User not logged in - login - register
Home Calendar Books School Tool Photo Gallery Message Boards Users Statistics Advertise Site Info
go to bottom | |
 Message Boards » » arghx's technical ramblings thread Page 1 2 3 4 5 [6] 7 8 9, Prev Next  
arghx
Deucefest '04
7584 Posts
user info
edit post

I wouldn't be surprised if modern greases hold up better than what they had literally 100 years ago.

10/11/2014 11:35:22 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Duke, maybe you can comment on the spark lever? I've sat in a Model T but never driven one. I think it's amazing that you were expected to adjust the spark advance, considering we are moving towards autopilot systems in the modern era.



I can't help but think that one day we will look on manual transmissions like I look on spark advance adjustment levers: "You mean people actually drove down the road picking their own gear? It wasn't done automatically?"

10/12/2014 8:18:24 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

High pressure fuel pumps on gasoline direct injected engines use solenoid delivery angle and current control in order to control pressure in the fuel rail. The newest GDI engines go up to about 250 bar fuel pressure, while the most common is 15 or 20 bar systems. The principle of the hardware is common among different suppliers of high pressure fuel pumps.



In the example above, you have three things. You have the red curve, the green box, and the shaded orange area. The shaded orange area is the portion of the delivery stroke which is being used. The green box is basically the duration of solenoid energization. The red is the current profile of the solenoid. The "width" of the green box is fixed. So, the ECU calculates how much of an orange shaded area it wants. The shaded area is based on the position of the green box. The green box is based on the current control (red dotted line). There are governing feed forward equations and calibrations, plus feedback gains. These are primarily set on a motoring rig and on an engine dyno.

The ECU has to calculate all three of these things (orange, green, red) and coordinate them with other actuators. Most of the newest HPFP's are 3 or 4 lobe, single plunger designs. This doesn't apply to the oddball pump in the BMW N54 engine that was notorious for failing.

10/12/2014 11:30:33 PM

y0willy0
All American
7863 Posts
user info
edit post

how come turbo compounding never took off in autos but only piston planes?

or did it and i just dont know?

turbos being geared via fluid coupling to the crank instead of delivering air to the intake is what i mean. ive always been interesting in large aero piston engines.

10/17/2014 5:59:44 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Are you talking about an exhaust driven turbo or a mechanically driven supercharger? Exactly which type of system. There are a lot and I don't have so much knowledge of what was used when, but I can dig into it and compare it to modern designs on production cars. But supercharger plus turbo and compound turbos are on production cars right now.

[Edited on October 17, 2014 at 6:40 PM. Reason : .]

10/17/2014 6:38:11 PM

y0willy0
All American
7863 Posts
user info
edit post

exhaust driven turbo geared to the crank, not used for air compression

10/17/2014 6:50:29 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Help me out and name a specific plane or engine that used that design. Most of the successful World War 2 aviation engines were similar to this Rolls Royce Merlin engine, a design widely used by the allies:



It had a crank-driven supercharger with air-to-water intercooler and blew compressed air into the engine to allow higher altitude and more power. It was geared, but that would be sort of like having a car with a Roots blower where you could switch between two different pulleys while driving down the road.

10/18/2014 8:21:59 AM

Dr Pepper
All American
3583 Posts
user info
edit post

^he's talking about the turbine shaft driving a gear on the crankshaft - the most recent that I am aware of is a Detroit Diesel engine

[Edited on October 18, 2014 at 8:28 AM. Reason : -]

10/18/2014 8:28:07 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

One reason why that's not used in regular passenger vehicles is complexity and packaging. You have to design a more complex oiling system and chain/gear drive. Think about how long it took for timing chains to be standard on dual overhead cam engines. Timing belts are simpler and quieter but have shorter life.

Some of it is inertia--why adopt that kind of design when nobody/hardly anybody has done it before? If I go to continental, Honeywell, Borg Warner, or Mitsubishi (main turbo suppliers) they are not going to have any experience with that application. They're not going to have any tooling for manufacturing it in a way that meets the requirements.

Another thing to keep in mind is that those 15ish liter Diesel engines run at very low engine speeds. Friction increases exponentially with rpm, so it's a much bigger factor on a 6000rpm passenger car engine vs a 2200rpm heavy duty diesel.

10/18/2014 4:08:15 PM

arghx
Deucefest '04
7584 Posts
user info
edit post



So this image shows motoring friction torque on a pushrod engine (so simpler valvetrain arrangement with low parasitic loss) connected to normal accessories, no blower or anything like that. The engine displacement is greater than 5 liters.

This is measured by hooking it up to an electric engine dyno. The engine dyno spins the non-firing engine to those different speeds. The number represents how much torque is required to hold that speed. It is equivalent to how much deceleration torque is generated when you coast in a vehicle (let off the gas pedal) with the clutch connected or torque converter locked up. You can see the friction torque increases at a greater rate as engine speed goes up. That basic relationship holds true on an any engine.

When you start adding additional hydraulics/gear/chains, like coupling the turbo to a gear drive, you will shift that friction torque up. Since passenger vehicles run at much higher speeds, the friction impact is greater than at an industrial engine that runs at very low speeds. Remember this relationship:

flywheel torque = combustion torque - torque wasted on pumping air through the cylinder - torque wasted overcoming friction

The term friction is sort of a "catch all" for various parasitic losses. Having a conventional mechanical supercharger on an engine would be considered a source of friction. We always want to reduce the friction and pumping losses if possible, if the tradeoffs aren't too great. When I put a bigger turbo on my engine, I am generally speaking reducing pumping losses at higher speeds/high power operation. If I use a roller cams/actuation vs solid cams/bucket actuation, I'm usually reducing friction. When I put stronger aftermarket pistons in my engine, I am allowing the engine to make more combustion torque, but also generally increasing frictional losses. The improved combustion torque outweighs the frictional losses at high loads, but I take a hit at part load efficiency.

10/18/2014 4:29:34 PM

sumfoo1
soup du hier
41043 Posts
user info
edit post

So its a hemi

10/18/2014 8:31:28 PM

y0willy0
All American
7863 Posts
user info
edit post

Wright R-3350 was the most popular engine to use turbo compounding.

Power Recovery Turbines is what they were called; simply a turbo geared to the crank via a fluid coupling.

10/18/2014 11:49:32 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

back on page 2 I discussed control of a modern port fuel injector:



Let's look at solenoid direct injection now. Direct injection is more complicated because they are run with peak-and-hold current control and booster voltage. They also are much trickier to control at low pulsewidths (typically under 1 millisecond). On a direct injector, we call the non linear range the ballistic range. This is because the lift curve looks like a ballistic missle trajectory. The valve is always in the process of opening or closing.



This is particularly important when you try to run high fuel pressure at low loads, requiring short pulsewidths, and when you try to divide the injection mass into multiple injection events of a shorter duration. Control of the valve and armature are important. With a lower flow DI injector, you flow less and may not need short pulsewidths as much at low loads.



So when the DI injector is being driven, you have 4 basic phases:

1) booster voltage phase -- a bunch of current and voltage applied to open the injector quickly

2) high current holding phase - current dithers at a high level as the injector reaches full open

3) lower current holding phase - the majority of the on time when long pulsewidths are used. the split between high current and low current hold phases has effects on injection noise and behavior of the spray pattern, especially with respect to emissions

4) closing phase - sort of the inverse of the boosting phase in terms of the voltage profile

10/27/2014 7:42:16 PM

sumfoo1
soup du hier
41043 Posts
user info
edit post

Shit hitting the fan @ fiat Chrysler over the whole bottom of the list thing.

10/28/2014 8:15:35 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

So BMW has published what they are calling an autonomous driving "roadmap." It really takes this idea of autonomous driving and breaks it down into more individual levels of capability.



Based on publicly available information (media reports etc) I'd say as of now we are capable of full automation in vehicle guidance in some circumstances, at least as a proof-of-concept. Bits and pieces of that capability are already available in production cars, with things like collision avoidance systems and stability control.

In terms of the "monitoring task" and "Performance limits" it seems we are around level 2 and sometimes level 3 in many cases. I'm not an expert on every autonomous driving system out there, but I can think of 2 basic kinds. One is heavily dependent on using up-to-date mapping to deal with a lot of the day-to-day issues you encounter in driving. Google for example is a mapping and software company. They use that technology heavily and it's easy for them to get information about what's going on in northern California where much of their testing is done. Where it gets really tricky is when you want to take an autonomous vehicle to an area that isn't so well mapped out with local traffic information, construction, local weather stations, etc.

A lot of the big auto suppliers that develop those things like collision avoidance, active cruise control, stability control, etc may not have as much experience with mapping software and data. However they have been developing model-based controls for vehicle dynamics and safety (like the Bosch stuff I mentioned in page 2 of this thread). So that's another approach to autonomous driving: using on-board models and a bunch of sensors to help make decisions.

Clearly there are advantages and disadvantages to both approaches (cloud data vs on-board calculations & sensing). I wonder in the future what blend of the two approaches will ultimately be used.

11/8/2014 5:03:24 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

I want to clear up some misconception about a common type of continuously variable valve lift system. The mechanical type used by Hyundai, Nissan, and BMW all have mechanical links in the valvetrain and an electric motor. These systems change the lift and duration profile at the same time. They do NOT:

1. Change the centerline of the cams. This requires a cam phaser, which is a separate but closely related system.
2. Independently control lift and duration. They both change at the same time.
3. Alter the exhaust lift profile in any way

Depending on which system we are talking about though, separate lift and duration profiles can be used for individual valves.

11/11/2014 9:52:23 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Here is a diagram of a basic car lead-acid battery:



You've got individual cells that add their voltages together to create ~12 volts. Each cell has a positive and negative plate, using lead dioxide and lead respectively. There's water and sulphuric acid that works as part of the chemical reaction storing charge. The vent caps release hydrogen gas that is created as the battery charges.

11/19/2014 10:18:03 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

The venerable Flathead Ford was one of the first mass produced V8 engines and its design was heavily influenced by Henry Ford himself.



They called it a flat head because a nearly flat piece of metal covered the top of the deck. Intake and exhaust ports were next to each other, meaning the mixture entered and left the chamber on the same side. This created a counterflow arrangement (as opposed to a crossflow, shown later with a Hemi engine).



The cam is in the valley between the two banks, and it drives flat tappets which then push on what we would now consider really long valves. There are no pushrods. You can't really see it here, but the exhaust actually flows around the head and water jacket, so that the exhaust manifold can be bolted to the outboard side. This design will allow exhaust heat to transfer back into the cylinder head and water jacket, and was one of the major downsides to the flathead Ford. Now, compare the flathead Ford to two later GM pushrod engines, which I previously posted about.



The 409 was an earlier overhead valve design, and the later Turbo-Jet family used the familiar wedge-shaped combustion chamber. The exhaust ports flow directly outboard of the engine to connect to the exhaust manifold. The wedge-shaped chamber is still used on the new direct injected GM pushrod engines, although the exhaust and intake ports of swapped sides in order to reposition the spark plug:



Now compare the flathead Ford and wedge shaped chambers to the original Chrysler Hemi arrangement from the early 1950s:



this is an overhead valve with crossflow head arrangement, where the air comes in one side and the exhaust goes out the other. Look at the valve actuation, with intake and exhaust valve pushrods on opposite sides:



The crossed pushrods are one of the reasons why a Hemi cylinder head is wider than a conventional wedge-shaped, counterflow head. The modern Hemis have a different radius of curvature in the head and two spark plugs to enhance burning characteristics and meet modern emissions standards:

11/22/2014 9:38:19 AM

y0willy0
All American
7863 Posts
user info
edit post

i figured i would post here before creating a new thread (or maybe someone could point me to an existing thread i dont know about)

small custom fabrication projects, such as building a small motor, completely from scratch in a machine shop setting

anybody here ever attempted something similar

11/25/2014 5:42:20 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

My brother is into that. He's built an a compressed air engine with some basic machine tools In his garage. I'm trying to remember what kit/process he used. He's commented that it's not easy to make the parts right.

11/25/2014 9:00:57 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Today's topic is conventional oil-pressure actuated cam phasing versus cam-torque actuated cam phasing. A conventional phaser uses oil pressure, essentially the energy from the oil pump, to move the vanes. When the engine shuts off the intake cam locks at the most retarded position.



A cam-torque actuated phaser (supplied by BorgWarner), like the ones you would find in the Ford Coyote V8 or a Subaru FA-series engine, still uses oil for actuation. Instead of using the hydraulic energy of the oil pump, the oil sort of “sloshes” in the chamber as the cam rotates, using the momentum of the cam for phasing:





Now instead of having the locking pin for the intake cam phaser at the most retarded position, these cam-torque actuated phasers can be put in a "mid lockpin" position (maybe not on every engine with cam-torque phasers, but the newest ones can for sure).



The solid blue line shows the intake cam profile in mid lockpin position. Instead of locking at the most retarded position, which could result in a very low effective compression ratio and starting problems, the cam is somewhere between the most advanced and retarded. This is to support late intake valve closing/Atkinson Cycle cam timing for fuel economy.



Another advanced lockpin system was developed by Aisin for Nissan. This system uses a spring, locking keys, and a hydraulic control system to enable operation:



This system was designed for reducing cold start emissions rather than to support Atkinson Cycle for fuel economy. When the system is engaged, the engine will crank over with more overlap. The overlap pushes unburned hydrocarbons back into the intake port to be burned again on the next cycle. New Nissan engines such as the updated QR25DE engine use this system to meet more stringent government emission standards.



[Edited on December 18, 2014 at 11:22 PM. Reason : added overlap diagram at end]

12/18/2014 11:22:05 PM

Hiro
All American
4673 Posts
user info
edit post

I love that you post these. Thank you for the good read!

12/19/2014 12:14:13 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

you're welcome

12/20/2014 10:46:09 AM

234sapphire
New Recruit
4 Posts
user info
edit post

Hi from the UK. Just paid $5 to register with Wolf Web so I could congratulate you on these excellent, interesting and informative postings. Seasons greetings and keep up the great work in 2015.

Any thoughts on simulated Atkinson cycles for fuel efficiency and in particular why the trend for engines such as Toyota Prius is to close the inlet valves well after BDC thereby ingesting air and then pumping it back into the inlet tract rather than just closing the valves before BDC. My guess is that it is to do with getting sufficient cylinder filling ?? Cheers

1/2/2015 4:27:37 AM

Hiro
All American
4673 Posts
user info
edit post

Cheers and welcome!

1/2/2015 3:07:29 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Thanks for the kind words. I don't feel like going through the tedium of reposting it here right now but I get heavily into Atkinson cycle and the hardware enablers for it on this thread: http://forums.nasioc.com/forums/showthread.php?t=2687550

As for late intake valve closing vs early intake valve closing for dethrottling via valve events, that's a huge topic. It depends on exactly what you are trying to do, what the application of the engine is. I feel the most sophisticated system for EIVC is any BMW Valvetronic 3 or later engine. They use EIVC and offset valve lifts (more lift on one valve vs another to add swirl). Look at the MAP on an N20 or N55 engine and you will see they drive around with 90-95kPa MAP at all times, basically making zero vacuum even at very low loads where it's hardest to pull off that strategy without extremely slow burn.

The problem with EIVC is that you have lots of time to heat up the air charge (making knock more likely) and a lot less time for mixture to form, which is a big problem on DI engines. With short cam duration you've gotta do all sorts of tricks to get the mixture to burn like a super restrictive head for high tumble flow. Or you accept less dethrottling by running at lower MAP, which starts to defeat the purpose of controlling load with valve timing.

The problem with LIVC has a lot to do with the implementation. Do you use a huge ass cam profile to do it (like some Hondas do)? If so, how do you control it? Do you use a modest size lift and duration, and then rely on phasing to do it? How do you deal with the startability issues?, and how do you deal with the drawbacks such as possibly friction and low effective compression ratio/starting issues? My preferred LIVC-capable system is the Fiat MultiAir aka Ina UniAir because of the huge cam profile that can be optimized for best lift and duration at full load.

Remember that the current 2ZR-FXE Prius engine is LIVC with cooled EGR (old 1NX didn't have CEGR), so there is a balancing act between the two where at some loads the cooled EGR works better at other loads it's better to do the LIVC with cam phasing.



[Edited on January 4, 2015 at 10:42 PM. Reason : Prius-specific stuff]

1/4/2015 10:23:10 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

You may have popped the hood on a vehicle before and noticed that, on the duct supplying dirty air to the airbox, there appears to be plastic chambers connected. These are resonance chambers meant to reduce noise on a stock configuration. They help cancel out sound waves.



They are typically deleted on aftermarket intakes to reduce restriction.

1/5/2015 9:09:20 PM

234sapphire
New Recruit
4 Posts
user info
edit post

Thanks for the link to all the Atkinson cycle stuff. Very useful and absolutely fascinating! There's far more to it than I had considered when you take into account EGR, VVT, turbos and direct injection. I am looking at a more simple application for a fuel efficient stationary engine project but a lot of this stuff is still very relevant. Cheers.

1/6/2015 11:39:00 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

So it's like a generator or power take off?

1/6/2015 2:16:55 PM

234sapphire
New Recruit
4 Posts
user info
edit post

Yes....a single cylinder for gensets and the like designed to work over a limited speed range.

1/6/2015 4:06:52 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

^ cool, good luck on the project.

1/6/2015 9:56:30 PM

dustm
All American
14289 Posts
user info
edit post

I'd like to hear more about it if you are willing to share, or already have a thread somewhere. You could always start a build thread here, though it will be relatively slow compared to other forums.

1/6/2015 11:13:08 PM

234sapphire
New Recruit
4 Posts
user info
edit post

I'll see what I can add to the forums as the project develops.
As an aside we are about to develop a 1950's classic car for historic racing. It's combustion chamber is just like the Chrysler Hemi talked about earlier in this forum. Before anyone asks the car in question is a rare 4 cylinder Armstrong Siddeley 234.

1/7/2015 4:06:48 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

So Volkswagen likes to use cooling systems that have some kind of split in flow or variation in control between block and head. One of their simple implementations of this idea is a dual thermostat setup found on one of their 3 cylinder engines.



The warmer block temperature helps to reduce friction, while the cooler head temperature reduces knocking tendency.

1/10/2015 8:15:53 PM

tchenku
midshipman
18568 Posts
user info
edit post

I should engineer them a single-thermostat-double-pathway setup




[Edited on January 11, 2015 at 1:14 AM. Reason : ]

1/11/2015 1:14:38 AM

dustm
All American
14289 Posts
user info
edit post

If that blue arrow doesn't change, cars 1 and 2 are going to collide head-on in intersection cccc

1/11/2015 5:16:11 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

Let’s talk about air induction systems, meaning here the intake system that connects to the throttlebody of a naturally aspirated engine or to the compressor inlet of a turbo engine. So we can discuss the components of a stock air induction system and how restriction can be understood. Then we’ll go into an n/a engine that uses pipe length tuning effects to benefit the engine and then discuss effects of induction systems on the operation of a turbo. The last section will cover the common aftermarket systems marketed as a “cold air intake” and a “short ram intake.”

Stock Air Induction Systems – who designs them?

First there’s something very important to realize. In many companies and/or development programs, the physical design of the air induction system that ends up in the car is the responsibility of the vehicle integration design team. The basic performance, fuel economy, and emissions of an engine are first developed by a core engine development team working in labs and with simulations, operating under some basic assumptions about what the intake and exhaust system would be in the vehicle. Designing that actual induction system falls upon the people whose job it is to take an engine sitting in a lab or on a stand and make it actually work in a car. So why does that matter?

Vehicle platform engineers don’t just think about performance. They only give high priority to performance if there is enough pressure on them to do so, like if it’s a performance-oriented development program (performance vehicle or performance package) from the beginning. Vehicle integration is going to care about cost, manufacturability, packaging, performance and sound. A less performance oriented, high volume application for an engine and vehicle usually gets developed first. Then the performance application (high output version for a sportier vehicle/trim level) comes after a number of design freezes. This limits a lot of the designers’ options for improving performance through induction system design.

You could make a nice airbox, but maybe the location of the battery has been frozen so that can’t be shifted around. You could make a large, straight dirty-side airduct (more on those in a minute) but the headlights would be in the way, or you can’t feed it fresh air well because the front fascia of the engine would need to be restyled and tooling has been frozen. A lot of designs cost $Texas to change after certain deadlines have been reached, because somebody would have to go figure out and build a new manufacturing process. Now with that context out of the way, let’s now look at the air induction system of a basic old 90s Civic.

Components of a Stock Air Induction System

The main elements of a stock induction system are a dirty air duct, an air cleaner box, the air cleaner (the filter), a clean air duct, usually noise cancelling features, and provisions for emissions such as a crankcase ventilation vapor hose.



In the image above found from some ebay image searches, we have the Honda airbox assembly on the left, the dirty air duct and what appears to be a large noise resonator chamber on the right, with the clean air duct below. The dirty air duct can draw air from a cooler part of the engine bay, which is a good thing. In the image above you can see a number of bends in the dirty air duct. That causes a restriction which is bad for performance.

Often you will see the dirty air duct connect from the ambient air source to an inlet on the bottom of the airbox. This makes air flow upward through a rectangular air filter and uses gravity to keep dirt and water out of the engine. The bottom-side dirty air inlet also reduces the chance of anything calling into the clean air duct during service. Leaves or rocks would just fall back into the dirty air duct. A lot of dirty air ducts don’t even draw in particularly cool air… they just get routed wherever in the engine bay, often to a fender area which is only slightly cooler. See my earlier comments about packaging conflicts, design freezes, etc.

The airbox houses a paper filter, typically square and replaceable. As I mentioned before, most conventional airboxes bring the dirty air in from the bottom of the box, so that the dirt and dust will tend to fall back down the dirty air duct during filter removal. The bottom-side inlet helps with keeping the engine clean but limits opportunities to reduce restriction in the ducting and airbox. The clean air duct routes air that has passed through the filter and will connect to the throttlebody or turbo compressor inlet. Often the clean air duct comes out the top of the airbox. It may contain a section for a mass airflow sensor, which usually need a straight undisturbed airflow to read most accurately. Often there is provision to connect a crankcase ventilation line and maybe some component related to fuel evaporative emissions.

From a performance perspective you can actually measure the pressure drop across the different components of the entire induction system separately on a flow bench. Then you can create a pie chart showing pressure drop of the total system under some given set of conditions (airflow assumption). The below pie chart is totally made up by me but it illustrates my point that it’s best to narrow down the restriction in the system. If there are tight bends in one of the ducts, or a very small airbox, you can illustrate its contribution to total restriction with a pie chart. So if I stick a better flowing air filter in there the benefit depends on how much the old air filter was contributing to the restriction in the first place.



So we’ve talked about restriction in the system, but what about the “cold” part? Remember that horsepower certification tests are done in a lab with controlled air temperature, so a cooler air intake source doesn’t really benefit the rated horsepower. If I stick a bigass lab air duct to the airbox and feed it a constant 25C air, for certification purposes it doesn’t matter that in the actual vehicle the dirty air duct pulls in hot air. That results in less incentive to create a dirty air duct with routing that draws in cold air. There isn’t such an easily discernable marketing benefit in the sense that the cold air advantage doesn't show up in a horsepower rating test.

[Edited on January 18, 2015 at 11:34 AM. Reason : .]

1/18/2015 11:32:41 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

Air Induction System Effect on Turbocharger Operation

Let’s talk about a pretty good induction system design for a turbo engine, the Mitsubishi Evo X.



Notice the short dirty air duct pulling cool air from the front of the vehicle (where the hood line is) and flowing axially through the middle from one side of the box to the other. Compare that to the dirty air “snake pipe” we see in the Civic that runs to the bottom of the airbox. You see a minimum of noise-cancelling aspects such as resonance chambers in the system. On the Evo X the turbo sits near the firewall (rear-facing transverse mount), so a few bends are necessary for the clean air duct to connect to the turbo. I don’t have any numbers to show but you would expect the majority of the restriction in the system to come from the clean air duct. So how does restriction affect the operation of the turbo?

There is an exponential relationship between mass airflow and pressure drop. So as I flow more air, restriction increases drastically and we get more and more suction at the compressor inlet. Changing the induction system hardware design results in a different shaped flow vs pressure drop curve:



The more pressure drop the harder the turbo works in the sense that the compressor pressure ratio increases. The pressure ratio is the Y axis of a compressor map. Remember from a previous post this old compressor map from the original Porsche 911 turbo?



The pressure ratio is the relationship between between the compressor outlet pressure and the pressure at the compressor inlet (labeled “p2/p1” on the Y axis above). People think the Y axis on a compressor map represents intake manifold pressure. That is not true, and yes that distinction can definitely matter. At very low flow, the inlet pressure is basically atmospheric. So ambient weather conditions and altitude are the only thing affecting it. As we try to flow more air through the turbo though the restriction curve affects plotting on the compressor map. Our compressor speed increases, potentially moving the engine to a less efficient portion of the map. Less efficiency means pushing out hotter air.

The biggest implication of a restrictive induction system is for turbo behavior near peak power, approaching the choke area of the map (map boundary in the upper right area). It’s much easier to overspeed the turbo with a restrictive enough induction system. The turbo will perform worse at altitude if you have a restrictive intake. It will have less margin in it to crank up the boost—basically, it will run out of breath earlier because it is working harder to overcome the pressure drop at the inlet of the compressor. In the compressor map above, there’s a thick line running through the map with grainy engine speed numbers showing where the engine falls on the map. If we have more restriction on the compressor inlet that line will shift upward, and suddenly the 6000rpm point would be beyond the choke line boundary.

Air Induction System Pipe Length Tuning on an N/A Engine

Now, on an n/a engine there are some tuning effects involved with these systems just like with intake manifolds. Remember the basic rule of thumb that, as far on the intake side (manifolds and induction system), longer pipe length tends to benefit torque at medium or lower engine speeds. Shorter pipe lengths benefit high engine speeds. The now-discontinued Mazda Rx-8 used a 5 stage intake system on its rotary engine:





We’re focusing on pre-throttlebody intake systems here, so look at the top (first image) and bottom left (second image) where you will see the “VFAD.” That’s the Variable Fresh Air Duct. It has a long pipe for low engine speeds and a short pipe for higher speeds, switched by a vacuum actuator. The chart below shows how stages 2 through 5 of the system work. The Renesis has primary, secondary, and auxiliary intake ports. Think of it as a piston engine with 3 intake valves per cylinder, each with their own cam lobe and intake runner that can be individually switched open.

The primary ports are always open and flow is controlled by the electronic throttle. Then the secondary ports open with a vacuum operated valve at 3750rpm. At 5000rpm we switch the variable fresh air duct to use a shorter pipe before the air box. The next stage opens the auxiliary ports (think of it as a 3rd intake valve) and finally the intake manifold runners are switch from long length to short.



I know a lot of people hate on Rx-8’s but this is basically a naturally aspirated Porsche-style intake system that’s very finely tuned. It’s one of the reasons why you couldn’t really mod a Renesis and pick up noticeable power. The stock intake system used almost every trick in the book.

[Edited on January 18, 2015 at 11:43 AM. Reason : .]

1/18/2015 11:42:36 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

Aftermarket Air Induction System - Cold Air Intake

Now let’s look at a popular aftermarket performance induction system marketed as a cold air intake.



It’s important to compare the aftermarket system to the stock one. The stock induction system might already draw in cold air from the front of the vehicle but perhaps the duct is small or has bends that make it restrictive. In that case you’re not really getting so much of a “cold air” benefit with the aftermarket system, but you are reducing restriction. The tradeoffs for that reduction in restriction are changes in sound, less protection from water or debris, and in some cases emissions implications.

In the aftermarket system shown above, the restriction compared to a stock system decreases by simply bypassing the dirty air duct altogether and dispensing with precautions to keep debris or water from getting into the filter. Noise cancelling systems (flexible accordion-like hoses and resonance chambers) are deleted, reducing restriction further. From the emissions perspective, evaporative or crankcase ventilation vapor flow rate or distribution could potentially be affected with the aftermarket system.

Perhaps the stock system only drew in mostly hot air from the edge of the engine bay, because of the way the stock dirty air duct was routed. In that case you would be getting a cold air benefit if the system you bought actually sucks in cold air.

Aftermarket Air Induction System – Short Ram Intake

Let’s say you bought a system that doesn’t draw from colder air. These are often marketed as a “short ram intake system”



This type of system still eliminates the dirty air duct and most likely uses a less restrictive air filter and clean air duct. The system also has a much lower risk of drawing in debris or water. The aftermarket system might also draw in hotter air than the factory system did. So now you have two competing effects – hotter air coming into the engine, but less restriction compared to the stock system. You might even have less restriction than an aftermarket “cold air” system. For n/a engines in some cases you could have a pipe length tuning effect with the shorter pipe – remember the Rx-8 Variable Fresh Air Duct? It’s hard to separate all these effects from each other though without a lot of equipment to take measurements.

Keep in mind that the systems pictured above are for an engine that doesn’t require a mass airflow sensor. Aftermarket systems often disturb the placement of the airflow sensor and preferably they would be accompanied with changes in the ECU to account for it. If we disturb the mass airflow sensor reading, it might result in the ECU calculating less airflow, resulting in leaner fueling and more advanced spark timing. That could have a number for results, anywhere from picking up horsepower, making the engine knock, and causing driveability problems.

Aftermarket Air Induction System – Dyno vs Real World

I will once again point out that the “cold air” aspect of the stock induction system really has very little to zero effect on the rated horsepower. The intake air is controlled in a lab anyway when you’re measuring horsepower on an engine dyno like the stock engine was when it was rated at some number. The restriction of the stock system shows up on an engine dyno with the right instrumentation, but the cold air aspect doesn’t, not in a real-world representative way.

In the real world on an actual vehicle we see different air temperatures with varying conditions. Temperatures depend on how the engine heatsoaks for example, with the radiator and A/C condenser and all that under the hood. On a Dynojet chassis dyno you will probably see the effect of the “cold air” aspect because you are drawing in ambient air. Even then you’ve got the hood open and some kind of fan blowing, and weather changes every day. The restriction aspect, changes in turbo behavior, tuning effects of volumetric efficiency can’t be as easily measured in a Dynojet environment because it’s not an actual lab with the right equipment.

Conclusion

So to sum up all that, all sorts of intake systems can go into a car but there are always constraints and tradeoffs. A typical stock intake system on a blandmobile doesn’t place high priority on performance because let’s face it, the average driver doesn’t care that much and the designers have competing concerns. We know a well-tuned intake system can help an n/a engine and a high flowing intake system especially helps a turbo operate more efficiently at its limit.

When you go aftermarket, you are making packaging, noise, emissions and actual cleaning of air lower priorities. You may see some advertised horsepower benefit of an aftermarket induction system. To understand the real effect of the aftermarket system it’s best to think about it in terms of changes in restriction, changes in the temperature of the air supplied to the engine, maybe even changes to the tuning of pipe lengths and pulsation effects. It’s also important to consider possible unintended effects on a mass airflow’s sensors measurement of intake airflow, which skew our understanding of performance benefits by changing the computer’s engine load calculation.

1/18/2015 11:48:23 AM

Dr Pepper
All American
3583 Posts
user info
edit post

can you delve into the evolution of control over (traditional torque converter equipped) automatic transmissions?

I'm comfortable with:
- the basics behind hydraulically controlled 3-speeds (and variants)
- the basics behind simple overdrive control & TC lock control via solenoids and what variables control (TPS, vehicle speed, engine RPM, etc)

Where I get lost is elegant control and application of the converter lock/unlock in various gears, and how the shift points are calculated. Have anything on that topic?

1/19/2015 5:40:12 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Great topic. This is sort of a controls question and sort of a powertrain systems optimization question.

Let me talk about the optimization of shift points for now, the big picture. In the old days everything was just some hydraulic design in valve bodies, a vacuum actuator thingie and cables. I honestly don't know what kind of process they used to set the shift points. I think a lot of it was limited by whatever mechanically you could make work. I don't know how fine of control you really have before computers. But hey, it's harder for me to understand a carburetor (at least, the more elaborate ones on cars rather than weed wacker ones) than fuel injection as fuel injection seems simpler to me.

The key to understanding shift points is that most of the work, the big picture idea of what gear the engine is going to be in and when, is done in simulation before the powertrain (engine, transmission, final drive, whatever) combination ever exists as physical objects. There are entire departments devoted to coming up with simulations for what the most optimum shift points are based on taking mostly stead-state data on engine and transmissions from labs and extrapolating them into predicted transient behavior.

Here's what I mean by that. You can't just make an engine and transmission, do all the development work to write the software and tune everything, drive it, and see what kind of fuel economy that makes. Besides all the money involved, there's no time for that. If it doesn't work, there's no time to do a big redesign.

It's a bit of a chicken-and-egg thing because in each application which came first, the transmission or the engine, or even the vehicle? So somebody in a lab comes up with 3D mapping of an engine's speed and load curve or even simulates it if it doesn't exist yet. Somebody else comes up with transmission characteristics, friction and gear ratios and such. Somebody else has to figure out what the weight and basic demand on the vehicle is, depending how it's going to be certified. All of this is geared towards fuel economy and emissions certification test cycles... we haven't even gotten to the real world, as cycle fuel economy and emissions are hard enough to predict.

So I can take a speed/load/brake specific fuel consumption 3D map and make it a scatterplot with points representing where you expect the engine to fall in a given cycle under certain assumptions about vehicle weight, gearing, etc. It might look something along these lines (plot just made up by me):



So the different points might represent 1 second spent at this speed and load on a given test cycle (different color is different test cycle). The main goals with shift mapping are to move the points into lower BSFC part of the map while staying within all sorts of constraints. There may not be any option to change any of the gearing if all those designs are carried over. You may be concerned about noise and vibration at certain points. On the engine side, maybe it's an old engine and you can't do much to optimize fuel economy at the speed and load points because the designs are mostly frozen. It all depends on those kinds of constraints.

Since you are into what we could consider medium duty vehicles (in terms of how they are regulated) I have to point out that CO emissions are huge concern for big gasoline trucks and of course NOx/Particulates are the concern for diesels. When you have to dump in fuel every time you do major acceleration on a gasoline engine you will pour out CO emissions, because CO is proportional to enrichment. Otherwise your cat will overheat and the vehicle will fail EPA in-use testing. With diesels of course you have particulate concerns for example from heavy acceleration.

That all fits into the constraints for the shift mapping. The government is cracking down on this stuff. Why do you think the Raptor is now going to a V6? Ford didn't want to dump money into making an old large displacement single overhead cam architecture meet emissions. The LEV II phase in has already done a lot, but California LEV III emission standards won't allow heavy enrichment for big work vans.

So that big picture has a huge effect on shift mapping, and the transmission development and shift points are there to execute the big picture approach and make it work in a mass production product that customers drive.

I think what you also have to think about are torque-based controls, which modern transmissions all use now. I can get into that more if anyone is interested... it involves on-board engine and powertrain torque models (including friction assumptions and such), CAN messages, torque reduction requests for shifts, that kind of thing.

1/19/2015 9:07:25 PM

Dr Pepper
All American
3583 Posts
user info
edit post

^you talking about 'defueling' or whatnot (last paragraph)?

1/20/2015 7:12:40 AM

sumfoo1
soup du hier
41043 Posts
user info
edit post




Rolls royce copied the batmowheel?

Seriously... does this mean it actually works?

1/20/2015 6:38:00 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Ok, it's time for part 2 of transmission controls. I talked about the big picture of how shift points and gear ratios are developed. There's a lot of simulation going on that incorporates lab data, way before the hardware has been matched and parts have been built to actually run around in a car.

All modern transmission and engine controls utilize torque calculations. Some manufacturers use it more than others. I will briefly explain engine torque models and show how it relates to transmission controls, especially the most modern ones that control to a target force/torque/power at the wheels. On a gasoline engine the basic torque calculation comes from airflow, spark timing, and air-fuel ratio.



In this type of system we raise or lower torque through two ways:

1. The slow path-this is mostly through controlling airflow through the engine. It's changing throttle or wastegate position mostly on a gasoline engine, and on a diesel it might be variable vane position or wastegate.

2. The fast path-on a gasoline engine that would be through spark timing and fuel cut. The primary function of the fast path is idle stability, torque reduction between shifts, and fuel cut for engine protection or emissions/fuel consumption reduction. On a diesel it might be fuel cut and injection timing.



The engine model goes into a larger powertrain torque model that accounts for gear ratios and friction/parasitic loss estimation:



As for shift schedules and coordinating with the engine, I actually covered that back as the 5th post on page 5. Basically, the most modern method of control is to calculate a target wheel torque/force/power. Then engine torque and gear ratios are coordinated to get there, and a bunch of algorithms can be used for smoothing that wheel torque delivery during shifts and hard accelerations. This feeds back into the torque-based engine controls, which figure out how to raise or lower the torque (fast or slow path? for a given path, which method?) based on the commands from the torque coordination system.



[Edited on January 25, 2015 at 1:25 AM. Reason : driving force block diagram]

1/25/2015 1:24:52 AM

sumfoo1
soup du hier
41043 Posts
user info
edit post

How do the concentric clutches in the Concentric dsgs work?

Concentric throw out bearings?

1/29/2015 9:33:13 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

So waaaay back on page 1 I had a diagram of an Evo X DCT showing the basic layout of the gearsets.



That's a transversely mounted DCT. That diagram doesn't show much of the clutch operation though. Remember that a DCT has two parallel gearsets for odd and even gears, each connected to their own clutch. The next gear will have already been selected and all the clutch needs to do is be connected after the previous clutch was released. Follow the powerflow of this BMW M6 longitudinally mounted DCT:







But wait, where's the throwout bearing? Remember that almost all the DCT's out there are using wet clutches. They are bathed in oil and there might be some slippage or friction between them. They don't engage and disengage with a throwout bearing connected to your foot with hydraulics or a cable(like a conventional manual). They aren't operated with some kind of servo like the old single clutch Sequential Manual Gearboxes. The DCT's clutches operate more like the multiplate clutches found in a slushbox. Remember that DCT is a sort of hybrid between a slushbox and a classic stick shift.



In your basic slushbox, hydraulic pressure pushes on a piston, which pushes on several clutch plates bathed in transmission fluid with some slippage between them. The clutches on a typical DCT operate a lot more like that than they operate like a conventional manual with a single dry clutch plate. Let's look at a VW DSG transmission:



You can see that the engine has a dual-mass flywheel and an input connection that goes to the DCT clutch carrier. The clutch carrier has an outer portion which is connected to the flywheel.



The inner portion of the clutch carrier is connected to the input shaft (either #1 or #2). Hydraulic pressure pushes the multiplate clutches, just like on a slushbox, and torque is transferred from the flywheel to the #1 input shaft here by connecting the outer and inner clutch carriers.



The same thing applies to the #2 clutch carrier. Hydraulic pressure works on a piston, which works on the multiplate clutches, which works on the inner carrier, which works on the outer carrier, which connects to the flywheel. Varying the hydraulic pressure allows slip and vehicle takeoff from a stop, but it's a lot trickier than with a torque converter, especially if you are concerned about fuel economy.

1/29/2015 11:27:40 PM

sumfoo1
soup du hier
41043 Posts
user info
edit post

Damn... i came up with a pretty sweet idea but it really needs concentric manual throw-out bearings lol.

Trying to figure out how to make it work.

1/30/2015 10:51:52 AM

arghx
Deucefest '04
7584 Posts
user info
edit post

This is what happens when you retard spark too much in a heavy load condition:



Cylinders will experience more and more misfire cycles as the spark timing retards and combustion becomes unstable. For the first two cycles, you can see the characteristic "double hump" of retarded combustion phasing (see my discussion of knock and combustion phasing on the previous page for more info).

The first hump is compression action of the cylinder, and the second hump comes from the combustion increasing cylinder pressure after it starts to burn. At the top of the pressure trace you can see a small peak in cylinder pressure from light knocking, and pressure oscillations at the end of the trace from cylinder pressure oscillations.

On the third cycle, the peak pressure is much lower, and equivalent to compression pressure. The second hump barely registers any rise in pressure at all. Enough of these misfire cycles (on enough cylinders) can cause a loss of engine torque that could be felt in the car, as well as potential damage to components such as a melted cat.

2/23/2015 10:16:24 PM

tchenku
midshipman
18568 Posts
user info
edit post

those DCT components are a lot thinner than I imagined. holy cow

2/24/2015 8:54:37 PM

arghx
Deucefest '04
7584 Posts
user info
edit post

Engine hard parts like pistons and rods are typically designed to a nominal peak combustion pressure limit. The limit is specified in terms of an engine average combustion pressure of a number of cycles, and a max pressure limit for an individual combustion cycle. So the engine should be able to withstand a few cycles higher than the average pressure limit, so if you're at 80 bar average you might be ok to 100 max on an individual cycle basis. The image below compares the 3.5L Ford Ecoboost and the non turbo Ford 3.5L Duratec engines' max cylinder pressure and rate of pressure rise at full load.




You can see the Ecoboost's pistons, rods, etc are rated to 80bar max combustion pressure average and the full load curve reflects that as peak pressure has been limited at high rpm. The peak pressure is typically limited by retarding spark timing. So the engine may still make the load but do it with less spark efficiency. The non turbo version of the engine shown in the image has a lower peak pressure due to the lower loads in the cylinder (lower specific torque output/lower indicated mean effective pressure). The rate of pressure rise is also slower for the n/a engine.

Generally high rates of pressure rise are associated with more combustion noise. That's partly why knock can be heard audibly. It's one of the reasons why diesels are loud. High rates of pressure rise are associated with autoignition in general.

2/26/2015 10:27:18 PM

 Message Boards » The Garage » arghx's technical ramblings thread Page 1 2 3 4 5 [6] 7 8 9, Prev Next  
go to top | |
Admin Options : move topic | lock topic

© 2024 by The Wolf Web - All Rights Reserved.
The material located at this site is not endorsed, sponsored or provided by or on behalf of North Carolina State University.
Powered by CrazyWeb v2.38 - our disclaimer.