts, the Porsche 908, being air-cooled, necessarily had a wider cylinder spacing, which naturally increased crankshaft length anyway. So to fit eight pairs of cranks into that space, together with the necessary main bearings, was surely easier than it would have been with water cooling. 

[Lee Nicolle]

Having raced 6 cyl tintops and speedway for 50 years my experience is that it is too easy to go too big on a header pipe. And that off the shelf headers will make better driveability by being split into two tailppes. Though peak power is often less. But I would prefer to have good smooth power between 3500 and 6500 than 5hp more at 6500 with less under 6000. More so on speedway where the engine works between say 4000 and 6500+

On pukka race engines it is probably quite different, though many of those have a very narrow power band and not a great deal of torque either.

[Lee Nicolle]

 

On second thoughts, the Porsche 908, being air-cooled, necessarily had a wider cylinder spacing, which naturally increased crankshaft length anyway. So to fit eight pairs of cranks into that space, together with the necessary main bearings, was surely easier than it would have been with water cooling. 

 

There is a few split rod journal V8s around. Cannot remember what but the journals worried me. 

Formula Holden with split journal V6 Buicks [or Holdens]  broke quite a few cranks early on and I believe went to earlier Buick  straight journal cranks. Not allowed in speedway Modified Sedans so they break cranks!! As do the street stock engines and wingless Sprint engines. And they are basically stock.

[Fat Boy]

Having raced 6 cyl tintops and speedway for 50 years my experience is that it is too easy to go too big on a header pipe. And that off the shelf headers will make better driveability by being split into two tailppes. Though peak power is often less. But I would prefer to have good smooth power between 3500 and 6500 than 5hp more at 6500 with less under 6000. More so on speedway where the engine works between say 4000 and 6500+

On pukka race engines it is probably quite different, though many of those have a very narrow power band and not a great deal of torque either.

 

You just hit on the compromise which produces all the head-scratching when designing an exhaust. Greater signalling (more connections, tri-y) between cylinders generally means a wider powerband at the expense of peak power. Reduced signalling (zoomies) means less back-pressure, which may gain peak output, but at the expense of power through the rev range. It usually comes down to "horses for courses".

[blueprint2002]

There is a few split rod journal V8s around. Cannot remember what but the journals worried me. 

Formula Holden with split journal V6 Buicks [or Holdens]  broke quite a few cranks early on and I believe went to earlier Buick  straight journal cranks. Not allowed in speedway Modified Sedans so they break cranks!! As do the street stock engines and wingless Sprint engines. And they are basically stock.

Took me a while to find what this reminded me of: the later version(s) of the BRM H16 actually had split journal crankshafts, but it was hardly what you would call a success. On The Yorkshire Ferret's site, there is a good description of the evolution of that engine with a picture of the original four-throw crankshaft together with the later split-journal successor.

But it is also intriguing to hear of a V8 with a split-journal crankshaft. Could it have been one of the very few which have been built with a cylinder bank angle other than 90 degrees? Those I know of are:

Miller (for the 4WD Indy car): 45 deg.

Honda RA302: 120 deg

Ford HB and Zetec-R: 75 deg

Ring a bell?

[Kelpiecross]

As above, just wondering what are the advantages of dual or triple exhausts (.e., a set of  3 or 2 cylinders having their own header which does not merge with the remaining cylinders) on a bank of six cylinders.  I presume the same would apply for a standalone inline six.

 

I can't imagine they would be willingly giving away scavenging performance, so I take it this arrangement performs better in these applications. Is it something to do with the high rpm and small displacement per cylinder of these applications?  

 

Matra, Triple exhaust per bank

 

Ferrari, Dual exhaust per bank

 

 

 

   That's not an exhaust system -  that's an exhaust system!   216 cylinders into  12 vertical pipes.

https://www.bing.com...&vt=0&eim=0,1,6

[7MGTEsup]

 

Ferrari, Dual exhaust per bank

 

I don't recall Ferrari F1 cars in the 90's having quad pipes like that, that must be something that has been made at a later date.