square wave pattern in datalog

[tbaggins]

Hi,

I was autotuning and datalogging my startup yesterday and noticed this square wave pattern pattern show up in my datalogs. The length is about 300ms and seems pretty regular. It shows up on the sensors that share VREF (MAF, TPS, MAP) . The log file is too large to attach.

Thanks!

[dans740iL]

Hi,

I was autotuning and datalogging my startup yesterday and noticed this square wave pattern pattern show up in my datalogs. The length is about 300ms and seems pretty regular. It shows up on the sensors that share VREF (MAF, TPS, MAP) . The log file is too large to attach.

Thanks!

Its a hall effect sensor. Normally on the cam wheel 30-2 or flywheel 60-2

Sent from my VS995 using Tapatalk

[tbaggins]

Both the CAM and CKPS are VR sensors for my car, I appreciate the response though. It seems like the period between occurring too slow for it to be a noise from an injector or COP (800 rpm = 13 revs/sec, an individual cylinder should fire one time per rev, so 13 times in 1 second).

The first thing I though was that it is either due to a datalog write or an autotune write? The little flashing bar at the top of the screen seems to pulse around 1 pulse every 1-2 seconds? The pulses are happening every 0.87ish seconds, and last 0.3-0.4 seconds.

Battery voltage is constant.

[turbo conversion]

Your TPS signal is noisy and it also needs to be calibrated if the throttle is closed in the log.

David

[tbaggins]

The TPS is calibrated to 0 (if you look at the MIN reading on the TPS in the datalog). I understand there is noise, but what would cause an 800ms noise signal that is a constant waveform? That really is the question I am getting at. If it were spark plugs or injectors, I would assume the noise would have a higher frequency.

[billr]

Is the output from your MAF sensor a frequency? Anywhere near what you are seeing (or a harmonic of it)?

[tbaggins]

It is a resistance based MAF, the tune is currently using speed density only. It is also full sequential, I don't have that in my signature. Also, for Turbo Conversion, I guess it does seem like the values for the TPS need to be calibrated. I did calibrate to 0 before starting the car, and the auto calibrate is on. I messed with the idle screw during this datalog, not sure if it was before or after this section. It is missing the stop nut, so it turns itself slowly over a few minutes.

A few things I have thought of as possibilities:
Something with autotune writes or datalog writes
The idle air control valve control, it was in open loop, but the wires run along side the TPS? (billr you mentioned a harmonic of the frequency, maybe this is a harmonic of the PWM?)
Alternator control, if the voltage output from the alternator is changing but the reference voltage is being displayed as the same
The stock computer is in parallel doing something strange
an actual oscillation of the throttle body blade (unlikely because the MAP would fall as the throttle body opens, but in the datalog it rises with the TPS voltage).

There is a clear ramp up, hold, and ramp down to the noise, it looks like capacitance building up, holding, and then discharging, but that is a shot in the dark. The left shoulder of the wave seems too sharp though.

[tbaggins]

The wiring is the stock wiring with the MS spliced in, so it is difficult to change to a piece of shielded cable to test if it due to noise near the TPS wires without hacking the harness up. As it is, it is pretty minor and doesn't seem to be affecting anything (haven't looked at a datalog while driving, but it drove better than stock when I tested it the other day). Just curious if anyone has seen something like this before and whether it needs to be/can be remedied easily.

I have an oscilloscope if someone has a suggestion on what to hook it up to and test.

[billr]

Try unplugging the IAC valve, that will stop current in those wires and should greatly reduce any noise they are causing, without having to attack the harness.

[tbaggins]

Good call, guess it's easier to check the low hanging fruit first. Thanks for the help.

[kaeman]

you said the system is connected in parallel, what is running off the megasquirt and what is running on the oe computer? Do you think you might be experiencing a ground loop, try running another ground from the engine to the battery, could it be the wide band heater cycling, where is the power for the wideband coming from, or a fuel pump controller?

[tbaggins]

Injectors, spark, dedicated CLT, MAP, AIT sensors. The only things shared by the stock computer and the megasquirt are the TPS signal, the MAF signal, and the Crank & Cam sensors. The signal for the TPS signal and MAF signal are shared, but the 5vref comes from the megasquirt. All of the sensors are on the megasquirt sensor ground.

Power for the wideband is on a separate fuse box that hooks directly to the battery and is switched with a relay. The stock ECU and MS draw 12v from the same source. Ground for the AFR heaters is on the body, sensor ground is at the MS (there are no blips in the two widebands that I can see btw...). Ground for the MS is connected to a star point on the body where the engine and battery also ground. Cabling for that point is 2 - 4 gauge wires that go from battery to grounding point, and a 4 gauge from ground to block. No other grounds to the engine unless it is a mounting point to the frame hooked to something.

I measured the VREF line the other day to sensor ground, and it was jumping around a bit (0.02 volts). That doesn't really help me though, if a current is being produced by a noise source, I would expect to see voltage fluctuation correct?

I unhooked the idle valve, which did nothing, and unhooked the MAF sensor harness, which made it worse. I moved TPS wires around (closer and further from coils and injectors, and idle valve) and no change.

[billr]

Have you tried removing the OEM ECU to take it out of "parallel operation"?

[tbaggins]

Not yet, I have two relays on my fuel pump, one is switched by the factory fuel pump driver module (so I could integrate the inertia cutoff switch), which is switched by the ECM. I can check into it though.

Is it normal for voltage to fluctuate more when one of the sensors in a series of sensors is unplugged? It shouldn't be floating, because two sensors are still attached (the TPS and MAP).

[billr]

Are you talking about the 5V ref? That should be pretty stable regardless of what sensors are connected to it, but will probably change a bit with each that is connected. A guess? Having the 5V ref change 50 mV when an active sensor like MAP or MAF is connected doesn't seem unreasonable to me.

Getting back to your "square wave", let's get a better handle on the magnitude of that issue. Measure AC voltage on that TPS signal, and the 5V ref, too.

[tbaggins]

When I unplugged the MAF sensor, the TPS sensor voltage fluctuation got worse, so it made the amplitude of that square wave pattern larger. I will get a datalog of it tonight. The reason I bring up the 5V VREF, it is because all of the sensors with the fluctuation are connected to the 5V VREF. Another possibility for the issue is the voltage regulator, so that is why I brought it up.

[billr]

The datalog is good, assuming you post the MSL and not just a screen-shot; but reading the ripple as AC voltage may be even easier to quantify.

[tbaggins]

There is an MSL file below the screen shot above, if you wanted to look at it, but point taken, I will post an MSL or both again.

[billr]

Oops, I missed that. Sorry. I'll take a peek at that MSL...

Edit: I'm back, real curious to see if that ripple shows up on the 5V ref

[tbaggins]

I will hook an oscilloscope up tomorrow, and maybe try to unhook the stock computer and see what happens if I can get the fuel pump relay hotwired.

So let's say there is an oscillation I find, what does that lead me toward? A bad regulator? It is a very regular pattern, and it doesn't seem to change with RPM, which is why I was leaning away from noise. And the RC shape of the waveform is curious as well. I gave someone advice about ruling out writes to the computer for logging, I will try and log to an SD card as well to eliminate my computer from the equation.

Thanks for the help working through this.

Here is another piece of a log with me driving it. The TPS gets better as it opens, but the MAP is still fubar.