Horrible engine noise in P, D, B, and R, but not N

Why, yes ... yes, it should. It, um, should be showing more air going in when the throttle opens more or the engine works harder, and less when it doesn't.

There's a possibility that "Car Scanner on iOS" just doesn't know how to get your MAF reading, but that seems a remote possibility. MAF reading is a standard PID that any self-respecting car scanner should be able to read.

That leaves the possibility that your ECM is getting a fixed unchanging reading from the MAF sensor. That would mean your ECM has no clue how much fuel to inject to make a combustible mixture, which will result in horrible noises (which you have) and codes about misfires (which you have) and about poor engine power (which you have, P3190, which has the MAF sensor as one of its suspect areas) and failure to start (P0A0F, likewise).

A voltmeter on the MAF sensor output would answer some questions the audience following at home surely have.

It's odd that you don't have (or haven't reported) any of the codes specifically reserved for a wacked-out MAF sensor (P0100, P0102, P0103). But those codes are looking for a sensor output that's either below 0.2 V or above 4.9 V, so clearly out of range.

Yours just looks like it's humming along at a never-ever-changing, oh, 250 kg/hour, eyeballing your graph. Toyota gives their ranges in g/sec, so 250 kg/hour is around 69 g/sec, which is right in the expected range of MAF values. So no P010x codes.

It sure isn't useful though. If your ECM had a way of knowing how much air was going in, your engine would run better.

There's a cute note in the repair manual next to the expected range (1.0 to 270.0 g/sec). The note says "The value must be changed when the throttle valve is opened or closed."

 

Here are three shots of my Prius maf and rpm using Car Scanner for IOS. Is it possible your scale units are set wrong in the app? (Edit: These were taken with "forced charging" in the driveway - zero speed - first two with changed throttle position)

 

Hmm ... while the range given in the manual is from 1 g/sec minimum to 270 g/sec all-out, if rjparker's readings were all around 7 g/sec (presumably idling, or just revving the engine in the driveway), then a fixed 69 g/sec reading out of the sketchy one would look to the ECM like a moderate respectable load. Well, about ten times what it should be at idle. Which would make the ECM inject a lot more fuel than it ought to.

 

While I was home from college one summer, I got an unexpected call from a classmate from a different state. She was in her VW Bug with her brother and they were making a multi-state road trip and the Bug had just then started shaking like the fenders would fall off, and they looked in our college directory to see if they knew anybody in the state they were in, and found they were only half an hour from me. (Our college gave out directories with our home addresses and phone numbers, how freaky is that?)

So they pulled into my driveway in full cartoon jalopy mode shaking and lurching and jumping and banging, and I was like "this car I'm driving I pretty much built, but I'm sure there's nothing I can do for yours, we'll go straight to the bug place". And I got in my car and we made a slow comical procession down to the specialty mechanic.

And he put on some new points and plug wires and it was just fine. Which I could so have done by myself, but I had let myself be too alarmed by all the racket it was making to even use my head and think it might be simple stuff.

 

I love the process of elimination here. This is definitely how it should be done, and sounds like you really know what you're doing with diagnosis. Way better than I could ever do at diagnosis.

 

I wonder if something is labeled funny or defined funny in that app.

The car has a sensor 1 and a sensor 2, and they should both be giving sensor readings, and then there should also be a fuel trim %, which is another number computed by the ECU based on the sensor readings and the injection timings the ECU has been generating. In fact there should be a short term and a long term trim. But the trim isn't really something from a single sensor, if you see what I mean.

I wouldn't be astonished if whoever it was collected the reverse-engineered Prius PIDs and populated that app with them maybe got one or two things mixed up.

Have you plugged the same app in to somebody else's Prius and seen a different result?