Engine technology thread.

[OrbitalEllipses]

Originally Posted by Homemade WRX

Ok before things get any worse with rod ratio and confusion I really ought to start a thread on this. Guy's that know me or my 3MI Racing stuff on NASIOC and can see my comments on Khiem's long rod article on MotoIQ, which he did a nice job on.

I'll be going through the bottom end combinations and publishing the data when parts come to market...maybe before.

I was waiting for you to pop in.

[arghx7]

There are so many factors involved with the rod length discussion. It's fun to talk about but you can't draw too many conclusions with basic calculations. For example the wrist pin and crank offset affects the side forces on the piston. BMW has optimized their N20 engine (for the F30 chassis 328i) for positive crank offset and negative wrist pin offset in order to reduce noise.




modern engines have gotten so optimized to meet strict targets of noise, oil consumption, emissions, and longevity that it becomes difficult to optimize something without a lot of expensive equipment, manpower, and software.

[arghx7]

Originally Posted by Ryephile

I might as well add this from my short time with the BRZ earlier this week. The O2 sensors are 4-wire heated narrow-band. Are there any other GDI cars that use narrow-band? Perhaps I'm too used to GDI + turbo having WBO2 these days. It's not the end of the world, I'm just a bit disappointed....it definitely shows the low-cost nature of the car.

I can guarantee you they are planar-type Denso wideband (or A/F sensor as the Japanese call it, as opposed to an H02S heated oxygen sensor). It's probably version 5.1 . The other Subarus are using some version of the same sensor You are thinking of the much better known pump-cell-type Bosch LSU sensor that is more popular on European cars (including Opel-sourced GM vehicles). I have a bunch of diagrams and technical info I can post... the reason why they have 4 wires instead of 5 is that they operate a little differently than the Bosch LSU sensor

Originally Posted by Dimman

Factory cars are coming with widebands these days?

They were on stock cars before they were available in the aftermarket. Remember that aftermarket wideband systems use mass-produced sensors, especially the Bosch LSU 4.2 . And most new German cars are using the 4.9 or the LSU-ADV. These newer sensor are mostly optimized for a much faster warmup to meet emissions standards.

You could name these three goals as having led to mass adoption of factory widebands:

1) improving catalyst efficiency--a wideband can more precisely control the amount of oxygen stored in the catalyst at a given time. With two narrowband sensors, you compare the amplitude of the front and rear waveforms to infer oxygen storage capacity. With a wideband you can increment an oxygen storage capacity model more easily and check catalyst efficiency more easily.

2) improving HC emissions at cold start-- widebands have smarter and more expensive heaters and can help control the mixture right after engine start. We are talking about closed loop control in ~10 seconds now.

3) improving CO emissions in transient conditions--on simpler/older engines, you have a closed loop condition where you oscillate the AFR around 14.7:1 (for pure gasoline). You have an open loop control system for heavier loads, and some logic for switching between the two. The problem is that it leads to excessive CO emissions especially. So the wideband can be used to hit a richer target AFR to keep the catalyst from overheating.

So off the top of my head here are some outlines of the control strategies used for a non lean-burn engine:

1) closed loop/open loop zones

This is common up through the late 90s at least. There are conditions to determine when the engine needs enrichment, like a throttle position or calculated load

2) closed loop/open loop zones with open loop delay

This is how Subaru has been doing it on the EJ engines for a while. You have a set of conditions which control some kind of delay timer. Once the delay timer counts up, the engine goes into enrichment mode.

3) Closed loop @ lambda 1, open loop delay, closed loop @ enriched AFR, open loop

You have a set of conditions determining when the engine goes into each mode. You have to balance fuel economy considerations, driveability concerns, exhaust temp/cat life, CO emissions, and knock to a certain extent. This requires a factory wideband.

4) Closed loop all the time

There are a number of ways to implement this, but it requires an especially accurate sensor and sophisticated models. BMW does this. Higher end cars from other makes do it... your probably won't find it on your mom's Yaris until the emission standards get too tight.

Some of the conditions that are taken into consideration are calculated engine pumping efficiency, rpm, measured/calculated catalyst temperature, water temperature, and a bunch of failsafe stuff

[Ryephile]

Originally Posted by arghx7

I can guarantee you they are planar-type Denso wideband (or A/F sensor as the Japanese call it, as opposed to an H02S heated oxygen sensor). It's probably version 5.1 . The other Subarus are using some version of the same sensor You are thinking of the much better known pump-cell-type Bosch LSU sensor that is more popular on European cars (including Opel-sourced GM vehicles). I have a bunch of diagrams and technical info I can post... the reason why they have 4 wires instead of 5 is that they operate a little differently than the Bosch LSU sensor

I knew you'd have excellent insight Japanese automotive isn't my forte', hence all my German and American references. Denso planar zirconia with accuracy between 12:1 and 19:1; you learn something new everyday. These seem like they're "quasi-wideband", and likely with a reasonable price to match.

Here's the picture I took of the 4-wire O2's connectors. Two black, one white, and one blue-sheath conductor. Look familiar to the Denso A/F?


RYE_3968.jpg by Ryephile, on Flickr

[arghx7]

Yup. Two black wires for the heater (power + ground, basically not polarized), blue for sensor signal, white for signal ground. That's definitely a Denso sensor and undoubtedly the planar A/F sensor type. The detection range depends on the model. I can't say I've tested/observed every model, and I don't want to go into too much detail here, but in my opinion they are accurate "enough." This is compared to some reference measurements from certain equipment. The Bosch LSU sensors are more expensive and complicated to control and also more accurate in enriched conditions.

I have also used both sensors to tune modded turbo cars, Subarus and otherwise. They get the job done--people get so fixated on AFR. It doesn't have to be super accurate if you are just trying to control preignition with enrichment.

I was talking about AFR control strategies. Here are some basic MS-Paint outlines:



This is your typical system up through the 90s. It's either in the closed loop zone or the open loop enriched zone.



This is what Subaru and a lot of other makes have been doing for a while. You have a region that's always closed loop at 14.7:1, a region that's always enriched, and a region that is subject to a delay timer and a number of other conditions.


This is a compromise system that basically adds a rich closed loop area at say mid rpms. The sensor has to be a wideband for this to work, but it doesn't have to be super accurate at a very rich mixture and it doesn't necessarily go with a really complex internal control strategy.

The fourth method is to be closed loop all the time, and model/measure catalyst temp, torque output, etc. You can use different types of feed forward and feed back control together. This is what the most advanced systems do. BMW does this using Bosch or Siemens engine management. BMW also sells expensive cars with expensive high-tech engines. There's still enrichment under certain conditions that could overheat the cat. These days though with direct injection, water cooled exhaust manifolds, and heat resistant materials you may see more and more of this on downsized turbo engines.

[serialk11r]

Thanks arghx7, tons of useful stuff as usual...
Offset cranks...doh forgot about those. Is there anything important about them that we should know? I imagine the offset reduces friction at low rpm by having the rod at a smaller angle on the power stroke, but what about at high rpm where the "inertial" forces are much greater?

Also I was wondering why engines usually run open loop at high rpm, if you don't mind.

[Want.FR-S]

Originally Posted by Dimman

I picked 138mm because I could've sworn that is the BEAMS 3SGE rod length. But yeah, I wasn't thinking about the space limits in the H4, with the asymmetric rods, but it could depend on piston design too.

But while looking for confirmation on the 138mm I ran into this page by accident:

http://trdparts.jp/english/parts_engine-3s-ge.html

270 PS (~266 bhp SAE) @ 8600 rpm all motor, all parts formerly available through Toyota Technocraft/TRD Japan.

This was when TRD parts were legit, not re-badge so-so other parts...

Just to chime in with two graphs I made with matlab to answer your question:



This is the acceleration graph with stroke 86 mm at rev 7000 and different rod lengths. 200 mm rod length was just chosen to show how the long rod would be.



This is the acceleration graph with stroke 86 mm at rev 8500 and different rod lengths. 200 mm rod length was just chosen to show how the long rod would be.

This is just number crunching. Nothing fancy here.

[arghx7]

Originally Posted by serialk11r

Thanks arghx7, tons of useful stuff as usual...
Offset cranks...doh forgot about those. Is there anything important about them that we should know? I imagine the offset reduces friction at low rpm by having the rod at a smaller angle on the power stroke, but what about at high rpm where the "inertial" forces are much greater?

It reduces friction and piston slap, but I'm not 100% sure if that effect is more or less significant at higher rpm. Check out this document https://docs.google.com/open?id=0B_j...c0hLTURYOWxyZw You can jump to page 38.

Also I was wondering why engines usually run open loop at high rpm, if you don't mind.

In the old days on a production car you had a narrowband sensor before the cat and that's it. That was when there were fewer drive cycles for emissions/fuel economy tests: the FTP75 (can also be referred to as LA4), the HWFET (Highway fuel economy test), and certain Japanese and European cycles. It's conceivable that you could drive through those cycles and never get out of closed loop once the engine warms up. Read this about a certain old 80s fuel injection system, roughly comparable to a simpler Megasquirt by my estimation: https://docs.google.com/open?id=0B_j...LTlBTjBJVWhXUQ

Then the government started adding more test cycles that apply additional load to the engine (SC03 which has the air conditioning on, and US06 which goes at higher speed and acceleration). Now you are going into enrichment more while at the same time emissions and fuel economy standards are getting tighter. So they introduce the closed loop delay to pass the tests and improve fuel economy. You have to be careful with the way it's set up or exhaust temperatures get too hot and the catalyst ages prematurely. As standards got tighter they had started putting in rear O2 sensors (to have a Catalyst monitor) and over time more factory widebands in the front.

Read this for some info on how the narrowband in front and rear system works: https://docs.google.com/open?id=0B_j...Z1ItUjlDWURBQQ

and read this for more info on one type of planar wideband in front + narrowband in the rear catalyst strategy: https://docs.google.com/open?id=0B_j...c3duZTZ5UTBQdw

With closed loop under all rpm and load points, in the end you can boil it down to cost. It takes more development and testing man hours, more sophisticated engine controllers and models, plus better sensors to do it that way. Multiply that by a gazillion cars and the costs add up. That's why the more sophisticated control strategies are mostly on higher end mass production cars and in areas with the tightest emissions standards (US and Europe).

Depending how you put it together, it's more than just having a target AFR table inside the ECM and some kind of basic feedback control. You really need to have rationality checks and monitors to make sure the model and the sensors are working correctly. If you go closed loop at all rpm and something goes wrong with the model (exhaust too lean), the catalyst could burn up from high exhaust temperatures or misfire, or you could severely degrade exhaust components, or even have a significant knock event. If the model is off and the exhaust is too rich, you are back where you started with open loop and you get high CO emissions. Then you can't certify the car, or if you do certify it you get nailed in 4 yours when you start doing in-use testing and the CO fails.

[serialk11r]

Thanks for the links! I guess my question was more, what makes it harder to keep things in check at high rpm?

[Grip Ronin]

Originally Posted by serialk11r

Not really, they're very different. Honda has both a system that will vary the cam timing, and 2 lift profiles to improve volumetric efficiency across a large range of rpm, which this lacks. The "VTEC kicked in yo" is not because of a cam advance, but because the engine switches to a long duration cam lobe that is meant for maximum power. The low duration cam lobe could make more power if they wanted it to, but they kept it mild to presumably reduce emissions and fuel consumption.

i think i see what your saying. the frs wont have the famous cross over of hondas and the mivec of mitsu? i do love my vtec lol

[arghx7]

Originally Posted by serialk11r

Thanks for the links! I guess my question was more, what makes it harder to keep things in check at high rpm?

I'm not sure I understand your question fully. At higher rpm, exhaust temperatures rise, and the mixture needs to be enriched to protect the catalyst and exhaust system. Usually about 900C is the max you want to go, and even that is way too high for a SULEV and PZEV certification in most cases because it will age the catalyst too fast. 800C is closer to where you want the exhaust temperatures in a lot of cases.

If you enrich the mixture, you have to have a control system. I described the different control systems above: open loop, open loop subject to delay, or closed loop. The richer you want to go, the better the model/control strategies and the sensor need to be if you want to go closed loop. At the highest rpms, exhaust temperatures tend to be the highest and thus enrichment requirements are greatest especially when ignition timing has to be retarded on crap gas to prevent knock.

Originally Posted by Grip Ronin

i think i see what your saying. the frs wont have the famous cross over of hondas and the mivec of mitsu? i do love my vtec lol

MiVEC as currently implemented in the Evo is essentially the same thing as AVCS on the current Subarus. It has cam phasers operated by oil pressure. There is no variable lift on the 4G63 or 4B11T. Mitsubishi probably has it on other obscure vehicles, but nobody gives a crap about anything but the Evo anyway.

[Grip Ronin]

Originally Posted by arghx7

MiVEC as currently implemented in the Evo is essentially the same thing as AVCS on the current Subarus. It has cam phasers operated by oil pressure. There is no variable lift on the 4G63 or 4B11T. Mitsubishi probably has it on other obscure vehicles, but nobody gives a crap about anything but the Evo anyway.

lol a kid i met at college had the fake evo things a rally art first time i noticed mivec because it had a loud crossover, but if its the same as avcs it shouldn't be changing the cam profiling to create the power increase?

[Homemade WRX]

As someone caught on already, the actual crank offset is the only thing I'm waiting on to do some bottom end finalization. That and I want to varify the change to the FA's compression height.

Just some quick work up, similarly done for the EJ/EG platforms.



^^^yellow line should read 83mm stroker (woops!)







You'll see for the simple derivations, I did them according to crank angle and time. I like leaving time in minutes so that people can realize how large a milisecond is when talking about combustion. I was actually joking about this at a 'green' combustion seminar/discussion this last week. We're talking nano and pico seconds for processing and guys were quoting mili for outdated controls functions and data aq.

Oh also that last graph is the best simple method of getting a visual for 'friction factor' that rod angle has on the skirt/wall/rings.


And now I wait to varify the FA20 and finalize some values and calcs...Then prototypes are ordered up and guinea pig testing is to commence.

[GrimmSpeed]

I've gone through this entire thread now, and i appreciate all the information, speculation, and random thoughts in here. I'll be contributing as the days go on as well.

HomemadeWRX, i also can't wait to see what you come up with. We'll have ours in a few months, and you better believe we'll have some stuff to bolt on.

Chase