I want to expand on one section above about why things can go wrong and my historical analogies:

The people making the big decisions are often very distant in a physical sense and in an organizational sense. When they learn about the status of the program, it is often in very formal settings that I would compare to a medieval audience with monarch at a royal court. There's a lot of protocol about who sits where in The Royal Great Hall, how The Royal Powerpoints are prepared, etc.

Executive drives are when they come down from their castle and mingle with the unwashed masses as they drive some specially prepared prototype, only to return to their royal quarters and feast at their special corporate banquet hall/cafeteria.

I think the Silicon Valley culture has a lot less of that though. But then again, it was probably very different in Henry Ford, Karl Benz, or Billy Durant (the founder of GM)'s day. And when Google or Tesla (or their successors) are 100 years old they may have many of the trappings of old, large, bureaucratic organizations. Osman, first Sultan of the Ottoman empire, rode around on a horse fighting battles. The last Sultan lived in a palace, making bad decisions overseeing a declining state. What we are witnessing in the auto industry is that classic organizational life cycle happen again. Young scrappy upstart vs formal, sclerotic declining institutions.

Keep in mind that the working level people know when it's a bad product. They know when they are making a piece of shit. Eventually though the whole thing turns into the July Crisis that preceded World War I, when the heir to the Austrian throne was assassinated. Everybody gets caught up in a rapidly moving series of events and nobody feels like they can stop the situation due to circumstance beyond their control.

[Edited on August 6, 2017 at 8:51 PM. Reason : be Osman I not Mehmed VI]

8/6/2017 8:46:21 PM

Currently I feel that the same things that have made tesla great are kicking them in the junk whilst building the gigsfacotry...

They refuse to do things the tried and true way and are learning a bunch of lessons... the problem is they are reinventing wheel that's staring to look a lot like a standard wheel after seeing many delays.

Gonna be a big ass wheel but right now it still has wooden spokes.

8/7/2017 10:04:55 PM

this clown just asked for another 1.5B in junk bond money. how the fuck does this guy keep getting billions in gov't money to make the good and then billions in tax incentives to buy the goods?

8/8/2017 9:38:53 AM

^ I can see that you're not drinking the Musk flavored Kool Aid.

8/8/2017 12:51:00 PM

Mazda's breakthrough

http://jalopnik.com/how-mazdas-holy-grail-of-gasoline-engines-actually-work-1797795428

And here is a great read on Mazda's love affair with the rotary engine

http://jalopnik.com/mazdas-cosmo-turns-50-today-so-lets-all-celebrate-wacky-1795669343

8/13/2017 4:51:07 PM

I do wonder if mazda's ICE advances are too little too late with so many companies doubling down on electric. Ford, mitsu, subaru and chrysler are definitely lagging in that department so maybe they'll license the tech

8/13/2017 10:41:21 PM

You can bet that Toyota is considering to deploy the HCCI by licensing. They are very conservative though and would want Mazda to test it in the market first.

The real question is how much engine operation can run in HCCI. If it only works when you are cruising at 30mph, which is basically what the Japan test cycles are like, there will be a big real world fuel economy risk.

From some previous material I've read, HCCI needs the supercharger to achieve even modest levels of torque. Except it to be a lower output engine.

And of course you know I love the rotary, did I mention the presentation I attended from MIT/Mazda joint research on reducing oil consumption in an rx8?

8/14/2017 10:50:20 AM

Also, it will be a long time before people drive full electric trucks. If you can make a v8 or turbo v6 more efficient it will be a big benefit.

8/14/2017 10:51:44 AM

We don't have electrical infrastructure to handle everyone getting an electric car. Not unless there is some power timer feature where the shit only charges from 11pm till 5 am... we might have some capacity there

8/15/2017 7:23:49 PM

Here's something to consider when we talk about what the "ideal" air fuel ratio is on a gasoline engine for E10 (confirmed to be actual 10% ethanol) fuel.



X axis air fuel ratio from a wideband, Y axis are various emission concentrations before any kind of catalyst as well as the brake specific fuel consumption.

So about 14.0 is supposed to be ideal AFR for E10. Well what does that really mean? If you look at 14.0, you will see that it doesn't have lowest fuel consumption (CO2 emission), it doesn't have lowest NOx emissions, or lowest hydrocarbon. What it does have is about equal CO and O2 concentration which is what makes the catalyst operate correctly. If you go leaner, all the emissions are roughly equal or better, and the catalyst starts storing oxygen. However, if you stay leaner than 14.0:1 for too long you end up unable to convert NOx because there isn't enough CO emission.

In the real world driving around, the engine cycles rich and lean of stoichiometry (here ~14.0:1) to control the tail pipe emission.

8/23/2017 6:28:23 PM

this is pretty fascinating. I thought this was the goal of bridge porting.

I wonder if we'll see HCCI rotary

9/6/2017 1:40:51 AM

^

First, the guy is using an older pre-Rx-8 engine. All the stuff about gas leakage was worse on those engines. The Rx-8 improved the sealing in various ways, such as adding an additional gas seal (cut off seal). Any HCCI engine isn't going to be 100% HCCI, just like on the piston engines. I had started writing up a really detailed article about Mazda's HCCI based on the patents but I guess I lost momentum and left it in my draft pile.

4:10 He says the combustion pushes against the motion of the rotor. It depends on the spark timing, and the flame propagation. The leading and trailing plugs are fired separately. Usually the leading plug (the lower one in the housing) fires before the trailing, but at idle the trailing plug fires first -- on the Rx-8 engines. On the older ones the leading always fires first.

4:25 the apex seals jump because they have to balance spring tension and sealing, just like a piston ring gas seal has to balance friction and ring tension. Now, with those older designs they had to make a pretty big trade off. From the discussion I had with an MIT researcher who was working on sealing for Mazda (two published SAE papers on this topic), part of the problems with sealing has to do with warping of the housings due to basically 1970s era tooling and manufacturing processes.

It's the equivalent of trying to meet emissions with a first gen Chevy Small Block.

4:45 what he's proposing is called an LDR combustion chamber in a rotary engine (Leading dynamic recess I think). Mazda's already tested it over 20 years ago. They tested it on an Rx-7 engine but decided that it didn't work.



X axis is "fuel consumption" which I think corresponds to Air fuel ratio. Y axis is HC emissions. HC emissions are proportional to misfire and poor combustion. If you look at the curve with the triangles, it had the worst burning/highest HC emissions. So Mazda scrapped that idea.



Also, having two leading plugs and 1 trailing plug means a LOT of gas leakage. Leading plugs have a large hole, trailing plugs have a small hole. The only time Mazda used 3 plugs was in racing applications like the famed R26B 4 rotor engine in LeMans, but that is a limited application.

6:15

You can rev the engine to the moon, but does it make power there? The Rx-8 was a naturally aspirated vehicle with a 5 stage intake manifold system similar to what you would find in a Porsche GT3 or something. That's how it kept making power up to almost 9000rpm. Boosted engines have a lot more trouble making power that high without sacrificing low end torque or using a complicated sequential system (which Mazda did use on both 2 and 3 rotor engines). The same problem applies to piston engines.

Generally, high revving engines have trouble making good fuel economy, which is why we are seeing lower revving, long stroke piston engines. I suspect the next rotary will be lower revving than the Rx-8. Mazda just recently announced a rotary range extender for 2019, so that's a really narrow use case again. The 16X prototype had the equivalent a longer stroke. I wonder if it revved as high as an Rx-8.

6:25 his concerns about spitback were improved by using the side exhaust ports on a rotary, but the problem of leaking past seals remains. By eliminating the overlap in the Rx-8 Renesis engine a lot of the unburned gases were re burned.

The trade off of sealing and spring tension still exists: too tight of spring tension and you get too much friction and sticking seals. Too little tension and you get blow by. That's why Mazda is working on new housing designs and other ways to improve the leakage.

I just found the latest paper Mazda has published about gas leakage. The lead author did a presentation about oil consumption that I attended. I didn't realize he had published more. I just sent him an email, so maybe he will respond.

http://gasturbinespower.asmedigitalcollection.asme.org/article.aspx?articleid=2467297

7:55 interesting anecdote about other companies working on rotaries:

I talked to this old guy who works at GM who has access to their internal archives. He was telling me about the old Chevy Monza rotary which almost went into production in the 70s. GM was going to double the displacement to a 4 rotor and put that in the Corvette.

The Monza engine was all ready to go to production, but they kept having premature apex seal failures. Turns out, the displacement was so high (like a 2.6 liter 2 rotor, basically double Mazda's last 2 rotors) that the rotor had very long travel through each stroke. There was too much friction/wear over each stroke and the apex seals were dying. For a long time GM during testing thought it was an assembly or manufacturing issue, but at the last minute they decided to scrap the production program and not dump any more money into figuring out a solution. They didn't want to go back to the drawing board.

8:00 Rotary exhaust temperatures aren't that hot by today's standards. 1050 degrees C gas temperature is standard limit on monoscroll turbos now. One of the most thermally difficult engine designs is a gas engine with inboard twin turbos, such as what BMW and Audi have.

[Edited on September 6, 2017 at 6:11 PM. Reason : HCCI still needs to run on spark plugs a lot]

9/6/2017 6:10:47 PM

So this is a cool video (gotta love the New Zealand accent) that covers a lot of topics:

Spark timing vs torque
Knock limit on E85 vs E0/E10 fuel
Combustion speed of ethanol's effect on MBT spark timing at part load

First notice how he's doing this tuning exercise. He's got an unnamed vehicle (I can't tell what it is, could be something we don't get in the US) running a full replacement standalone Motec ECU. He's running on a chassis dyno that can hold speed and load. He's holding a specific manifold pressure, most likely by commanding the electronic throttle angle. He's also not running full WOT because that will generate too much heat. So with this kind of near steady-state condition he can do full spark advance sweeps. I'm not sure how the Motec's knock judgment is configured but it probably has some baseline signal processing preloaded that he is using.

In the video you will see at about 4:00 he's at part throttle with E0/E10 fuel and his minimum spark advance for best torque is about 32 degrees. Around 7:30 he does the spark sweep again and finds MBT at about 29 degrees on E85. This is because, at least at lower engine loads on port injected engines, E85 tends to burn a little bit faster than E0/E10. In my experience it can burn a little bit slower at higher load conditions which can amplify the number of additional spark advance degrees needed to find knock limit.

Notice too the hook shape of the curve. When he's far off from optimum spark, the incremental gain is high. 1 degree of spark gets a lot of torque. As he gets closer to optimum though you get diminishing returns. In the knock limited conditions he decides to get to take the spark where the ECU judges the knock limit and then set the actual value in the spark map about 2-3 degrees retarded.



The image above shows wide open throttle on a port injected non turbo engine running E85 with spark set at knock limit and then E10 87 AKI/92 RON gas (conservative tune). You can see the additional spark put in there, although it varies with rpm according to the knocking tendency of the engine and the combustion speed. Also note that the additional spark increases the cylinder pressures, even in a non knocking condition, as seen in the upper graph.

9/16/2017 11:55:30 AM

^he says it's a turbo FT86 early in the video

but DAMN! That MF got 312 lb*ft @ 2500rpm from a 2.0L @ 3 PSI. WTF?! Is it (mostly) the 12.5:1 CR at play?

[Edited on September 18, 2017 at 11:20 AM. Reason : not fair]

9/18/2017 11:20:01 AM

12.5:1 geometric compression, intake and exhaust variable valve timing. If you advance the intake cam enough you get a very high effective compression ratio. Remember that the 12.5:1 only applies if the intake closing timing is right around bottom dead center. A BRZ has something like a 50 or 70 crank angle degree range of intake cam phasing, so when it's at 0 (or rather negative, since it is an intermediate locking phaser), it's closing very late and the effective compression ratio is too low to make much torque at that speed unless you can boost a lot and go miller cycle.

Also if you retard the exhaust cam enough you get a very high expansion ratio and scavenging to spool the turbo. And then E85 helps with the knock.

9/18/2017 9:05:50 PM

I want to talk a little bit about OBD 1 vs OBD 2 level diagnostics and failsafes. The case study is the 92-95 Mazda Rx-7 (similar to 89-91 for this system), which were actually OBD I cars but had some OBD 2 level sophistication buried in them.

The feature I'm talking about is the oil metering pump, which basically drips crankcase oil on to the combustion surface of the aluminum rotor housing where the tips of the triangle slide. In this era of pumps, the pump is driven by a gear. Then a valve controlled by a stepper motor (a bunch of magnets basically) increases the flow from some minimum amount up to what is required to keep the main gas seal (apex seal) from overheating and getting worn out. There is a position sensor to let the ECU know where the valve position is.

So there are three diagnostic codes related to the oil metering pump:



So we've got 3 codes.

Code 20 and Code 26 are dumb "OBD 1" style codes. They basically detect an open circuit (something disconnected) or a short circuit (a positive/hot wire touching metal). Code 20 is for the position sensor and Code 26 is for the stepper motor.

Now, look closely at code 26 and code 27. What's the difference between the two? "Position Sensor Inaccurate" is in code 27. That's what's called a rationality monitor. It goes beyond just knowing whether the thing is plugged in or not or if the insulation has melted. It's OBD II level sophistication in a car that was designed in the 80s.

Rationality Monitors - Modern Examples

Here's a comparison with a modern car. If you unplug the o2 sensor in a modern car, it will throw a code. You can't just pop a resistor in there though and keep driving. It will figure out that it's being fooled. The ECU compares the measured value to an expected value to see if what the sensor is reporting is making sense. That's how a rationality monitor works.

Here's another example. On an OBD II engine that has a mass airflow sensor and no MAP sensor, the ECU can diagnose a bad mass airflow sensor by looking at a model. It can look at the modeled manifold pressure based on the mass airflow sensor and the modeled manifold pressure based on the throttle position. If the MAF based calculation is outside an acceptable range, we can trigger an error code, and the engine will go into some sort of failsafe condition. That's an example of a mass airflow sensor rationality monitor.

Active vs Passive Diagnostic Monitors

Now there are two main types of diagnostic monitors: passive monitors and active/intrusive monitors. Typically they are not activated until certain criteria are met, such as coolant temperature. Usually the OEM will try to run the monitor as little as it can get away with so as to not impact the customer, and CARB will try to push the OEM to run it more so that they can detect an emissions problem. If CARB gets mad enough they will fine the OEM and force a recall to make the diagnostic monitor run more.

Passive monitors are running in the background comparing the current value to the expected value. They don't interfere with engine operation. What I roughly described above for the mass airflow sensor monitor is basically a passive monitor. The throttle position sensor (especially on an electronic throttle equipped engine like a modern car has) is considered the dependable signal and if the mass airflow signal is off then it means that sensor is bad, or so the logic goes.

Active monitors directly interfere with engine operation to make sure a component is working right. They can also be called "intrusive monitors." In the case of an OBD II car, it might force the engine into a specific target air fuel ratio pattern ("square wave" pattern) and then confirm that the O2 sensor is reading as expected. Intrusive monitors can have an impact on emissions or driveability so they are used sparingly. I don't know for sure, but I suspect that on that old Rx-7 (and maybe the 2004-2008 Rx-8) the ECU might force the oil metering pump to step open and closed and confirm the sensor is reading as expected. However that could cause additional oil consumption, so it might only run once in a while, like when the battery has been disconnected and the learning values reset.

So with a rationality monitor you are storing some learning value in the ECU to make that comparison between expected and actual. Keep in mind that on old ECU's there were especially limited resources for doing extra computations and storing memory values. For example, I was talking to a guy who had reverse engineered some old 90s Dodge Neon era Chrysler ECUs. The 1996 OBD II ECU had basically twice the memory and processing of the OBD I ECU, even though the basic functions were the same, because OBD is so demanding on the controller.

Failsafe Modes

Getting back to Mazda's designed in the 80s: Since the oil metering pump is a critical engine component Mazda decided to devote memory and processing to it. The service manual actually instructs the technician to unplug the battery because there are stored diagnostic values, similar to a "keep alive memory" function you might hear about on your old 80s Mustang with EEC-IV controller.

If the component is judged to have failed by either of the passive or active method, the engine goes into a failsafe mode. Check out this chart from the Rx-7 service manual showing a list of trouble codes, the condition of failure, and the failsafe. Below notice that code 27 talks about a "sticking" sensor rather than only an open or short circuit. This meaning that the ECU is smart enough to have a rationality monitor.



Now think about the failsafe condition. Since the oil metering pump still dribbles a little oil even when the electrical system fails, the engine can run but is torque limited to keep the apex seal temperatures within acceptable range. Mazda actually published an SAE paper which describes the method they use in lab testing to determine the minimum amount of oil pump lubrication needed to run safely. Based on that info they could determine the torque limit when the oil metering pump is at its minimum position. Notice in the image above it says "basic fuel injection fixed" and "basic ignition timing fixed." It's limiting engine torque by changing spark and fuel (fuel cut etc). Only the stock Mazda ECU is smart enough to have a rationality monitor and a failsafe system for a bad OMP.

It's common to remove the stock Mazda ECU on these cars for performance purposes. Later owners will start hyperventilating wondering if their oil metering pump has gone bad unbeknownst to them. The failsafe mode ("limp mode") on the stock ECU is a highly engineered feature, not just an annoyance.

The service procedure in the manual has the technician checking how the position sensor voltage moves in accordance to acceleration and deceleration conditions. This is to confirm that something isn't stuck in the system such as a failed stepper motor/valve or a sensor that doesn't react as engine speed and load changes.

[Edited on September 18, 2017 at 9:35 PM. Reason : minimum oil necessary]

9/18/2017 9:29:56 PM

Here's a video from BorgWarner turbo systems in Asheville (NC). They take a stock manual transmission 2.3 Ecoboost Mustang. They bolt on a 48 volt electric turbo to improve spool. Then they put on a bigger turbo (one of their EFR performance aftermarket turbos) and use the eBooster to compensate for the lost low end torque and spool.



Mercedes has just put this device on their new inline 6 gasoline engine. One thing to point out is that the eBooster shown here gives a temporary boost for a few seconds and doesn't run continuously. So you couldn't lug it at 1500 rpm up a mountain in 6th gear for 10 minutes and have this thing working the whole time.

From a fuel economy and emissions perspective, if you use one of these eBoosters to help spool the turbo, you could change the variable valve timing so that you don't need to use as much overlap and scavenging, which is better for CO2 and NOx emissions.

9/22/2017 10:27:54 PM

I hate start stop systems, from the perspective of driving the actual vehicle. By start-stop systems I mean on non-hybrid vehicles where a starter is used. I think a lot of people agree with me. But there are a few interesting technical things added to vehicles to support the technology.

BMW for example talks about how they added a special coating to the connecting rod bearing (half of it, the side that contacts the rod). This coating helps maintain lubrication during all the starts. Since automotive engines use crank-driven oil pumps, there is limited pressure and flow during cranking to protect the bearings. With all the additional starting on a start-stop engine there are extra lubrication needs.

10/21/2017 1:47:43 PM

^^ I remember reading about the mazdaspeed rx8 rumors, saying mazda was working on a very similar system. Not sure there was any truth to it.

10/22/2017 9:58:37 AM

A lot of f work is being done but It comes down to cost of course, and what overall mix of pure EV, hybrids, and combustion engines are projected for the fleet. Automakers with fewer pure EV, like Mazda, may be more attracted to the idea.

10/22/2017 10:47:11 AM

Depending on your local rules, you will probably need some kind of emission inspection annually. I'm going to break these down into 4 categories, and focus on light duty vehicles with gasoline engines.

The first and most common type today is simply plugging into the diagnostic port on any 1996 or later (OBD II) car. If the car doesn't have any permanent DTC's, and the on board diagnostic monitors have all completed (usually they will let you have 1 not complete), you will pass. The whole idea is that the ECU should be smart enough to throw a code if any system that affects emissions significantly (usually 1.5 times the legal limit) is failing. The diagnostic monitors are different types of tests that the ECU runs on the engine, such as checking for misfire (which can damage catalysts), checking for a leak in the fuel system (which means more HC emission), and making sure the oxygen sensors are working.

The second type of test is what is called a visual inspection. The technicians are supposed to look under the car and check for any tampering that could impact emissions, such as deleting emission equipment. They may also check the functionality of parts such as an EGR valve. In practice, the enforcement of this varies drastically by locality. In places like California they are very strict, but in other areas with less emissions enforcement and less training for technicians the visual inspection doesn't get a whole of effort put into it.

Emission Measurement Tests - Steady State

The third type of test I'm going to talk about are steady state emission concentration tests. The idea is to hook up equipment that can inexpensively measure what's coming out of the tailpipe to find obvious sources of pollutants.



The state of NC for example used to have an idle-only tailpipe sniffer test for pre 1996 cars. This will catch completely gutted cats for example, but there's a lot of stuff that could slip by, like failed EGR valves. FYI, EGR valves were common in the 90s, then went away in the 2000s, and now are coming back with EGR coolers (current WRX has this for example).

Here's the last emissions inspection report for my current 1995 Rx-7 when it was still in California:



You can see the section on visual inspection, and then at the bottom are the steady-state emission test results. The vehicle is run on a dyno at speady mph and then the emission concentration is sampled ("MEAS") and compared to the legislated limit ("MAX"). You also get an average of what other passing vehicles got ("AVE"). Here you can see that my NOx was higher than average and my CO was lower than average. This indicates a deteriorating o2 sensor that has shifted the closed loop fuel control system lean. I haven't replaced this sensor as the leaner mixture results in a slight bump in fuel economy, and as you all know, Rx-7's can use all the help they can get in that department. Since it's an OBD I car it can only detect whether the sensor is unplugged/shorted, not whether it is deteriorated.

The 15mph and 25mph points have more load than an idle-only test. So they are going to be a better representation of the condition of the engine, such as whether it has high blowby, or a failed EGR valve, or a bad o2 sensor. Of course you need a dyno to test this, so the cost is much higher than idle-only testing.

Here's a handy guide (not sure where it came from originally) for addressing failed steady state emission tests:



These tests can be hard to pass for older cars, but you still are not actually testing emissions during an acceleration. But what if your area's testing includes actual transient acceleration tests?

Emission Measurement Tests - IM240 Warm Engine Transient Test



While it's certainly not an EPA or Euro cold start emission certification test, this type of test can be much more challenging for cars to past than an idle or steady vehicle speed test. It's also a much more expensive test for an emissions inspection station to run. The driver has to be trained to follow the pattern shown above on his monitor, and shift gears according to an indicator (if it's a manual). The equipment has to be able to capture the emission on a second-by-second basis and calculate a final result.

Here's an example of how a transient emission test is more representative of how a car really operates than a steady state.



I was helping this guy out with his failed IM240 test. He failed CO emissions but all the stuff mentioned in the chart above checked out. His o2 sensor was replaced, secondary air injection/smog pump worked fine, his cat had been replaced, all that stuff was good but he still failed. So I asked him if he had any chart showing the emissions on a second-by-second basis, and it turns out he did:



I'm using a bigger image here so you can see what's going on. On the x axis is time. On the Y axis you have the vehicle speed and the specific emission plotted. This is a typical plot used in vehicle emission development, so I knew how to help.

You can see I put a red box around three areas where the CO emission trace looked not like a spike but "rectangular." That means the tailpipe is producing so much emissions it is maxing out the equipment or the scale of the graph. To meet the emission limit, which is actually pretty generous compared to the equivalent of what a 2017 model year new car would ever have meet, he needed to fix those really big spikes.

When you deal with emission measurement on a second-by-second basis, you have to assume some kind of delay. So if you look at the 3rd "box" from the left, the big one, you can see it's depicted as a few seconds after a major acceleration from 30 to 45 mph. That's just the delay in the measurement. The car was clearly going into boost and blowing a ton of CO out the tailpipe.

So the owner of this Rx-7 recently got it to pass this IM240 test he had been failing for CO. There were two things he did to pass.

1) He turned the boost down. His car was close to stock, but the boost control system had some issues, and it was running more boost than stock. More boost-->more air-->more fuel with rich mixture due to load-->more tail pipe CO.

2) He asked the inspection station to give him a less aggressive driver. Yes, that's the dirty little secret of emission testing like this. It's highly dependent on the driver, at least for CO emission. A "Good" driver, from the perspective of whoever is trying to make a vehicle pass, follows the trace carefully and doesn't out-accelerate it much. The EPA knows all about this, and they have ways of testing for worst case drivers, but Joe's Inspection Station doesn't get into that kind of intricacy normally. They just want to get you in and out as fast as possible.

10/28/2017 1:35:01 PM

So Ford announced the 2018 Mustang Ecoboost for Europe a few months ago. Power is down from 313ps to 268ps (ps is similar to horsepower). I don't know all the differences between model years but my understanding is that it is primarily caused by the addition of a Gasoline Particulate Filter (GPF). This is because Europe regulates "Particulate Number" , instead of "Particulate Mass" like the US. So if you have a gazillion tiny particles you can't certify the engine in Europe. In the US you can as long as they don't weigh too much in total. If PN regulation is strict enough it requires a particulate filter like a diesel. So now you've got a big restriction in the exhaust, like adding an additional catalytic converter, that can also get clogged up.

Recently NGK published a paper (Ogata, "Particle Number Emission Reduction for GDI Engines with Particulate Filters," 2017) about the impact of a particulate filter on a 1.4L and a larger engine which I suspect is a 2.0 . They didn't mention the manufacturer but it is likely something they can get easily in Europe. In my opinion it is very much a rosy scenario they are describing where CO2 emissions (fuel economy) are barely affected, backpressure is much higher and yet barely affects horsepower.



You can see their measured data showed a 30% increase in backpressure with a fresh GPF. The study said it was a 170kw engine which is 227 horsepower. I suspect that on a higher output engines (especially high output boosted 4 cylinders) the impact gets amplified significantly, as we've seen with the European 2.3L Ford engine. Temperature limits of the GPF probably don't help horsepower. There's a bunch of stuff in here about test cycle fuel economy that I won't get into, but I doubt they measured the impact of regeneration (forcing the GPF to burn off the soot when it gets too clogged) on fuel economy.

I have no doubt that in some scenarios the impact of a particulate filter on fuel economy and power is negligible. But I'm annoyed when a company that stands to profit from selling various GPF related parts says these kinds of things and less than a year later we see a Mustang adopt a GPF and lose so much power.

[Edited on December 15, 2017 at 7:56 PM. Reason : .]

[Edited on December 15, 2017 at 7:57 PM. Reason : .]

12/15/2017 7:55:56 PM

Recently I was asked about the process by which knock sensors and knock control settings are configured during development of a gasoline engine. The basic process consists of running an engine in a lab in both knocking and severe knocking conditions, then looking at the characteristics of the vibration generated.

Take a basic naturally aspirated engine and run it at wide open throttle at a lower engine speed (say 2000rpm or under) continuously in an engine lab. I hook up lab grade knock sensors (Basically accelerometers) and hook up cylinder pressure measurement to it (drilling holes and inserting pressure transducers in the combustion chamber). The accelerometers give me a voltage, and the cylinder pressure sensors will look at the cylinder pressure spikes during knocking to give me a "knock peak."



The image above has crank angle degrees ATDC firing on the X axis (axes aren't fully aligned but you get the idea). It shows cylinder pressure in bar (the "hump") and then the calculated knock peak at each crank angle (the wavy line). The highest knock peak of each cycle can then be overlaid with the accelerometer voltage in a scatterplot:



X axis shows the knocking peak. Y axis shows the accelerometer voltage. The horizontal lines show background accelerometer voltage in non knocking, and the scatter plots show individual combustion cycles in a severe knocking condition. This kind of measurement and plotting is what generates data for the ECU to determine the noise / knock thresholds. You might take all the statistically gathered knock sensor/accelerometer data and use it to set a threshold (10 times the standard deviation of the signal during knocking or some other such general rule of thumb).

Note that you can do this kind of testing without using cylinder pressure. You basically replace the X axis of the graph with some other kind of knocking severity metric, typically a subjective rating (somebody rates it 1 to 10 or whatever based on the sound heard through a microphone or other such device). In that case you would use a variety of spark timings that are both very advanced and very retarded.

Keep in mind that knocking noise generally increases at an exponential rate with rpm:



Also note that the different colors represent different knock sensor locations. Some locations are more sensitive than the other depending on rpm and the particular cylinder. That sensitivity is based on how far away the sensor is from the cylinder in question. It can also be affected by other items in the way such as cooling water jackets and rotating mass such as timing chains or valve trains. Here is an image of a modern turbo direct injected engine that uses 2 knock cylinders for better accuracy: one for the front two cylinders, one for the rear two. Notice that they are near the center of the block to reduce exposure to noise.



In order for the knock control system to work the ECU has to look for a particular frequency (unless the frequency is baked into the design of the sensor instead). To do that, a fast fourier transform is used. This is similar to what you see on a sound equalizer.



The actual frequency is going to be a compromise due to the varying noise profile of the engine. Also, signal amplification can be used to increase or decrease the signal under specific conditions due to the known sources of noise mentioned above.

1/5/2018 11:12:53 PM

Thanks for creating and continuing to update this thread.


This is easily in the top 10 for most educational threads I follow.

1/9/2018 10:14:32 AM

^ Thanks for the kind words. It keeps me motivated to continue posting

So a friend of mine picked up a brand new Aprilia RSV4 motorcycle recently. I'll admit, I had never even heard of Aprilia before, but I'm not super knowledgeable about motorcycles, and it's less of a household name in the US.

He started telling me about the wheelie control system and how he can ride the thing in cold conditions more safely because of the super customizable traction control system and I was pretty intrigued. Here's a pretty cool video explaining how the traction control portion works:



Electronic throttle and spark retard based traction control are certainly not new ideas. However it seems to me that the best and most unusual thing about it is the fact that it can be dynamically adjusted in such fine increments.

Quote :

"ATC Aprilia Traction Control The exclusive traction control system developed by Aprilia to get maximum grip out of any type of surface and to give the rider greater confidence while at the same time improving safety. Two speed sensors allow the control unit to determine bike speed at any given time and use a sophisticated CAN-bus communication system to interact with the engine ECU. If the rear wheel is found to be rotating faster than the front wheel, the control unit determines the slippage of the rear wheel and interacts with the ignition advance and injection timing systems to limit the amount of torque produced by the engine thereby preventing loss of grip. The 8 different settings (3 on the version not combined with the APRC package) can be selected for different types of road surface and riding style. They allow the rider to extract the utmost performance from the bike whether on the track or provide enhanced safety on the road."

There's more info on the wheelie control and other systems here:

http://www.aprilia.com/en_EN/technology/aprc/

1/14/2018 2:05:14 PM

In a modern gasoline powered car, basically something from the last 30 years, how does the stock computer handle the load of the air conditioning compressor? Let's break it down into steps:



1. Driver hits the A/C button. For simplicity's sake, let's say there isn't an automatic climate control.

2. ECU receives the command and enables a delay before switching on the compressor. If I hit the A/C button and the compressor immediately comes on without any intervention from the ECU, the idle is going to drop due to the additional torque required by the engine.

During this delay, the ECU will command additional idle airflow. This has evolved over the years. In the really early days (early 80s lets say) it was an ON/OFF vacuum actuated valve with a solenoid. Later the additional idle airflow will come from increasing the duty cycle of an ON/OFF (duty controlled) idle air control motor. Here's a Ford style one you might find on an old 5.0 Mustang:



That passage cycled ON and OFF at different % ratio, with more ON time associated with higher airflow. It could also be a pintle style, stepper motor type, similar to what you would find on a Gen II small block (LT1) Camaro.



That valve opens and holds to a specific position based on the operation of magnets (stepper motor). Cars from the past 10-15 years are going to have an electronic throttle which doesn't use these kinds of external valves. Inside the ECU there may have been some kind of table in the ECU that converted duty cycle/position into a physical airflow, taking into account the throttle position and the size of the throttlebody.

So the "triangular airflow" is big to account for the initial load of the A/C cycling on, and then tapers down so that the rpm doesn't surge.

3. The compressor cycles on according to the ECU command. The delay has ended, the "triangular airflow correction" is in process, and now the compressor can come on without the engine possibly sputtering or stalling. Also note that spark could have been adjusted depending on the control system. Spark responds faster than airflow and is often coordinated with electronic throttle or idle air valves.

4. The A/C is on and the additional airflow is now at a near steady value. There can be some small adjustments made by the feedback system to keep the engine at a slightly higher target idle speed.

5. The A/C "OFF" command comes from the driver, and we begin the inverse process of disengaging the compressor.

6. During the delay period the airflow is going to be reduced, and possibly the spark retarded. We want the rpm to decrease as the compressor cycles off, so that the rpm doesn't surge.

7. Compressor is now off and idle stabilizes at a lower setpoint.

1/21/2018 9:56:20 AM

Perhaps it's been discussed in this thread already...

How does EGR lower combustion chamber temperatures? As I understand it, you are recirculating hot exhaust air to mix with fresh intake air... Wouldn't that increase the IAT that's being fed into the cylinders? Or does the recirculated exhaust air have a post combustion chemical characteristic that helps provide a less volatile fuel air mixture, thus reducing the chances of knocking? Would like to know the science behind this; I understand what and EGR does, just don't understand why it works like it does.

Is it that the exhaust gases reduce the amount of Oxygen available for combustion, thus the fuel:air is a bit on the rich side? Lean mixtures run warmer, thus a little more fuel in the mixture helps keep things "cooler?".

[Edited on February 13, 2018 at 10:13 AM. Reason : .]

2/13/2018 10:10:14 AM

^ I'll get to your question... I've got a post in my mind I want to write down while I still have inspiration/motivation.

Airflow Measurement Methods

There are basically three ways we can figure out how much air is entering the combustion chamber of a spark ignition engine.

1. Directly measure it with a mass airflow (MAF) meter in the intake air stream, understanding that there is some time delay between when the air passes through the sensor and actually enters the engine. The ECU will have a function showing frequency or voltage from the sensor vs actual airflow. Here is the transfer function for an LS1 Camaro:



2. Calculate the airflow based on pressure sensors, primarily the Manifold Absolute Pressure (MAP) sensor. To do this the ECU has to be pre programmed with look up tables and equations that explain how "full" the cylinder should be at a given engine speed and manifold pressure. The simple method is to have one basic Volumetric Efficiency table which helps the engine calculate airflow based on speed and manifold pressure. Again here is a table from an LS1 Camaro (from the stock ECU's tune in the HP Tuners software):



3. Calculate the airflow based on the throttle angle, called Alpha-N. In simple implementations, such as what you find in most aftermarket ECUs, these consist of look up tables based on engine speed and throttle angle. So if I'm at 2000rpm and my throttle position sensor is 15 degrees, give it this amount of spark advance or this injection pulsewidth. The more sophisticated OEM level implementations convert throttle angle to actual airflow. It does this by knowing the displacement of the engine, and the relationship between opening area and airflow. This image demonstrates the concept of calculating airflow through the throttle:



Putting it all together: here is an image showing the basic concept of a manifold filling model (from Bosch documentation on their control system called ME7). You don't have to think too much about the math and equations. Just understand that MAF, MAP, and Alpha-N each are part of the puzzle shown in the diagram, and that you can use more than one method at a time.



For example, GM uses Mass Airflow and Manifold Absolute Pressure calculations. They sort of check each other, and manifold absolute pressure can be used in transient conditions. Some Japanese OEMs will use Mass Airflow and Alpha-N instead. Basically the Alpha-N calculations are used to correct mass airflow readings under specific conditions.

Airflow Measurement Methods with Variable Valve Timing

This is where shit starts to get real. Now a few simple math functions or look up tables aren't going to cut it if we want really strong accuracy for best fuel economy, emissions, and transient response (a drag car is another story). Take a modern Dual Overhead Cam engine with variable intake and exhaust camshafts:



The exhaust camshaft can retard from its base centerline (the crank angle at which peak valve lift is achieved), and the intake can be advanced (and in many cases now retarded) from its base position. Now one main VE table with a few compensation factors isn't going to work, and even using a MAF sensor still needs some kind of compensation for best accuracy and response (throttle angle or MAP sensor based compensation). Chrysler published a paper back in 2004 that explains the challenge.



And this approach of multiple look up tables or equations is what we see most manufacturers use. On a Coyote 5.0 Ford engine there are a bunch of different look up tables and math functions depending on the variable valve timing position.



So for example, one math function for the VVT set to 0 on intake and 0 on exhaust, one that might be for intake fully advanced and exhaust at 0, etc. The ECU then has to interpolate among these to calculate cylinder airflow filling. These tables are basically populated by engineers and technicians running engines in a lab, taking a gazillion measurements, and filling out the tables using Matlab scripts with a little massaging along the way.

Chrysler's Unusual Approach for VE calculation on VVT Engines

Chrysler/Dodge/etc took a different approach to calculating VE. They used a form of machine learning called an artificial neural net. I'll be the first to admit that I'm not an expert on machine learning, but I do get the concept. Basically, they run the engine in the lab and collect a lot of data. The image below shows the basic process:



This shows how the engine dyno had to collect a bazillion data points for many combinations of valve timing angles, manifold pressure, and engine speeds (more efficient model based methods exist today):



Then using learning algorithms in Matlab, they "train" the engine to calculate airflow based on the input factors. In the Artificial Neural Network the ECU doesn't know "why" physically there is a given amount of air entering the engine. It just knows how to calculate it, provided that the persons tuning the algorithm sort of turn the right knobs (# of nodes, weighting factors, etc). They look for smooth calculation maps when possible. Then they compare calculated airflow to measured airflow, measured either by lab equipment or by the vehicle MAF sensor:



The above image shows how engine air flow varies with the position of the intake cam centerline. A higher number indicates more retarded intake valve timing. The red dots show measured airflow in the lab, and the blue lines show calculated airflow from the artificial neural net used in the ECU. The dataset name "ANN 4-8-8-1" represents particular parameters of the machine learning algorithm that were adjusted by whoever was tuning it to make the calculations work.



A similar concept is shown above, except based on rpm, manifold absolute pressure, and cam position.



Finally, you can see a transient manuever performed by the engine dyno. The chart is time-based. Before the 4 second mark, the ECU commands a sudden throttle opening. The commanded VVT positions change according to some look up tables. The ANN smoothly predicts a change in airflow inside the cylinder. The mass airflow sensor, located in the intake piping, has some time delay and also pulsation effects causing some oscillations.


So basically VVT makes things a lot more complicated. The more sophisticated controls used pretty advanced models and labor-intensive tuning techniques during development for more accurate prediction, but less sophisticated mass airflow sensor based controls exist as well *cough* Subaru *cough*.

2/17/2018 11:03:32 AM

Quote :

"How does EGR lower combustion chamber temperatures? As I understand it, you are recirculating hot exhaust air to mix with fresh intake air... Wouldn't that increase the IAT that's being fed into the cylinders? Or does the recirculated exhaust air have a post combustion chemical characteristic that helps provide a less volatile fuel air mixture, thus reducing the chances of knocking? Would like to know the science behind this; I understand what and EGR does, just don't understand why it works like it does. Is it that the exhaust gases reduce the amount of Oxygen available for combustion, thus the fuel:air is a bit on the rich side? Lean mixtures run warmer, thus a little more fuel in the mixture helps keep things "cooler?"."

EGR can get complicated because, in terms of what's typically been in production, there are a few kinds: External hot high pressure EGR, External cooled high pressure EGR, External cooled low pressure EGR, and internal EGR. Here are some helpful images to conceptualize each type of system:



This is your old school EGR system from 80s and 90s port injected gasoline engines. It uses a vacuum to regulate EGR flow rates on an engine, and it might have a temperature or pressure sensor to help control flow or detect valve failures for OBD.



This is early-mid 2000s internal EGR for a 4 valve, dual overhead cam engine with intake variable valve timing. My mom's 2007 Corolla had this along with electronic throttle. You can see that the intake valve is opening at 40 degrees BTDC intake while the piston is still rising in the exhaust stroke.



This is the high pressure cooled EGR system from a 2nd generation Prius, also in the mid 2000s. It was one of the earlier gasoline engine to adopt cooled EGR. It's high pressure because it comes from the exhaust manifold before the catalyst, goes through the cooler, and flows to the intake manifold. I'll get into more detail about that later on as it relates to the original temperature related question.



This illustrates a low pressure cooled EGR system that you would find in a few engines such as the facelifted Juke. It's low pressure because it brings exhaust from after the catalyst to the inlet of the turbo, and is primarily useful for boosted operation.



There are two main effects from EGR. The first is the dilution effect, and you get that from any of the above categories of EGR. I'll admit I don't know all of the details on a chemical level, but basically the act of diluting the combustion chamber with noncombustible gases will slow down the burn and lower the combustion temperatures.

The second effect of EGR is actual cooling due to the temperature difference between the exhaust gases and the temperature inside the combustion chamber. You need cooled EGR to get a significant improvement. Cooled EGR is big on diesels, but is also on gasoline engines now. A current WRX uses a high pressure type (good for medium loads) and a Juke uses a low pressure type (Good for boosted operation).

The image below shows the impact of cooled EGR temperature and flow rate (expressed as basically % of combustion chamber gas) on max spark timing and brake specific fuel consumption for a 2nd gen Prius.



The charts both represent a steady speed and engine torque point where a Prius could be operating during normal operation. Recall that the Prius uses a power-split type hybrid architecture where a simple planetary gear set controls combustion engine and electric motor speeds. There are plenty of Youtube videos and articles about it, but suffice it to say that the Prius and similar full hybrids can be optimized to run the combustion engine in a narrow range of operating points where it would be most efficient.

On the X axis is the temperature of the cooled EGR gases which are flowing into the intake manifold. It's likely that these were controlled by a lab-based cooling system. For the top chart, the Y axis is spark timing. There are two datasets: one showing an EGR rate of 11% and one showing it for 15%. Basically, the higher rate indicates that the EGR valve is open more and thus more dilute gases are flowing.

Look at the solid black dots representing 11% EGR. As the EGR temperature decreases (more cooling applied), the spark can be advanced. This primarily reflects the reduced knocking tendency. A 2nd gen Prius engine runs high compression ratio with port injection and regular fuel, so it is prone to knocking. If you look at the corresponding solid dots in the lower chart, you will see that brake specific fuel consumption decreases (lower means more fuel efficiency) as the EGR gets cooler.

Now, look at the black circles representing 15% EGR. The same temperature trend holds: cooler gases means more spark advance and lower BSFC. BUT notice how the overall spark advance is higher at 15% EGR than at 11% EGR? This is due to the dilution effect. The additional dilution slows down the burn, requiring the spark plug to be fired earlier, even while running the same EGR temperature.

2/17/2018 2:58:13 PM

100k views

3/23/2018 6:13:59 PM

Any ideas for topics for future posts?

5/8/2018 5:37:45 PM

Are direct injection injectors fired like a typical injectors ?

Aka how long will it take for stand alone ecus to be developed for the 3rd gen coyote engines.

They are currently cheap in junk yards

5/9/2018 6:45:57 PM

do cross plane cranks really have a mechanical advantage for cornerspeed based on power delivery? yamaha has claimed this in their marketing material and their race machines tend to have much higher cornerspeed than any others in the class

5/10/2018 12:04:01 AM

on page 6 I covered direct injection vs port injection, including high pressure fuel pumps. They are more complicated to control and more expensive. As for standalone ECUs, Motec can run direct injection. They already have a system for the BRZ for example. I'm confused though. The 3rd gen Coyote is the one that came out for the 2018 Mustang isn't it? How is that cheap? The earlier Coyote engines are port injection only.

As for the video, they are saying basically that their cross plane V8 has less torque fluctuation/vibration, which in my understanding is the main argument for cross planes in general. I'm not sure what they mean by throttle response though. Do they mean that they have a better low end torque? People throw those terms around a lot.

I would want to see torque curves, volumetric efficiency curves, and torque fluctuation (standard deviation of brake torque let's say) between their cross plane engine and a comparable flat plane engine. Then I'd be interested in seeing the rate of torque rise in a given manuever.

I suspect they figured out a way to make a cross plane engine cheaper than a flat plane to achieve whatever goal they had, and then came up with an engineering rationale for what was essentially a business decision. That's pretty typical. It might even be more expensive on a component level but cheaper on a system wide level because then they can reuse some tooling for some other part of the bike instead of redesigning xyz things that would ultimately achieve the same target but in a different way. "We couldn't change this or that, so we figured we could switch to a cross plane crank to get there for less than money"

5/10/2018 5:38:42 PM

They are cheap because no one can shove them into a hot rod yet.

5/10/2018 8:09:30 PM

what is this actuator assembly on a Cayman S transmission?


[Edited on May 10, 2018 at 10:28 PM. Reason : ]

5/10/2018 10:15:32 PM

https://www.americanmuscle.com/frpp-coyote-engine-control.html

If you want to swap a Coyote 5.0 into something, just get the 1st gen Coyote and then buy the kit Ford performance sells. It comes with an unlocked engine controller, accelerator pedal, wiring harness, and is already tuned to work with return fuel systems. That's a way better deal than spending thousands on a higher end ECU (because 8 cylinders cost more) and tuning.

As for the transmission above ^ , it looks like an automatic with mechanical shift linkage to me. A large portion of automatics today are shift by wire , but 10+ years ago you still had cables shifting to park, neutral, drive, etc.

5/10/2018 10:53:35 PM

That's a 6-speed manual

I'm having no luck looking through Porsche forums. I just want to know for curiosity's sake. I was thinking some sort of gear change damper to take the shock out of quick shifting

5/11/2018 5:57:36 AM

^ are we talking about the same thing? can you circle it? a manual transmission needs a shifter linkage too, unless the shifter goes directly in it (only on longitudinal transmissions, and typically older ones).

I received this question by PM:

Quote :

"Why is it that pushrods motors (eg Small Block Chevy LS motors) make more low end torque, at the expense of higher rpm bandwidth, compared to an OHC setup"

The short answer is, pushrod engines are 2 valve engines and don't flow as well as 2 valve engines. To make high rpm power you need airflow (200 lb ft at 2000rpm needs much less air than 200 lb ft at 6000rpm). It increases exponentially with rpm.

These pushrod motors tend to be larger in displacement so it's easier to make low end torque. High displacement, low airflow capability will generally make good torque but low power and low specific output per liter.

I recommend you check out these videos:

5/11/2018 5:21:30 PM

It’s the 2018 truck motor that’s cheap, because there is no control pack for them, they haven’t had many issues and the aluminum bodies total out trucks relatively easy there is a supply with no demand. That and I just wanted to learn while I work.

So next. What makes a coil more powerful (a spark less likely to blow out) higher Ampacity or a higher step up voltage ?

Since most aftermarket coils suck which oe coil is the strongest? Are the ones meant to fire 2 plugs, like the hemis and the 6.2 ford higher powered than the others?

7/12/2018 8:41:20 PM

What kind of ignition system are we talking about here? Modern individual coils?

7/28/2018 4:05:56 PM

Some of you may be familiar with one pedal driving on battery electric vehicles.


(Bosch Automative Handbook 11th edition)

Essentially when the accelerator pedal is not depressed, the vehicle naturally brakes. You need to depress the accelerator enough past the middle "coasting" part of the range, into the "acceleration" part. Then you can gradually let off to engage more and more regenerative braking with respect to pedal position. Typically at low vehiclespeeds the actual friction brakes kick in. What can be a little annoying is that max regenerative braking varies with the temperature of the battery.


(don't remember where I got this, on a forum somewhere maybe)

That's a chart showing the maximum regenerative braking power on a Tesla Model S after some software update. I don't even remember the details anymore, but the point still stands: the battery needs to be warm to get full regenerative braking, and the deceleration force from releasing the pedal can vary if the vehicle doesn't compensate with friction braking. So it comes down to the way the pedal mapping was tuned to account for the temperature differences. Of course you could heat the battery to get better regenerative braking to get more efficiency that way, but that would probably be a net energy loser.

Personally I like a medium-strength regenerative braking feel but I can't fully get into one pedal driving. I think I'm too used to coasting with my foot off the pedal and I find holding the neutral position tiring.

6/28/2022 2:13:50 PM

So I always wondered why forward facing ultrasound sensors (usually for parking assist) tend to get sort of weird phantom obstacles at low speeds. Apparently it's a known issue, and part of the solution is to have the sensor intentionally operate more in the horizontal direction than the vertical direction.






[Edited on June 29, 2022 at 9:45 AM. Reason : .]

6/29/2022 9:44:56 AM

Yesterday was a pretty good day for my solar production. I have a Panasonic system rated at about 5 kw, with Enphase inverters, running to a single Tesla Powerwall (around 13.5kwh nominally).

So as background most of my major appliances are gas (dryer, stove/oven, furnace, water heater), plus I have a natural gas grill on the patio. I've been able to look at the the major energy consumers by taking live readings of energy consumption. If you're trying to minimize grid usage by running off solar and battery it's important to consider energy used (KWh), but also instantaneous energy flow (KW). Where I live the power company basically gives you half as much money to sell them power as you pay to buy it. So there is an incentive to be off grid.

Here are a few things I've observed as far as power consumption in units of flow. The flow is important because solar energy is also a flow. You want to stagger your energy consumption throughout the day if you can, so that you are drawing from the solar panels first, then the battery if it has some juice left, and finally from the grid if needed.

For context, house is only a year old, 2 story, ~1500 square feet not including finished basement and attached garage, single HVAC zone with forced air (A/C and natural gas furnace).

My A/C: about 2.2 KW when it's running.
Heat: about .5 KW - basically the fan itself uses around 500 watts, even if it's burning gas for heat.
Dishwasher - around 1 KW depending on the cycle
Clothes Washer: maybe 800 watts? Trying to remember. It's a high end large unit.
Clothes dryer: a couple hundred watts? that doesn't include the heat itself which is gas
Lights: not a lot. A few hundred watts if a turn on basically every light in the house. LEDs and CFL bulbs are way less than incadescent

Car charging: More on that later, can be up to 8kw if I want to, but I typically throttle it to 2 or 3 KW, running 240V and moving the amps between 8 and 12 amps.



There's my consumption yesterday. It was a sunny, not very humid day, max temperature around 78F (~26C) That big spike starting at 10:30 AM is the car charging, and that extra little bump on top later is the dishwasher kicking on. That little sawtooth hump overnight I think is the refrigerator cycling. Around 6PM you can see the A/C kicking on as my thermostat schedule cools down for comfortable sleeping upstairs. Part of the issue is that since I have a single zone and the main bedrooms are upstairs, I need to kick it on for comfortable sleeping upstairs even if the ground and basement are comfortable.

6/29/2022 10:18:03 AM

I really wish that the solar system I had installed on 4/9 was online. But it's not, because Duke still hasn't come to inspect it


f*cking assholes

6/29/2022 10:23:16 AM

Now let's look at my solar production as a flow:



You can see the curve is pretty smooth, and the little noisy jagged parts are just a few clouds rolling by in the afternoon. Peak production is during the middle of the day, so it makes sense to run my most energy intensive processes during the day such as car charging and major appliances. That's a bit counter intuitive because time of use power plans encourage you to do the opposite, by tapping juice at night when the grid is under less strain. Here's the energy consumption again but showing its energy source, directly from solar or battery. No grid usage is shown here because it was a mild but sunny day.

6/29/2022 10:25:44 AM

Now let's look at some sum totals.



Using the micro inverters' reporting system, I can break it down by panel. The side of the house is south facing and is pretty clear of obstructions, so it gets pretty good power throughout the day. The front of the house is east facing and primarily gets power in the morning and early afternoon. Its production falls off in the afternoon, but I get power through the back of the house. You can see that two of the panels on the back aren't producing as much, around 1.8kwh each. Those are partly obstructed by trees. I had a bunch of tress trimmed and removed to improve the solar production, but I don't want a completely tree-less back yard and I can't just cut down my neighbor's tree.

For the month of June (up through June 28th) I have generated 111% electricity of what I have used. Keep in mind that is not running A/C in the morning, taking short trips and mostly working from home (or charging at work when I do go in). Cooking, drying clothes, and heating water are done off gas, although some clothes are air dried. I also have relatively fuel efficient gas cars in the household used for some cruising around and commuting.



For the month I'm 89% self powered. The panels only make so much power, and the battery only has so much juice. My electrical bills are around $12 a month right now with some fixed fees built into that.



I'm sure if I sat down and did a more full energy analysis looking at natural gas and gasoline consumption (converted from gallons into KWh) it would look less self sufficient. In the winter with less sun and more heat running it will be worse.

[Edited on June 29, 2022 at 10:38 AM. Reason : .]

6/29/2022 10:37:55 AM

So the way these lane detection systems typically work is that they scan horizontally across the oncoming road and look for changes in the light given off from a surface. When there is a distinct change they can detect the beginning and end (horizontally) of a line line. That's why when you get merging areas, broken pavement, etc they at best guess where the lane is supposed to be. They can also just disengage, or start ping ponging and/or swerving.

7/12/2022 12:05:33 PM

So I recently went to the Auburn Cord Duesenberg Museum in Auburn, IN. Those aren't exactly household names these days but they were independent brands that were mostly around in the 1920s and 1930s as notable automotive manufacturers. Their headquarters was in Auburn and you can see their design office. It's worth checking out if you're in the area.

They had a display of their front wheel drive model Cord models that were around in 1936 and 1937. You can see the engine is longitudinally mounted with the transmission in front of it, with half shafts coming out. Today when we think front wheel drive we normally think of longitudinally mounted engines, and a lot of the traction problems that come with torque steer and having weight shift to the back of the car.

Back then it was more about having extra weight over the driven wheels to help traction in inclement weather. I remember my dad telling me how he got his first front wheel drive car (1970s? 1980s? don't remember). The biggest thing he noticed was how much easier it was to take off in the snow. Having owned a few non traction control front and rear wheel drive cars, I can agree with that.

7/12/2022 12:21:51 PM

Holy fucking shit. You ARE alive.

9/25/2022 4:36:38 PM