Cam signal idea for full sequential
 

I asked this question a month or so back on miata.net forum and although I got some helpful answers I didn't get a direct answer to whether my suggestion would work so I am asking here. Here's the plan:

I'd like to loose my modified Mk1 cam sensor which is supplying the required cam signal for sequential injection, as it takes up room at the back of the engine and looks messy.

My plan is to remove 4 of the 5 stainless steel studs holding one of my Flyin` Miata adjustable cam gears and replace them with titanium or aluminium studs. Setting a mag sensor in front of the cam gear aligned with the studs should then provide me with one cam signal per engine rotation.

So what do you all think? What are others running full sequential doing?

we are running cars with built in CAS :)

you could get one of the future Boundary Engineering adjustable cam gears that has cam sensor tabs on it.

Sigh, walked right into that one.:laugh: Wouldn't you like the extra space and cleaner look though?

Anyhow I checked boundary engineerings site just before posting this and couldn't see any reference to the sensor tabs?

This is interesting...

Truthfully, I've never really thought about the behavior of a hall-effect sensor in the presence of non-ferrous trigger materials. But come to think of it, aren't the stock NB cam gears made of something other than steel? The NA gears certainly seem to be.

Actually, I think that the same principle as an eddy-current brake would apply here, since the hall effect sensors we typically use have an integral permanent magnet. Eddy brakes use non-ferromagnetic rotors or rails, and they work just fine.

So I question whether replacing a steel stud with a titanium or aluminum one would have any effect upon the sensor's ability to detect it.

(Also, I don't see anything about the Boundary cam gear that shows a sensor tab.)

I'm using an Electromotive GM magnetic sensor. Given how the hall effect works I wonder what the difference is between a magnetic sensor and a hall effect sensor (separate options within the tuning software so there must be some difference - perhaps voltage output?).

steve, check the boundary engineering section or search for posts by TravisR about the cam gear. it's too new for the site I guess.

https://www.miataturbo.net/forum/t33797-4/#post416806

Is it a variable reluctance sensor perhaps? Or perhaps something stupidly simple like a reed switch?

How many wires coming out of it, and is there a shield?

Shield plus 2 wires, GND and signal. Here is the EM sensor sensor page. The magnetic sensor I am planning on using is 255-72218 (the same as the sensor used in the 60-2 tooth trigger wheel setup).

electromotive calls it a "magnetic sensor". the trigger wheels are steel. the sensor has two wires and a shield.

looks like they say any bosch motronic sensor will work... and have a list of sensors. seems easy enough to figure out.

That's a VR sensor. Basically a coil of wire behind a small magnet, same as I'm using with my 36-1 wheel (mine's an OEM Ford piece). It outputs an analog waveform.

Yes, puts out something very close to a square wave (hand held oscilloscope). So do we think this would work? I'd just need to mount the sensor just in front of the cam gear.

The original CAS uses a hall effect switch and it is going to give a digital square wave pattern with a 0-5 volt range.

A magnetic sensor similar to ABS and 36-1 tooth crank wheels output an analog ac signal.

So long as the new sensor outputs a pattern that is recognized by the PCM it should be ok.

BTW, I just took my Tekscope out of moth balls, let me know if you need any waveforms.

You can build a digital circuit that takes the miata's cam and crank sensors, and outputs a TEC friendly cam sync signal.

not if you dont have them... i'm guessing he's got an early car without crank+cas and wants to get rid of the dual duty cas

A wacky idea that I came up with a while back while thinking of how to add CMP back to my car w/o a CAS:

What if you mounted a hall-effect sensor on the top of the VC such that it read one of the lobes of a cam?

wouldn't your waveform be a little soggy? you wouldn't have the rapid change in flux like you do with a square-cornered tab I'd think.

Yes, I don't have them. The engine is a bit of a Frankenstien, '96 short block with FM 2L stroker whilst the top end is a fully worked '99 with VICs.

I guess what I need is confirmation that mounting a sensor in front of a cam gear with 4 titanium and 1 stainless steel stud will give me 1 signal per engine cycle. The TECgt only uses this signal to provide a rising voltage between 180 and 6 deg BTDC compression for No. 1 cylinder - it's just a sync pulse, so no real need for accuracy (60-2 tooth crank trigger provides this).

This method would be a little cleaner (no requirement for pull-up resistors, less to go wrong etc) than the OEM CAS mod on there currently and give more space at the rear of the engine.

you have a 99 head... do you no longer have the front of the cam cover?

Could you just use a taller bolt in that one hole? Or put some sort of funky washer?

Also, you could use a run of the mill hall switch and glue a magnet to any number of your existing bolts. I sort of like that. Should be really straight forward.

No, but I am hoping there is enough left to fabricate a stable mount for the sensor. See here and here.

It should be fine, glue a magnet on.

My friend chopped his cover pretty bad, we got the OEM sensor back on... But you want a hall and a magnet I would think.

If I were to use a sensor other than that supplied by EM I would need to check that it would provide the correct signal. Not sure what I need right now, but glueing a magnet on is an attractive idea. How small could it be? IOW I'd want to upset the balance of the cam gear as little as possible.

Thinking along similar lines could I use the standard EM mag sensor and a thin steel steel 'sheet' glued on there? How small and thin could the sheet be? Would the studs passing nearby result in interference? Any electrical engineer types out there or anyone with the right equipment to test this kind of idea?

You might have to do a little signal tweaking, but I don't think it would be a big deal. I'm not sure how hard it would be to fake a VR signal, and I guess you don't want to do soldering in your case?

If you don't mine, then, I'd pull the gear, drill a hole in it, glue the magnet into the hole (should offset the weight a bit, you could even put a bolt on the opposit side for balance but I don't think you'll need to bother. It could be inboard of the adjuster bolts for minimal unbalancing.)


I guess the first thing to figure out is if it's a VR sensor for sure, or a hall. If it's a hall it's putting out a processed signal anyway. My guess would be VR though.

Anyone think we could take a hall signal, put an offset on it, and use the same "zero crossing" input from his VR sensor?

Soldering I don't mind (although for wiring connections I prefer crimp terminals).

I like the magnet in the cam gear idea. I have to take all this apart some time soon as I want to fit an ATi damper. Where to source a suitable small magnet from though? Also I am not sure what parts (sensors) I would need to make this idea work.

I'm pretty sure any old strong magnet would work. Some rare earth. The sensor will likely have a spec. Hmmm. I wonder if SparkFun would have something like this?

Digi-Key - 620-1185-1-ND (Allegro Microsystems Inc - A6850KLTR-T)

SparkFun Electronics - Dual Axis Magnetic Sensor Eval Board - Honeywell HMC1052L

But probably something from McMaster would make more sense. Honestly, I'd get one off the distributor of an 80's dodge. They used Hall sensors for exactly this application, these speeds, these temperatures.

SparkFun Electronics - Photo Interrupter CNZ1120 A tall bolt would work with this well.

Digi-Key - 620-1045-ND (Allegro Microsystems Inc - A3230LUA-T)
This is basically what you want. There's tons of signals out, but you want a power, ground, signal, and just glue/bolt it inside the cover, call it good.

Edit: Interesting, that last one is designed for ring magnets - i.e. you could use a round magnet on the face of the gear, instead of a single embeded spot. Personally, I like the single dot idea, but it's a way to go. You'd get a rising edge on one cycle, and falling on another.

The TEC will accept a 0-5V signal for the cam sync.

FWIW, Toda cam gears have a steel insert on tye backside to trigger the NB sensor. I stumbled across a pair of these one off gears ebay a few years ago. The hubs were on 90-97 gears initialy so I had a friend machine an NB gear for the hub on the intake side.
http://forum.miata.net/vb/attachment...3&d=1187143401

Thanks for the links. Trouble is I'm not familiar with electronics so I'd need an idiots guide to how to set this up to reliably generate the sync signal for the TEC. At the moment I suspect that it is beyond me.

So this is just a stud/steel insert on the cam gear for the NB cam sensor to read? I wonder if that NB sensor generates a TEC friendly signal? I presume Mazda sell them for a price?

That's a stock NB cam gear outer, fit onto an aftermarket inner hub which is where the adjustability is.

You see the two raised bumps at the top of the outer section and the one raised bump at the bottom (by the white dot)? Those are part of the OEM casting, and they're what the cam sensor reads in the '99-'00.

Why not use a factory NB sensor? Just use one magnet in the right position, feed it 12V (or is it 5V?), voila!

TEC I believe feeds +5v to the sensor (easy enough to check) - anyone know what the NB sensor requires?

If it is +5v then I just glue a small magnet to the cam gear, add a factory NB sensor to the recipe and the TEC will see a signal that it likes?

should be 12V as it is fed straight from the main relay terminal D.

I could be mistaken, but I was pretty sure that sensor sees the same +5V the TPS sees. I don't think I'm feeding +12 into my little circuit.


Also, the miata sensor is a VR sensor, not a hall sensor. It "emulates" hall sensor output (i.e. the conversion to digital is done in the sensor, hence the 5V feed), but it's actually a VR.

But hall sensors are not hard to find. I listed one which should work just fine, and number of those will work.
-Abe.

+12, straight from the main relay:

http://img31.picoodle.com/img/img31/...ym_004c788.gif

Output is open-collector.

thanks guys for doubting me.

That's not a Toda in the picture. Its some home brewed one off that I found on ebay. Very well machined, made from good materials and cad plated. Trusted it enough to put on the car. What caught my eye on the original NA pulley was the adjuster disc being inside the circle of the timing nubs so I could copy the modified outer with an NB gear to take advantage of that. Makes cam timing changes on the dyno easier if you're using the OEM hall sensor.

What year CAS are we talking here?

The CAS from 1990-1997 outputs two separate signals to drive the injectors and ignition coils.

One signal is a square wave 0-5v. The other if I remember correctly is an analog ac type signal.

There is no VR sensor in a Miata CAS from 90-97

The VR sensor is on the OBDII cars from 1996 on...but that is at the crankshaft, not the CAS.

I am going to look through my waveforms and see if I have a CAS signal. If I don't I will get some and post them here.

WE're talking '99 cars, no CAS at all. The VR crank sensor (with the open "hall" output), and a similar one on the cam.


Oh, yeah, that about explains in. Funny, I remember something not working right when I have my 5v ref messed up, but maybe it was something else. Obviously it was.

Whether the factory sensor is powered by 12V or 5V, I'm pretty sure the output is 5V, which the TEC will like.

Slowmx5, here are some waveforms from a 1996 CAS vehicle. On earlier vehicles the wire color changes but the signals are the same. Hopefully they help on your quest.

The bottom signal on both screen captures is the Yellow/Blue wire and the top is the white wire from the CAS.

Both captures were at idle. The difference between them is the time scale only.


EDIT: I removed the waveforms.

That looks like a magnetic pickup output, just like the TEC's crank pikcup.

It looks similar but it's not.

A magnetic or VR sensor outputs a pure sine wave ac output. Meaning it oscillates between negative and positive. If you notice it never goes negative.

Look at the wave form, the bottom signal is triggered from battery voltage as a source. A magnetic sensor doesn't need a source voltage, it develops it's own voltage from induction ie. from the passing of a ferrous trigger across the face of the sensor.

mrtonyg, you say those captures are the two outputs of a '96 CAS? Something isn't right with those traces. The lower trace (CMP, which you call SGC) should not be a little spikey thing, but a nice squarewave with unequal duty cycle on alternating events.

Edit: here's what a healthy NA CAS looks like when both outputs have +5 pullups applied. The top trace is CMP, the next one is CKP, and the third & fourth traces are ignition outputs A and B.

(apologies for the hand-written traces, this was done on my four-channel scope when the floppy drive was broken.)

http://img33.picoodle.com/img/img33/...Pm_11fbb80.gif

Not to be argumentative, but that's only because the output is returned to 12V instead of GND. So if you think of 12V as "ground", it *is* outputting AC.

Joe, I agree with you. I was surprised with the output of the bottom trace.

The car this was taken from runs perfect, and I double checked when I saw the output.

I will check on my other Miata.

Alternating current means negative and positive oscillations. Going from 12v to 0v is not ac. Going from +12v to -12v is ac.

You can't think of 12v as ground because it's not...that argument doesn't make sense.

Just checked on my other Miata...I have a faulty CAS sensor which led to the odd looking trace!!! Hard to believe, the car runs perfect!

Below are off my 1994 Miata.

Both captures were at idle. The difference between them is the time scale only.

I believe the top signal with the shorter pulse is the #1cyl TDC. When I get back from a 2 day vacation, I will sync this signal to the injectors and coils to see their relationship.

Yep, that's what they're supposed to look like. CMP always rises before CKP, however on one cycle it falls before CKP, and on the other it falls after. If your ECU is capable of looking at the relationship between two pulses then you can use this to determine #1 TDC.

This is one reason why the MS1 cannot do full sequential, as its trigger inputs are edge-detecting only and require all edges to be evenly spaced. So we ignore the trailing edge, read the leading edge of CMP (which has a fixed relationship to CKP) and use it to reset the cycle counter.

In relation to newer car terminology, essentially the top trace is CAM sensor and the bottom trace serves as the Crank sensor (CKP).

If we want to be really technical here, we need to bear in mind that a VR sensor, by itself, has no reference. It's a balanced device where the signal on one wire is always of equal but inverse magnitude to the signal on the other. You could connect it to the primary side of an isolation transformer and it'd still work just fine.

It's not until you connect the sensor to an input circuit that you really define a reference. Typically, one side of the sensor is connected to ground. In such an arrangement, the voltage on the opposite wire will swing both above and below ground, having a positive potential for half the cycle, and a negative potential for the other. I'm pretty sure that this meets everyone's accepted definition of AC.

As Obi-Wan would say, it's all relative. +12 can most certainly be thought of as ground, if you are using it as the common point in a circuit. Years ago, cars had positive-ground electrical systems, where the battery's positive terminal was connected to chassis and the negative terminal was connected to the fuse block. A particle physicist might argue that this is actually the most correct usage, owing to the fact that in a DC circuit the physical flow of electrons is from negative to positive.

The fact of the matter is that in a car, there's really no such thing as ground, because the vehicle is not electrically connected to terra firma. There is only chassis, which we have all agreed to use as the common reference point for all circuits in the car.

I was LOOKING for that comic but couldn't find it in time DAHHH

AC is not ac unless the voltage goes in reverse polarity. An oscilloscope picks up reverse polarity when the voltage goes negative or below 0.

That is the definition of AC.

That is incorrect, if you connect one side of a VR sensor to ground it will not generate the expected ac sine wave.

The only wiring that goes to ground on a vr sensor is the shield wire, not the signal wires.

Has anyone tried using the disc out of an injected Festiva distributor in a NA6 CAS to get an isolated cam position signal for sequential injection/ignition?

mrtonyg, the sensor coil itself outputs AC, but one side is connected to 12V instead of to the chassis. Mazda may have done this in order to easily trigger on the falling edge, threshold at 5V or whatever. The giveaway is the exponential (decay) return to 12V.

Here you can see the signal coming out of my VR sensor, as well as the signal produced by the decoder.


http://img262.imageshack.us/img262/7...nvertedfh3.gif


One leg of the VR sensor is connected to ground at the same point as the shield. The other leg of the sensor is connected to the input of the zero-crossing detector. In this trace, the tip probe of the scope is connected to the input stage of the detector, and the ground clip of the probe is connected to circuit ground.

Both inputs of the scope are set to DC, so they are showing you the absolute voltage on the line at any given time. At the left, the green 1> icon shows you the 0V point for channel 1. It is quite clear that the signal coming from the VR sensor is positive with regard to ground for half of the cycle, and negative with regard to ground for the other half.


Do you contend that the waveform shown in channel 1 is not an AC sine wave?


I really can't think of any alternate interpretation of this. Again, this is a "pure" VR sensor, rather than the sensor which Jason is describing. However it would not matter whether the first leg of my sensor was connected to ground, to +12, or to the hot leg of one of the phases of AC power coming into my house, so long as whatever circuit was doing the zero-crossing detection was also referenced to that point.

What sensor are we talking about here. The VR sensor doesn't have a need for reference or input voltage.