Port and Polish by Comiittee thread (Cool pics throughout)
01-20-2015, 02:11 PM
#241
Chronic Tool Dropper
Lifetime Rennlist
Member
Lifetime Rennlist
Member
Curious whether there are other valves (perhaps from other non-Porsche cars...) with the right diameter and stem length but already shaped with the flatter tulip under the head.
A couple decades ago I found some hi-po aftermarket Chrysler valves for a Rolls-Royce engine that were drop-in replacements at a tiny fraction of the R-R prices. Just needed to use the Chrysler keepers and retainers. There's a pretty amazing supply of valves in the aftermarket.
A couple decades ago I found some hi-po aftermarket Chrysler valves for a Rolls-Royce engine that were drop-in replacements at a tiny fraction of the R-R prices. Just needed to use the Chrysler keepers and retainers. There's a pretty amazing supply of valves in the aftermarket.
01-20-2015, 04:26 PM
#242
As this is the porting and polishing by committee thread, I have volunteered to some secretarial duties. Ake has been working on the stock S4 port and valves to see how much flow improvement one can generate without the expense of new seats and valves. This doesn't mean that the 39mm valves wouldn't be better for large high rpm engines, they would be. This is in the original spirit of the thread of what can be done when the heads are off without new, expensive components. And just to remind everyone, I've done none of the work, it's all Ake's efforts and expertise. I am just documenting.
2x37mm valve flow study
The below describes some beneficial modifications to the 928 S4 heads. The figures in the bottom of the post show three ways to interpret the test data from Ake's experiments with 37mm intake valves, to see how beneficial and under which conditions various modifications are.
There are three select experiments, all three evaluated in each of the three graphs. All experiments are on a 100mm bore and in comparable conditions.
The first experiment is labeled "STOCK INTAKE PORT 2x37mm, stock valves". This is as the head and the valves came from the factory.
The second experiment is labeled "BUDGET INTAKE PORT 2x37mm, stock valves". This is tested combination #16 in Ake's data set.
The only change is a five-angle valve job that gives a larger throat diameter which is then blended to the port. In particular, this experiment is run with unmodified stock 2x37mm intake valves. Valve seat ring ID increased to 33.3mm (90% of the valve size). Cross sectional area (CSA) at the throat per valve is 871 - 33 = 838 mm^2. Valve seat angle 45 degrees x width 0.8mm. Upstream the port 60 degrees x 1,2mm width and 75 degrees x 2,5mm. The upstream port matched to the increased seat ring ID. In the combustion chamber, 32.5 degrees x 1.2mm and 15 degrees x 0.5mm cuts.
Attachment 904331
Attachment 904332
The third experiment is labeled "BUDGET INTAKE PORT 2x37mm, radiused cuts, short-side and long side modified, modified valves". This is tested combination #26 in Ake's data set. Modified intake port with 2 x 37mm modified valves.
The starting point is the same valve job as in the second experiment (Ake's #16). The bottom two cuts are then radiused and rounded.
Attachment 904333
There are many opinions about radiusing intake valve job angles. The consensus among writers appears to be that it usually hurts with carb'd 2-valve engines and may hurt or help with fuel-injected 4-valve engines. How much and how far downstream to radius, if at all, depends on how important it is to reduce reversion for the engine in question. Leaving the sharp edges in supposedly reduces reversion. From the flow perspective, radiusing has at most a marginal effect.
Valves are modified. Nail-shaped flat valve, back angle of 8 degrees and back cut of 30 degrees. Valve back rounded and smoothed, the 45-degree valve seat in the valve reduced from over 2mm in the stock valve to 1.0mm in the modified valve. The valve weight goes down as a positive by-product.
Attachment 904334
In addition, the port small-side radius is moved 1mm closer to the seat. The purpose of this modification is to increase the short-side radius, allowing air to turn more gently. The modification is made between 10:30am and 2pm on the left hand side and between 10am and 1:30pm on the right hand side valve. This will increase the valve throat somewhat upstream of the blended cuts, the distance between the short side and long side is 34.3mm.
Attachment 904335
Attachment 904336
The above short-side radius optimization is in my opinion the most challenging part of the third experiment. There are conflicting requirements of keeping the cross-sectional area down, providing air a straight section before the seat to settle down, and increasing the short-side turn radius. Ake can probably give much more detail on this step if readers so desire.
The long-side radius is also modified by making the divided ports slightly wider and taller in the area where the valve guide is protruding.
Attachment 904337
In terms of numbers, here is some guidance to what the cross-sectional areas are supposed measure at after these modifications:
Here are the flow results from the three experiments:
Attachment 904338
The first experiment has no surprises, since it is the stock port.
The second experiment increases the throat-to-valve diameter ratio, which helps the high lift flow but at best doesn't lose any low-mid and mid-lift flow. Still, with big cams, the second experiment is already a big improvement. With small cams, not so much.
The third experiment involves a lot more time-consuming modifications. The flow results show that those are worth it. Basically, the third experiment takes the flow curve from the second experiment and shifts it up across the lift range. This is a very impressive result with 37mm valves. Flow rates of 350+ CFM at 28 inches of water stand comparison against most other ported 928 heads, regardless of the valve size.
In terms of port efficiency, the picture is similar. Recall that the below graph defines port efficiency as the ratio of actual flow to the predicted flow from a simple compound restriction formula that takes into account the valve curtain area and the throat area.
Attachment 904339
The stock port is very efficient at the 5-9mm lift range. This is mainly due to the very low 83% throat-to-valve diameter ratio. If one is limited to low lift cams, then one should modify Ake's valve job angles that result in a slightly smaller throat-to-valve diameter ratio.
The third experiment is more efficient everywhere than the second experiment. Again, the modifications are paying off.
The final graph is port velocity at its fastest point. This is not the velocity of the port in a running engine, instead it's the computed velocity at the flow bench. One could also directly measure velocity at some points with a pitot-static probe, but this is not that kind of velocity. It's simply the flow rate vs. the minimum cross-sectional area.
Attachment 904340
I am assuming above that all three ports still have 1400 mm^2 MCSA upstream in the port. This estimate will likely be revised in the future with the benefit of better measurements. Downstream I am accounting for the seat ring area and the valve curtain area.
As one would guess based on the efficiency graphs, the stock port has high velocity at the minimum cross-sectional area (MCSA) at 5-9mm lifts. Also as expected, the 90% throat valve job loses to the stock port in velocity below 9mm and beats it above. Impressively, the third experiment matches the stock port in velocity at 6-7mm lifts and beats it rather clearly both below and above that lift range.
My personal conclusion from Ake's study: These experiments show that one can improve the 928 4-valve ports with some thoughtful modifications expertly implemented with good tools, without having to buy expensive trick parts.
2x37mm valve flow study
The below describes some beneficial modifications to the 928 S4 heads. The figures in the bottom of the post show three ways to interpret the test data from Ake's experiments with 37mm intake valves, to see how beneficial and under which conditions various modifications are.
There are three select experiments, all three evaluated in each of the three graphs. All experiments are on a 100mm bore and in comparable conditions.
The first experiment is labeled "STOCK INTAKE PORT 2x37mm, stock valves". This is as the head and the valves came from the factory.
The second experiment is labeled "BUDGET INTAKE PORT 2x37mm, stock valves". This is tested combination #16 in Ake's data set.
The only change is a five-angle valve job that gives a larger throat diameter which is then blended to the port. In particular, this experiment is run with unmodified stock 2x37mm intake valves. Valve seat ring ID increased to 33.3mm (90% of the valve size). Cross sectional area (CSA) at the throat per valve is 871 - 33 = 838 mm^2. Valve seat angle 45 degrees x width 0.8mm. Upstream the port 60 degrees x 1,2mm width and 75 degrees x 2,5mm. The upstream port matched to the increased seat ring ID. In the combustion chamber, 32.5 degrees x 1.2mm and 15 degrees x 0.5mm cuts.
Attachment 904331
Attachment 904332
The third experiment is labeled "BUDGET INTAKE PORT 2x37mm, radiused cuts, short-side and long side modified, modified valves". This is tested combination #26 in Ake's data set. Modified intake port with 2 x 37mm modified valves.
The starting point is the same valve job as in the second experiment (Ake's #16). The bottom two cuts are then radiused and rounded.
Attachment 904333
There are many opinions about radiusing intake valve job angles. The consensus among writers appears to be that it usually hurts with carb'd 2-valve engines and may hurt or help with fuel-injected 4-valve engines. How much and how far downstream to radius, if at all, depends on how important it is to reduce reversion for the engine in question. Leaving the sharp edges in supposedly reduces reversion. From the flow perspective, radiusing has at most a marginal effect.
Valves are modified. Nail-shaped flat valve, back angle of 8 degrees and back cut of 30 degrees. Valve back rounded and smoothed, the 45-degree valve seat in the valve reduced from over 2mm in the stock valve to 1.0mm in the modified valve. The valve weight goes down as a positive by-product.
Attachment 904334
In addition, the port small-side radius is moved 1mm closer to the seat. The purpose of this modification is to increase the short-side radius, allowing air to turn more gently. The modification is made between 10:30am and 2pm on the left hand side and between 10am and 1:30pm on the right hand side valve. This will increase the valve throat somewhat upstream of the blended cuts, the distance between the short side and long side is 34.3mm.
Attachment 904335
Attachment 904336
The above short-side radius optimization is in my opinion the most challenging part of the third experiment. There are conflicting requirements of keeping the cross-sectional area down, providing air a straight section before the seat to settle down, and increasing the short-side turn radius. Ake can probably give much more detail on this step if readers so desire.
The long-side radius is also modified by making the divided ports slightly wider and taller in the area where the valve guide is protruding.
Attachment 904337
In terms of numbers, here is some guidance to what the cross-sectional areas are supposed measure at after these modifications:
- CSA at seat ring ID 871+ 30 - 33 = 868 mm^2.
- CSA at valve guide 855 + 30 + 30 - 33 = 869 mm^2.
- CSA where the divided ports begin 804 + 36 + 54 - 20 = 874 mm^2.
Here are the flow results from the three experiments:
Attachment 904338
The first experiment has no surprises, since it is the stock port.
The second experiment increases the throat-to-valve diameter ratio, which helps the high lift flow but at best doesn't lose any low-mid and mid-lift flow. Still, with big cams, the second experiment is already a big improvement. With small cams, not so much.
The third experiment involves a lot more time-consuming modifications. The flow results show that those are worth it. Basically, the third experiment takes the flow curve from the second experiment and shifts it up across the lift range. This is a very impressive result with 37mm valves. Flow rates of 350+ CFM at 28 inches of water stand comparison against most other ported 928 heads, regardless of the valve size.
In terms of port efficiency, the picture is similar. Recall that the below graph defines port efficiency as the ratio of actual flow to the predicted flow from a simple compound restriction formula that takes into account the valve curtain area and the throat area.
Attachment 904339
The stock port is very efficient at the 5-9mm lift range. This is mainly due to the very low 83% throat-to-valve diameter ratio. If one is limited to low lift cams, then one should modify Ake's valve job angles that result in a slightly smaller throat-to-valve diameter ratio.
The third experiment is more efficient everywhere than the second experiment. Again, the modifications are paying off.
The final graph is port velocity at its fastest point. This is not the velocity of the port in a running engine, instead it's the computed velocity at the flow bench. One could also directly measure velocity at some points with a pitot-static probe, but this is not that kind of velocity. It's simply the flow rate vs. the minimum cross-sectional area.
Attachment 904340
I am assuming above that all three ports still have 1400 mm^2 MCSA upstream in the port. This estimate will likely be revised in the future with the benefit of better measurements. Downstream I am accounting for the seat ring area and the valve curtain area.
As one would guess based on the efficiency graphs, the stock port has high velocity at the minimum cross-sectional area (MCSA) at 5-9mm lifts. Also as expected, the 90% throat valve job loses to the stock port in velocity below 9mm and beats it above. Impressively, the third experiment matches the stock port in velocity at 6-7mm lifts and beats it rather clearly both below and above that lift range.
My personal conclusion from Ake's study: These experiments show that one can improve the 928 4-valve ports with some thoughtful modifications expertly implemented with good tools, without having to buy expensive trick parts.
Talk about a mountain of work by you and Ake, the hours that must have gone into these studies is mind boggling and I hope to benefit when my 4 V heads are done by Peter over here. Btw will shortly post new pics of the bigger 2 valve heads.
Quite incredible flow numbers being achieved and the power potential both N/A and forced are staggering. We haven't yet seen what the 42 mm valves can do and remember we are at just over 400 cfm with 39 mm valves. Also the larger bore of 4.25" has not been used has it?
Tuomo as I don't have my pipemax program anymore, can you run a simulation with the following specs, 7.2 litre bore 4.285" stroke 3.76" 435 cfm intake flow 300 cfm exhaust flow. (N.B we don't have Ake's exhaust work yet but just guessing) and 7,750 rpm? It would be interesting to see where we will come out. Cheers for all that hard free work you are sharing.
01-20-2015, 04:29 PM
#243
Rennlist Member
Thread Starter
I am sitting there before I sent my heads to get machines, and I was thinking about the movement of the valve through its cycle.
Its at max lift once. It's at all other lifts twice. So it seems to me that in the second graph, we need to figure out how to get the blue line higher BELOW 8mm.
Its at max lift once. It's at all other lifts twice. So it seems to me that in the second graph, we need to figure out how to get the blue line higher BELOW 8mm.
01-20-2015, 05:16 PM
#244
Nordschleife Master
I am sitting there before I sent my heads to get machines, and I was thinking about the movement of the valve through its cycle. Its at max lift once. It's at all other lifts twice. So it seems to me that in the second graph, we need to figure out how to get the blue line higher BELOW 8mm.
It's true that the camshaft goes thru the below peak lift points twice, but you also have to take into account that because of the camshaft profile being what it is, the valve goes thru the lift points right below the peak lift twice and very slowly. The heuristic of going thru once or twice fails because it doesn't take into account the how fast the valve moves at each lift.
Last edited by ptuomov; 01-20-2015 at 09:17 PM.
01-20-2015, 05:18 PM
#245
Nordschleife Master
The work was distributed unevenly with Ake doing 100% of it and me doing 0%...
Pipemax doesn't take head flow as inputs, it gives some indications of required head flow. I can try to use another program that does it the opposite way.
Pipemax doesn't take head flow as inputs, it gives some indications of required head flow. I can try to use another program that does it the opposite way.
Talk about a mountain of work by you and Ake, the hours that must have gone into these studies is mind boggling and I hope to benefit when my 4 V heads are done by Peter over here. Btw will shortly post new pics of the bigger 2 valve heads.
Quite incredible flow numbers being achieved and the power potential both N/A and forced are staggering. We haven't yet seen what the 42 mm valves can do and remember we are at just over 400 cfm with 39 mm valves. Also the larger bore of 4.25" has not been used has it?
Tuomo as I don't have my pipemax program anymore, can you run a simulation with the following specs, 7.2 litre bore 4.285" stroke 3.76" 435 cfm intake flow 300 cfm exhaust flow. (N.B we don't have Ake's exhaust work yet but just guessing) and 7,750 rpm? It would be interesting to see where we will come out. Cheers for all that hard free work you are sharing.
Quite incredible flow numbers being achieved and the power potential both N/A and forced are staggering. We haven't yet seen what the 42 mm valves can do and remember we are at just over 400 cfm with 39 mm valves. Also the larger bore of 4.25" has not been used has it?
Tuomo as I don't have my pipemax program anymore, can you run a simulation with the following specs, 7.2 litre bore 4.285" stroke 3.76" 435 cfm intake flow 300 cfm exhaust flow. (N.B we don't have Ake's exhaust work yet but just guessing) and 7,750 rpm? It would be interesting to see where we will come out. Cheers for all that hard free work you are sharing.
01-20-2015, 06:22 PM
#246
Addict
Rennlist Member
Rennlist Member
01-20-2015, 07:03 PM
#247
Nordschleife Master
Curious whether there are other valves (perhaps from other non-Porsche cars...) with the right diameter and stem length but already shaped with the flatter tulip under the head. A couple decades ago I found some hi-po aftermarket Chrysler valves for a Rolls-Royce engine that were drop-in replacements at a tiny fraction of the R-R prices. Just needed to use the Chrysler keepers and retainers. There's a pretty amazing supply of valves in the aftermarket.
For these tests, we were just interested in what one could do on a budget, given stock heads in good condition and tweaking them in a cost effective way.
01-22-2015, 07:56 AM
#248
Nordschleife Master
A question for the committee: Any thoughts on how the exhaust port should be ported and how to evaluate whether the modifications are an improvement?
Kind regards, the Secretary of the Porting and Polishing Committee
Kind regards, the Secretary of the Porting and Polishing Committee
01-22-2015, 12:42 PM
#249
Rennlist Member
Thread Starter
I started this thread long ago in an attempt to get people's input into the low hanging fruit of porting. I got an initial answer early on, and went with that with two sets of heads that have not yet seen combustion.
After doing that work, I realized that I love this science, but I also love knowing there are subject matter experts, and that I can rely on. The amount of time and effort to do some of this, for me personally, is better put into actually getting the system where I want it first.
But this is great. Much appreciated. Great pics. I will also say this - all this is for naught when the car pulls 3-9 degrees of timing when a squirrel sneezes in the next county thinking its knock.
After doing that work, I realized that I love this science, but I also love knowing there are subject matter experts, and that I can rely on. The amount of time and effort to do some of this, for me personally, is better put into actually getting the system where I want it first.
But this is great. Much appreciated. Great pics. I will also say this - all this is for naught when the car pulls 3-9 degrees of timing when a squirrel sneezes in the next county thinking its knock.
01-22-2015, 07:28 PM
#250
Nordschleife Master
Some theory ramblings on exhaust porting
I've been thinking about exhaust porting and trying to understand what is happening there.
On the one hand, some ported 928 4-valve heads have the exhaust ports opened all the way to their physical limits. On the other hand, probably the highest powered normally aspirated 928 engine runs 1.75" OD headers and small exhaust ports with stock size exhaust valves. What's going on?
Here's how I've rationalized these two very different approaches. It comes down to gas exchange at overlap vs. pumping losses.
If you run cams which have a lot of overlap, then you perhaps counter intuitively want small exhaust ports, small exhaust valves, and small headers. I say perhaps counter intuitively because at first blush you'd think you'd need big holes and pipes to get the exhaust gas out of the chamber so it doesn't go into the intake port during the big overlap. Turns out the opposite is true. If the exhaust valve, ports, and headers are small, then the combustion chamber gets emptied extremely well at overlap. This is because the small port means high velocity, high velocity means low static pressure, and low static pressure means the intake port sees "suction" from the cylinder at overlap. This is especially powerful effect if the static compression ratio is high.
Of course you could in theory make the exhaust ports so small that the cylinder can't be emptied despite of the high velocity, but if you are starting with factory ports that are large the odds are this will never happen.
The other side of the tradeoff is pumping losses. The small exhaust ports mean that the piston will have to do more work pumping the exhaust gas out, to generate that high velocity. Some of that can be combated by opening the exhaust valve early, which will lead to more exhaust gas escaping the cylinder before the piston bottom dead center. But some of it will remain and the smaller port will force the piston and crankshaft to do a little more work keeping the exhaust flow going.
So it's a tradeoff between pumping losses and gas exchange. There's some optimal point at which the benefits of high velocity and costs of a lot of pumping work together net out to maximum power.
Is there ever a case when "large" exhaust ports are optimal? I think large exhaust ports might help power if the cams have minimal or no overlap. On the one hand, the benefit from the high gas velocity is not as large, since the impact of exhaust pulling in fresh charge is much smaller when the intake and exhaust valves are not open at the same time. On the other hand, the big ports still reduce pumping losses or allow opening the exhaust valve later. So when the cams have no overlap, the tradeoff shifts towards larger exhaust ports.
Much like on the intake side, one has to have some idea about the cams that will be run on the engine before one can port the exhaust well. Conversely, one needs to know a lot about the exhaust valve, port, and headers before one can pick the optimal camshaft profiles.
That's at least what I think. Does that make sense?
On the one hand, some ported 928 4-valve heads have the exhaust ports opened all the way to their physical limits. On the other hand, probably the highest powered normally aspirated 928 engine runs 1.75" OD headers and small exhaust ports with stock size exhaust valves. What's going on?
Here's how I've rationalized these two very different approaches. It comes down to gas exchange at overlap vs. pumping losses.
If you run cams which have a lot of overlap, then you perhaps counter intuitively want small exhaust ports, small exhaust valves, and small headers. I say perhaps counter intuitively because at first blush you'd think you'd need big holes and pipes to get the exhaust gas out of the chamber so it doesn't go into the intake port during the big overlap. Turns out the opposite is true. If the exhaust valve, ports, and headers are small, then the combustion chamber gets emptied extremely well at overlap. This is because the small port means high velocity, high velocity means low static pressure, and low static pressure means the intake port sees "suction" from the cylinder at overlap. This is especially powerful effect if the static compression ratio is high.
Of course you could in theory make the exhaust ports so small that the cylinder can't be emptied despite of the high velocity, but if you are starting with factory ports that are large the odds are this will never happen.
The other side of the tradeoff is pumping losses. The small exhaust ports mean that the piston will have to do more work pumping the exhaust gas out, to generate that high velocity. Some of that can be combated by opening the exhaust valve early, which will lead to more exhaust gas escaping the cylinder before the piston bottom dead center. But some of it will remain and the smaller port will force the piston and crankshaft to do a little more work keeping the exhaust flow going.
So it's a tradeoff between pumping losses and gas exchange. There's some optimal point at which the benefits of high velocity and costs of a lot of pumping work together net out to maximum power.
Is there ever a case when "large" exhaust ports are optimal? I think large exhaust ports might help power if the cams have minimal or no overlap. On the one hand, the benefit from the high gas velocity is not as large, since the impact of exhaust pulling in fresh charge is much smaller when the intake and exhaust valves are not open at the same time. On the other hand, the big ports still reduce pumping losses or allow opening the exhaust valve later. So when the cams have no overlap, the tradeoff shifts towards larger exhaust ports.
Much like on the intake side, one has to have some idea about the cams that will be run on the engine before one can port the exhaust well. Conversely, one needs to know a lot about the exhaust valve, port, and headers before one can pick the optimal camshaft profiles.
That's at least what I think. Does that make sense?
01-22-2015, 10:46 PM
#251
Rennlist Member
Thread Starter
RE the ex port size and primary size (and length)
Using the existing port size and the s3 cam specs, on the dyno software Todd uses, I found that I lost very little tq down low going from 1.6x primaries all the way to 2.2. Shortening them did affect power lower down..
The pumping losses you suggest would exponentially increase with rpm no?
Using the existing port size and the s3 cam specs, on the dyno software Todd uses, I found that I lost very little tq down low going from 1.6x primaries all the way to 2.2. Shortening them did affect power lower down..
The pumping losses you suggest would exponentially increase with rpm no?
01-23-2015, 07:43 AM
#252
Nordschleife Master
RE the ex port size and primary size (and length). Using the existing port size and the s3 cam specs, on the dyno software Todd uses, I found that I lost very little tq down low going from 1.6x primaries all the way to 2.2. Shortening them did affect power lower down.. The pumping losses you suggest would exponentially increase with rpm no?
In any case, I don't have any real world experience on this, so those are just my thoughts and speculations and nothing else.
01-23-2015, 10:14 PM
#254
Nordschleife Master
In particular, here's the one difference I can think of. One of my (copied from the internet) theories is that when the exhaust valve opens, the flow goes sonic at the small opening between valve seat and valve. There's enough heat and pressure there. If so, we have a supersonic convergent-divergent nozzle that produces supersonic exhaust gas flow. The shape that the flow sees after the valve seat determines how much of the pressure and temperature of the exhaust gas is turned into kinetic energy of the exhaust gas flowing out of the port.
Since I am not aware of any supersonic flow benches with continuous flow, what the flow bench reads will be relatively unrelated to how the supersonic exhaust gas flow actually behaves at the beginning of the exhaust blowdown. We either have to get huge tanks pressurized at 200 psi and do burst flow experiments with sonic flow (dangerous, valve pieces in the frontal lobe slow down your thinking) or rely on unproven theories until we see how the engine runs.
Rocket scientists, literally, design ideal shapes for the divergent sections that turn the maximal amount of pressure and temperature energy into kinetic energy. While those shapes are generally interesting, one specifically interesting observation is that they are usually different from the shapes that flow well with subsonic flow. With subsonic flow, expanding the cross-sectional area after the throat slows the flow down, which makes intuitive sense. With supersonic flow, expanding the cross-sectional area after the throat accelerates the flow, which doesn't make intuitive sense to me (but the math is a tyrant). Therefore, the shape that accelerates supersonic flow (good within reason) is also the shape that decelerates the subsonic flow (usually bad).
This is a bit of a problem because as the valve opens more to higher lifts, the flow turns from supersonic to subsonic and a different port geometry should be appropriate.
So one potential explanation to why the tulip shape exhaust valves seem to work well is that at low lifts when the valve, valve seat, and port form a de Laval nozzle. That's a good shape as long as flow is supersonic. This puts requirements on the shape of the exhaust valve back more than anything, as that's the only thing that move location with lift. Something along these lines if you are into theory: http://exploration.grc.nasa.gov/educ...et/nozzle.html
Then as the valve is lifted more and the tulip back of the valve is far into the cylinder, the port is exposed without the valve tulip back in it and now the geometry is different and likely more suitable for subsonic flow. Seems to me that the only way to change the geometry from supersonic accelerating nozzle to subsonic velocity-maintaining nozzle is to put a lot of the geometry into the valve since that valve moving is the only thing that can change the geometry.
That might explain some of the counter intuitive but proven exhaust port rules out there.
If this is correct, one should test exhaust port at the highest possible pressure differential that the flowbench can push and then largely ignore the low-lift flow numbers.
Maybe. Any thoughts?
01-23-2015, 10:34 PM
#255
Rennlist Member
Thread Starter
More suposition:
Exhaust science in non turbo seems to be about scavenging and pumping losses ( as you also mentioned previously).
Much is written about primary size and length, as well as how one would wish for the primaries to be arranged in the secondaries (collector)....
The turbo f1 article posted essentially says Mercedes did great things with a very small (relative) one piece turbo manifold.
Todd said something interesting a few years ago. He strayed a bit with his last engine (and acknowledged this); but he essentially said, relative to budget - what's your goal? Driving or spending 30k or 50k on an engine that can still blow up? If you just design some really great heads and intake, the bottom ends can be plucked out and in when they melt down. Spend the money on the stuff that makes the engine better, but that is also "reusable". Take it as you will. Over time, for him, I see it as the 100s of hours in tuning.... That's repeatable. Bottom ends come and go....
Hence the reason for my ultra high CR rat motor that is almost done. Cheap, used parts throughout. For what may happen to it with 12cr and a Novi, it should be cheap.
I've digressed.
Exhaust science in non turbo seems to be about scavenging and pumping losses ( as you also mentioned previously).
Much is written about primary size and length, as well as how one would wish for the primaries to be arranged in the secondaries (collector)....
The turbo f1 article posted essentially says Mercedes did great things with a very small (relative) one piece turbo manifold.
Todd said something interesting a few years ago. He strayed a bit with his last engine (and acknowledged this); but he essentially said, relative to budget - what's your goal? Driving or spending 30k or 50k on an engine that can still blow up? If you just design some really great heads and intake, the bottom ends can be plucked out and in when they melt down. Spend the money on the stuff that makes the engine better, but that is also "reusable". Take it as you will. Over time, for him, I see it as the 100s of hours in tuning.... That's repeatable. Bottom ends come and go....
Hence the reason for my ultra high CR rat motor that is almost done. Cheap, used parts throughout. For what may happen to it with 12cr and a Novi, it should be cheap.
I've digressed.