13Beast REW

Did some testing on the crank angle sensors last night. For lack of an oscilloscope I wired them up to the microphone jack of my laptop and recorded the signals emitted while cranking. I'm realizing that by default both signals are noisey regardless of shielding and ground separation. Both signals are also very distinct but accurately depict engine RPM. Because of the noisey signals I first opted to use WavePad to record the signal into a wave file which more people would be able to look at or listen to. Tonight I'll be running the test through a more appropriate PC oscilloscope.

Here are the wav files:

NE Signal - 1.4MB
G Signal - 1.4MB

This is a link to download WavePad.

Both samples were run at 96000 samples per second then saved in a 48Khz 16-bit uncompressed sound file. Each is about 15 seconds; 5 seconds before cranking , 5 seconds while cranking, then another 5 seconds after cranking.

Peek voltage on the NE signal was .006 volts and peek voltage on the G signal was .003 volts. Two things... I may have that backwards and I also don't trust that voltage but thats what I got last night.

All of this was informative but as far as I'm concerned its utimatley useless. Once I run the actual oscilloscope program and recheck the voltages I should have more usable information. At the current moment though I'm already thinking that some resistor tweaking will necessary.

muythaibxr

not surprising... also useful would be the resistances of the sensors themselves.

13Beast REW

Its too damn cold! Didn't work on the car last night but I did buy some more parts for my car.

The purchase was an in-dash flipout VGA touch screen pc monitor which will display megatune (and other computer stuff) instead of havin to make space for a big *** laptop screen while i'm working the MS issues out. I guess I'll stick the laptop under the passenger seat until the amplifier case for my mini pc motherboard comes in. The monitor should be here either tonight or tomorrow. I'll post pics of it running megatune after I install it.

muythaibxr

Cool... sounds like it's going to look nice!

13Beast REW

Interesting night. Ran tests on the crank sensors again with much better software and got much more usable information.

First off here are four oscilloscope graphs

Graph 1 - Ne Signal
Graph 2 - G Signal
Graph 3 - Ne Signal (Osc negative probe on G- signal wire)
Graph 4 - G Signal (Osc negative probe on Ne- signal wire)

I don't know how relevant graphs 3 and 4 are but for some reason I felt I should post them. I think what I might have been looking for was evidence of crosstalk between the Ne and G signals.

Anyway, looking at the Ne graph shows 23 similar waves then 1 shorter wave. I've taken that to mean the FD has a 24-1 wheel.

The G signal graph I haven't been able to decipher yet but its obvious there is a pattern there.

The two primary issues I have with the data I got and the graphs that were made from them is one, I was only able to graph one signal at a time instead of being able to read both signals at the same time and plotting them both on the same graph. Second, I think I may have been using too high a sample rate (192KHz) so when trying to load the data into Excel to have more control over the graph there was more data than Excel could handle and the graph turned out to be mostly useless. I might resample tomorrow after getting or making some new cables for the osc.

I still have the raw wav files that the osc saved the tests to if anyone wants them, otherwise I'll just get rid of em if I resample with a lower sample rate (11-44KHz).

EDIT: One more thing... Megatune has always shown 2 RPM when my engine is not even running. I orignally took this to be as a result of the noise generated by the electrical system whenever the key is in the ON position but I took it a step further out of curiosity and disconnected the ignition connector (the DB-9) and notice MS still showed 2 RPM. I reset the system with the DB-9 still disconnected and megatune still showed 2 RPM. That can't be normal...

muythaibxr

I accidentally edited your post, but I saved those links and will repost them when I'm more awake.

I looked at the graphs, and there is definite crosstalk...

The G signal should be 1 pulse for every 12 Ne pulses.... you have a 12+1 wheel....


The 2 RPM problem isn't a problem... nothing more than a minor bug in megatune... I've had the MS on the oscilloscope when megatune is showing 2 rpms, and there is no fuel being injected, nor is the megasquirt outputting and ignition signal.

13Beast REW

Ken... So what are you thinking regarding dealing with the cross talk? Heavier VR2 resistor or do you think its more related to the shielding?

muythaibxr

actually, do your Ne and G sensors still share a common ground?

The resistors might help a bit... but I'd really need to have it sitting in front of me so I could try some things. I've been reading around to see what other people have done when they have sensors that share a ground... I'll post here when I find something interesting.

Ken

13Beast REW

No, when I have the MS connected the grounds are separated. I reconnect them when using the stock computer because it won't run without them together.

muythaibxr

ahh ok... well, whenever you have time we'll experiment with several things.

1) adjusting load resistors and input resistors. (Load resistors are on Ne- and G-, input resistors are on Ne+ and G+).
2) trying both separate ne/g grounds and grounding them together.... (I don't think grounding together will help, but if it's not that hard to try, it'd be worth it).


I think #1 will be the one that gets us what we're looking for. We just have to experiment. Right now on the circuit I built for you, there aren't any load resistors (that works fine with the 2nd Gen CAS).

TwistedRotors

iTrader: (3)

I'd just like to say that I'm watching this thread like a hawk and intend to setup an MS on my FD as soon as I finish my FC. 13Beast or muythaibxr, is anything special done to the MS hardware wise for the FD ignition?
I did have my TII running on a MS back in the day way before the smart guys figured out timing and staged injection (dual table code and stock ecu for me!). Now I've got a Microtech controlling it but after doing some research it looks like the MS is damn near plug and play for the 2nd gen CAS! Very exciting stuff, I just may have to ditch the MT someday.
-John

muythaibxr

It IS plug 'n play for 2nd gens (S4's for now, S5's will come as soon as I find my extra connector) I'll take pics of my plug 'n play setup (that I'll be selling at some point) and post them in another thread...

For the FD, we're still in the process of figuring out what to do to make the dual vr sensor conditioner work right with the FD's VR sensors. Right now there's crosstalk between the VR sensors so the Ne signal is being affected by the G sensor and the G sensor's signal shows input from the Ne wheel.

But yeah, there are several "special" things done for both the FC and FD.

TwistedRotors

iTrader: (3)

Awlright, well I couldn't wait anymore. Ordered my kit and harness from DIYAutoTune. I'm really looking forward to putting this thing together. It's simply amazing what Al, Bruce and the whole MS community have done!
muythaibxr - What kind of IAC does the FD have stock? Will it work with the MS's IAC controller?
-John

muythaibxr

the MS doesn't have an IAC controller stock. You need to get a ztx688 transistor from digikey, and put that in place of the standard MS's transistor.

I believe the IAC is just a standard 2 wire PWM IAC on the FD right?

13Beast REW

I believe so. I'll verify when I get home.

TwistedRotors

iTrader: (3)

Any updates?
-John

13Beast REW

Sorry, time and money has been going to expensive *** lawyers to try and resolve a child custody case out of court. Hoping I can get back to this not this weekend but possibly next.

13Beast REW

Damn it, Ken I want to go ahead and get this thing working right. When will you have some time for me to bring the FD and MS over to work out the issues?

13Beast REW

Ken...

muythaibxr

hrmm, I should have some time next week. Maybe Monday during the day after I fix the grounds on the GXL.

13Beast REW

OK, I'll call to make sure

muythaibxr

I'm going to try to fix my grounds tomorrow so we can have more time monday. Just call before you come over.

13Beast REW

gotcha

13Beast REW

Ok... finally another update. After a rather simple modification to the MS VR circuitry like we were discussing weeks ago, we've gotten a stable tach signal from the stock FD ignition system. I haven't scoped the ignition signals since the change yet but we're confident that we have what we wanted now.

So now we have clean RPM readings in MegaTune.

I have had some other electronic projects going on in between working with the MS which may have been the cause of some of my MS wiring in the car not being consistant but that will be the next thing I fix.

The car has tried to start several times so, wiring issues aside, it seems that its just a matter of getting the right amount of fuel, dwell, and/or spark advance. I'll be working that out throughout the week. Hopefully the FD will be idling with the MS by this weekend.

muythaibxr

For those who want to know, I fixed the tach signal by putting a 0.01 uF capacitor inline on the G+ input into the lm1815 circuit for that signal.

I decided to do that after looking at 13Beast REW's osiclliscope output and finding that the signal looked like random noise followed by a good pulse every time the G tab on the e-shaft wheel passed the sensor. The capacitor essentially filters out enough of the noise that the lm1815's adaptive filtering can clean up the rest of the signal.