Rpm signal loss

[NewZed]

With a Hall sensor don't you just increase the voltage to get better S/N?  Maybe your input voltage is low.  Seems like you're missing the benefit of going to Hall sensor over VR.  I think the main purpose of developing the Hall effect sensors was to get a consistent square wave signal of pre-determined voltage, or strength, much greater than any noise on the line.  Isn't there something in the data logs that shows output voltage from the sensor?  Seems like you should be able to better quantify and optimize  the signal you're using, than just spinning the trigger wheel and hoping the pulse count is right. 

 

The problem seems to be more difficult than it should be.

[seattlejester]

...I appreciate the response, but I think you may be confused with how the hall effect sensor is wired in. It gets a straight 12 volts powered from a constant 12 volt ignition source. Without the car running it's hard to get more than that. It does generate much more of a consistent square wave, but it's still a sensor, and sensors need to be tuned to ignore extraneous noise. I can't remember if I uploaded it, but if you look at the signal the VR was picking up, the signal strength fluctuated quite a bit based off of RPM. With the hall effect sensor, the signal strength is actually fairly flat lined in fact I think it hits the same height each time. I mean I have already reaped the benefits. It works much better then the VR sensor I had, and the car even started for quite a bit. From not running, to running, benefit gained...no?

 

I might be missing some intonation or something not conveyed through typing. I've been taking tooth logs which are pretty representative of what the sensor is actually outputting, there isn't much hope going on in the process.

 

It is not really a difficult problem by any means, I think Akrev hit it on the money and I am in full agreement that it has to do with the reading of noise of some sort which would account for inconsistent tooth count on startup. Which will have to be reduced by moving the sensor closer, shielding the signal wire better, or playing with the potentiometers on the megasquirt board.

[NewZed]

My thought, more fully-formed, is that tuning the MS side for problems picking up the teeth is pointless (pun!) if you're getting a square wave from the sensor and MS uses it correctly.  As you said, you're getting a solid zero or twelve volt square signal at MS, so MS has what it needs.  Noise at the sensor would just add more zeros and twelves, noise on the wire to the sensor would add some intermediate voltages that the VR circuit would use to trigger.  Shielding on the wire from MS to the sensor might help.

 

In other words, if you're getting a clean sensor square wave, just too many of them, then the problem is on the wheel side of the sensor.  If you're getting too many voltage spikes that the VR circuit sees as a trigger, then the problem is noise in between the sensor and the MS board.  I don't know if MS datalogs will pick that up.  An oscilloscope probably would.

 

Just trying to break it in to logical pieces.  Don't try to work on all of it once.  Seems like you're trying changes on both sides of the sensor, and at the sensor.

[seattlejester]

That makes a bit more sense.

 

I do agree that the problem probably lies wheel end of the setup, but I think it is also important to set the noise threshold so that it can ignore any stray noise that tries to work its way into setting off the sensor. I'm not sitting here trying to tune a mechanical problem out, that would indeed be silly.

 

I think the problem is an inconsistent reading of the number of teeth as mentioned. I've given this a fair bit of thought I just don't think people want to be bored by my theories. But here goes.

 

My thought on the matter is the missing tooth makes the ECU reset the count regardless of how many teeth it sees, thus it fluctuates given the catching/stumbling of the motor, until the engine fires and the wheel is spinning at a consistent speed. Example would be a stall. Sometimes the motor is cranking and then it stalls as it tries to catch, to the ECU if the sensor is between teeth it can misread it as a the missing tooth and restart the timing.

 

Physically, the bowing in of the wheel around the pulley may be pushing some teeth out of the plane of the sensor. The sensor is pretty wide so in my mind it shouldn't matter, but it could play a part in not being at the correct angle to generate enough of a field to deflect the electrons which could cause an incorrect signal.

 

Electronically, it could be the fluctuations in cranking voltage as it spikes as the engine starts to catch and powers the alternator briefly. The sudden increase in voltage could possibly cause the stream of electrons to be increased making the sensor skip a tooth or something of that nature.

 

My plan is...

 

1. Take the wheel and the spacers off. Check the height of the spacers and add an extra washer so that it is actually floating off of the pulley so that it can't be bowed in. Flatten the wheel back to straight if it is bent, and then add the correct thickness washer behind the center bolt as well so that it is a flat plane.

 

2. Gap the hall effect sensor so that it is indeed less then 1.5mm from the teeth.

 

3. Play with the potentiometers to see if I can reduce sync loss during cranking (fuel injector fuse removed) and get it to read a constant value without loss.

 

4. Route a temporary exhaust using flex tubing outside so fumes are not a problem.

 

5. Confirm continuity between oil pressure sender and gauge and then confirm continuity between sensor and engine block.

 

6. Go over LC-1 wiring and power up the stand alone gauge so I have two reference points for AFR

 

Pretty sure the solution will be achieved with the next visit. 

 

Do appreciate your input. 

[seattlejester]

So flattened the wheel.

 

Gapped the sensor.

 

Noticed the wiring diagram I had said the oil pressure wire was YW, when I visually looked it is YB, so changed that. Still no oil pressure, so will have to check for continuity at the sensor for ground.

 

Used aluminum flex tube which did not work (back fire ripped the aluminum wall), will have to go purchase some steel flex tube

 

Still have to look at O2 sensor

 

I tried cranking the motor to get it to run for a bit last night, it stumbled a few times but wouldn't catch for long. My thought was that I should be playing with the priming pulse, when I read on the msextra site that the priming pulse is for clearing air out of the injectors and that I should instead be playing with the cranking values.

 

According to the website if the motor tries to catch but does not start I should be playing with the after start enrichment curve as well? Visually the RPM's break out of the 300 cranking start area a few times.

 

I looked over the values I have copied, and it looks like my values are taken exactly from their website for cranking and ASE. My thoughts based on the stumbling is to richen the mixture up a bit. It is listed in % so it's a bit hard to really understand what value it is multiplying exactly.

 

Things that definitely need to be addressed is timing the motor to see if it still exhibits the funny spark.

 

Playing with the pots to lower extraneous signal now that wheel has been looked at.

 

Getting an AFR reading to see if engine is dying from rich or lean condition.

 

Check oil pressure continuity.

 

Find alternative method for venting exhaust.

 

Doesn't really feel like the list has gotten much shorter.

[AkRev]

Took a moment to play around with the ASE and the cranking settings from your msq on the last page. Take a look and see what you think.

To be honest, I would not do much until you get your oil issue sorted out, spark checked, and afr confirmed (I still think your settings are too rich... but better rich than lean.) 

https://drive.google.com/file/d/0ByA2lDE2waj_MTFFb1djdDg4d1U/edit?usp=sharing

[seattlejester]

Thanks a bunch AkRev, I owe you a virtual beer.

 

I agree, I am flying in the blind until I get all those sensors up and running, will work on getting those running and then move around to doing a tooth log to tune the signal and then apply the adjustments (don't have tuner studio on work computer).

 

Next update should be Wednesday evening.

[seattlejester]

It looks like you turned down the pulse width% and lowered the temperature as well? Hard to exactly grasp, but looks to me that you lowered the PW% right about where it would start, so end result would be making the engine run a bit leaner on start up (makes sense to me). Hopefully good news later this evening.

[AkRev]

It looks like you turned down the pulse width% and lowered the temperature as well? Hard to exactly grasp, but looks to me that you lowered the PW% right about where it would start, so end result would be making the engine run a bit leaner on start up (makes sense to me). Hopefully good news later this evening.

Yep, judging by the fact you are getting smoked out of the garage and flame out by the transmission tunnel. My assumptions is that you are running too rich. 

In general (bad timing excused), fire out of the exhaust is a sign of too rich. The opposite of that is the lean "pops" that can be heard through the intake. Due to the fact that I am not there, I just slightly leaned out the cranking pw and priming pw. I would sugest more leaning, but try that first and see if it was an improvement or a step back.

 

After you get your car running and start to adjust the ve table/bins, you will need to revisit those settings (since they are based on % of the VE bins.)

 

Good luck today.

Edited April 2, 2014 by AkRev

[seattlejester]

Note to self, when working on car, good idea to take your keys...

 

Second note to self, when you remove the wire to tighten the nut on the oil pressure sending unit, put it back together...

 

Third note to self, when you remove the power wire for your afr gauge and o2 sensor, make sure to put it back before going about an engine swap...

 

So, supposedly got all my sensors all hooked up, won't know until this saturday. Since I was stuck anyways, I pulled the plugs which were extremely sooty. Wiped them off and burned off excess soot with propane torch. Definitely looks like it was running rich. 

[AkRev]

Note to self, when working on car, good idea to take your keys...

 

Second note to self, when you remove the wire to tighten the nut on the oil pressure sending unit, put it back together...

 

Third note to self, when you remove the power wire for your afr gauge and o2 sensor, make sure to put it back before going about an engine swap...

 

So, supposedly got all my sensors all hooked up, won't know until this saturday. Since I was stuck anyways, I pulled the plugs which were extremely sooty. Wiped them off and burned off excess soot with propane torch. Definitely looks like it was running rich. 

Nice, glad you got those sorted out. Good luck on Saturday. 

[seattlejester]

LC-1 connected, reading within 0.1-0.2 of the gauge, but more then good enough for now. Unfortunately the way I have it wired, during cranking it resets (enters heating element mode), so I don't get a reading until after has caught and died, I think I might need to get an external power source to keep it running so I can find out exactly what is happening. I have seen it show 13-14 while running rough so the fuel mixture seems to be decent.

 

Still no oil pressure, checked that engine has oil, checked continuity to block, just have to check continuity to signal wire.

 

Actually killed my little battery with all the cranking activity. Pretty sure my charger also bit the fan. Not sure if one caused the other.

 

Bought a replacement standard acid-lead battery and surprisingly that cleared up quite a bit of problems. I can get the car to catch and enter into the non crank settings. It doesn't stall as it did before.

 

I think I finally got my sensor in the correct orientation as well. I got my sync loss down to essentially 0. Only get sync loss when I stop cranking.

 

Here's a link to the timing video which I think is the source of all my current problems.

 

I believe this is with the fuel injectors disconnected so I could just look at the spark timing. I'm not sure why it uploaded in vertical instead of landscape, but the timing marks are to the right in that orientation. I added a piece of white tape near the timing mark, the point is pointing towards the TDC mark, ignore the angle of the tape if possible.

 

 

You can see the problems I encounter:

1. The before and after TDC timing firing pulses

2. The empty sparkless cycles

 

Currently the car runs so rough that I had to put it on the ground so that it does not shake itself off of the jack stands. It does not stay running for very long at all, maybe 5-10 seconds at most.

 

What I expect is seeing one mark at around 10BTDC, and then no mark (the mark should be 180 out), and then a mark at 10 BTDC alternating from the wasted spark setup. What I am seeing is a mark at 10btdc, and then another maybe 20atdc. Then long gaps of no spark. Then the correct pattern etc.

 

This is all being read out of the tach input wire for the first coil. I've used this same cheap gun before in the past without problem, so I'm not likely to believe it is the gun especially with how rough the engine is running. If someone has seen cheap timing lights exhibit misfires, I can try a nicer one.

 

I've played around with the center pot to try and tune out some noise, but I've found that the more I turn it the less teeth in general it picks up. According to the tooth logs, if I am reading them correctly, I am missing out on some teeth? (Tooth log will be uploaded tomorrow)

[AkRev]

Sorry to hear. After looking at the tooth logs last time, I noticed that it was still counting the teeth right, just with a lot of sync loss.
So, to go from here I would recommend a few things;

 

First, confirm wiring to coils. (could you post how you wired it up so we could see?)

Second, You know what cylinder one is outputting (10btdc and 20 atdc) what are the other cylinders showing?

 

I am looking through the settings from before and nothing is jumping out to me about the spark settings. Maybe someone with more MS3 experience can chime in and help confirm your settings inside MS.

[madkaw]

Don't have enough experience to help much, but my set up has been rock steady. I will say that with my 3.57 board i had to install a pull up resistor and a jumpers + pot changes for my hall sensor to work properly. It was a bit confusing to understand what MS wanted as far as board requirements for what sensor. I run a 1K resisitor for pull up voltage, but I use the 3X harness which uses a 5v wire. Which sensor exactly are you using? 

[seattlejester]

Sync loss, has been much improved on this last one, I literally can get it down to only reading a sync loss when it actually looses sync, i.e. when I stop cranking. I will upload that during lunch today.

 

This is read off of the tach signal wire for cylinder 1, so I would have to go check the other ones. I think I am getting what you are thinking of? Maybe the inputs have been switched for the other coils so that it is running with essentially what would be the equivalent of messing up your spark plug order? I can go take a look this week. My gun is supposedly a dial back light, so I should theoretically make it able to read what it is firing at.

 

The coil packs are from a 2zzfe engine I believe 200X Toyota Matrix.

http://www.miataturbo.net/useful-saved-posts-8/upgrading-coil-plugs-all-years-cop-writeup-12704/

That is the wiring diagram I followed.

Ground wire is grounding all coils to the same place as the sensor ground for megasquirt (chassis)

Power wire is taken from my makeshift relay box and fed to the power wire of the coil

Spark A is taken from pin 37? Spark B is taken from PAD3 and Spark C is taken from PAD2 if memory serves, I think it was discussed at some point on here, I'll have to go find the thread.

I have run test mode and I can audibly here the coil firing. I haven't tried to confirm exactly where the firing is coming from though, so you may be onto something.

 

Madkaw: any new input is appreciated. Where did you install the jumpers and the resistor if you don't mind me asking? My sensor is the threaded sensor from DIYautotune :

http://www.diyautotune.com/catalog/hall-effect-threaded-body-crankshaft-position-sensor-p-489.html

[seattlejester]

Oh my god, I missed out on the capacitor portion on the parts list for the coils...could that be the problem?

 

"10,000uF capacitor (install on 12v and GND)"

[madkaw]

You might want to read this:

 

http://www.msextra.com/forums/viewtopic.php?f=131&t=53371

 

This line also bother me:

 

Ground wire is grounding all coils to the same place as the sensor ground for megasquirt (chassis)

[seattlejester]

Interesting some new information there, 7 turns instead of 5, and he is adjusting pot R56 instead of R52 as I did. There's some talk about 12v and 5v differences, or connecting it to Vref. Mine is connected to the ignition signal which I believe is 12v.

 

I'll have to go look through the pictures I took of the board to see how the jumpers are set. But he did add a 1000k resistor on the pads marked Pad57.

 

Honestly I had mine put together by diy, so I am unfamiliar with the settings or modifications they did. I did tell them I was planning on using a VR sensors before I switched to this hall effect, so there may be some differences there. I had read that the settings should be the same though.

 

It seems like he was not getting any signal and now he is? I on the other hand have been getting a signal so perhaps these mods are already done.

 

So things to do...

Buy the capacitors and add them to the ground and power wires for the coils per instructions

Check the board to see that jumpers are in the right spot and look for the resistor

If resistor is missing acquire resistor and place jumpers in correct orientation

Turn both pots 7 turns counterclockwise and turn R56 2 turns clockwise

Acquire new battery charger

Test timing output via test mode to see if the correct coil is firing, and during cranking to see if they are firing at the right location/this will also test to see if the capacitors are helping any

Move coil ground?

 

May I ask why the coils grounding to the same place as megasquirt is a worry? I really am curious. Should they be at separate locations? Or grounded to the block or something? None of my other sensors seem to be fluctuation (coolant temp, air intake temp, AFR, TPS). Do the coils just ground a lot of amps and should be located to its own dedicated ground?

 

Tooth log attached. One thing I did notice, was that the time between teeth have gone down with the addition of the new battery by a few ms. I guess my little battery was having trouble.

crank wheel adjustment.csv.zip

Edited April 7, 2014 by seattlejester

[madkaw]

The pull up resistor difference might be because your are using direct ignition voltage. Read those pages carefully-don't want to mislead anyone. At first I was getting rpm for crank and idle and then it went away, so then I started reading.

[seattlejester]

I see, please don't worry, I wouldn't go about adding resistors willy nilly. Thank you for the heads up though. I will definitely take a look.

 

Any thoughts on the reason for the ground being problematic?