So, how many teeth do I really need?

[SQLGUY]

This is my current tach pickup setup: two VR sensors triggered by a single tooth on the rotor.



As I've mentioned in another thread, with this setup I see significant timing error at lower (<2000 RPM) engine speeds. I would like to improve this. My question is, how many teeth do I need to sufficiently improve the low-RPM accuracy? Would it be enough to add six teeth to this existing rotor (one tooth every 45 degrees)? This would change my trigger offset to either -45 or -135, depending on whether I skipped the next tooth or the previous one.

My second choice would be to cut disks from sheet steel, cut notches into them, and weld them either to this rotor or the original one. If I were to do that, how many teeth would I need in the new wheel?

These are my two options because I don't want to replace the VR pickups and timing plate, which I like would have to do in order to use an available trigger wheel. The only one I've found that's close is a 3" from triggerwheels.com (UK), but I really need about 80mm (3 1/8"). Vauxhaul used a 3 1/8" wheel on one of their engines, but I have no idea where to get one of those.

Cheers,
Paul

[newtyres1]

Short answer - If you can double the tach inputs per crank revolution, then the timing deviation would be pushed down to <1000 rpm where it probably doesn't matter.

How often do you ride the bike under 2000rpm? Is the timing deviation causing ridability problems? On most bikes that is a transition zone from idle to riding, so very little time spent riding in that zone.

If you can't find a 3 1/8" Vauxhall wheel, and use a 3" wheel, do you have enough metal on the VR mount to slot it 1.6mm (1/16") to enable adjustment of the gap?

Ian.

[SQLGUY]

The timing errors don't seem to cause ridability issues, but they do cause idle and starting issues. Idle is pretty rough and uneven, and when starting, the error sometimes results in kickback.

To double my tach inputs, as I understand it, I'd have to add one tooth 120 degrees from the current one, and then change my trigger offset to either -120 or -60. I don't know whether this will neccessarily be twice as accurate, as the pickup scheme (and corresponding code) will also change from dual tachs to "3 tooth" trigger wheel.

The current setup has the VR pickups held by screws into tapped holes in the timing plate. To make them adjustable I would need to cut slots in place of the threaded holes, and then mount nuts behind the plate. I don't know if there is clearance for that.

I am currently thinking of trying a pair of 20 tooth bicycle sprockets welded together for extra thickness. Grippo, at one point, posted that I wouldn't have this problem if I had "...even 12 teeth on my trigger wheel", this makes me think that 8 (or 3) wouldn't be enough, but 20 should be.

[grippo]

If you are only having the problems you indicated during cranking, then I think 6 teeth will easily get you over that. Just doubling the teeth might not be enough because the tooth-based algorithm switches over to time based at least one tooth away from the next tach tooth, so 2 teeth instead of 1 might not buy you anything. If you were in competition racing I would want 12-36 teeth.

[SQLGUY]

I set the advanced options for 8 teeth, 1 missing tooth, 4 skip teeth, 0 delay teeth, and single crank wheel. I left my trigger offset at -90 and my advanced offset at 0. This picture is with #1 at TDC.

Initially I had some other settings that didn't work. Then I changed to these. When the engine didn't want to start after first setting these options, I figured it was probably flooded, so I pulled the plugs and cleaned them. Then I put them back in, but left one out so I could visually check for spark. Then it occurred to me that it wasn't going to work, because, before, when I was cranking, uS didn't see me as cranking and didn't run the fuel pump. Out of curiosity, I hit the starter button anyway, and, to my surprise I got spark and the engine even ran pretty well on three cylinders.

So, I put the fourth plug back in and tried starting the engine again... nothing - no RPM, no running of the fuel pump, and, when I pulled the plug again, no spark. Now, the only thing I changed since the bike was running this morning was to add those six teeth to the rotor, so I am very confident of my fuel settings, spark table, etc. I am quite stumped as to why this would work fine on one try and not at all on every other try. Initially my basic ignition settings were set for a time mask of 0 and % mask of 50. I since tried the recommended 0.2 msec and 10%, but no difference there either.

What am I missing?

P.S. I tried a datalog while cranking a couple of times. RPM is a consistent 0, but I do see non-zero delta T's. I don't know what that means, but it sounds significant to me. Any thoughts?

A bit more information:

1. My original setup had the upper (cylinders 2/3) pickup wired to VR1, and the lower (cylinders 1/4) pickup wired to VR2. This didn't really matter since the two ignition channels work independently in dual tachs, so the ECU didn't care whether it was firing 1/4 or 2/4. Now, with a single crank wheel, it does make a difference, so I would guess that my delay teeth should be 4 with a -90 offset, or 2 with 0 offset. This makes me that much more confused as to why the engine fired up that one time.

2. While RPM's are showing 0 during cranking, and the ECU is not activating the fuel pump, the ECU does activate the IAC whenever I crank, so it is getting some input from the pickups. It does the same if I tap the upper pickup with a screwdriver.

3. Here's a picture of my scope while cranking. This is on pin 32 relative to pin 33. P-P is in the 15 to 18 volt range. (sorry about the lousy quality of the picture; this is about a half-second exposure with one hand on the camera and the other on the starter button).





Thanks,
Paul

[hwright]

Ok,

So if I get this right you have changed from dual tach inputs to a 8-1 tooth wheel.

So you only now have a SINGLE VR input. Disconnect the 2nd VR sensor.
You need to set to single crank wheel and pick the VR circuit that gives you the smallest -ve trigger offset. Otherwise use delay teeth to move it around.
Also if you have a normal 4 cylinder engine then you don't need any 2nd channel offset. Also what value resistor (if any) do you have on your VR input. You might need one. Try 10K and move up from there, probabily 40K will be a maximum.

Check my latest msq. attached in V-Twin ignition woes. That was also for an 8-1 tooth wheel. The bike idles at 1100-1300 RPM and is a Vtwin so should be rougher than your 4cyl.

[SQLGUY]

Ok,

So if I get this right you have changed from dual tach inputs to a 8-1 tooth wheel.

Yes.

So you only now have a SINGLE VR input. Disconnect the 2nd VR sensor.

Are you saying that having it connected, even if uS isn't using it, is causing problems? I'd rather leave it connected, in case I can't get this setup to work I can always grind off the other six teeth and switch back to dual tachs.... I suppose it's easy enough to pull the pin from the Ampseal for now, though, and leave it hanging for testing purposes.

You need to set to single crank wheel and pick the VR circuit that gives you the smallest -ve trigger offset. Otherwise use delay teeth to move it around.

It is set to single crank wheel, and I have tried with both 2 delay teeth (=0 trigger offset, and 4 delay teeth =-90 trigger offset). The thing is, it's not that I'm getting out of time spark; I'm getting no spark, no crank, no RPM. Other than triggering the IAC, uS doesn't seem to recognize the signal at all. I think there's something about my setup that doesn't make sense to uS as a trigger wheel. The fact that it started that one time makes even less sense.

Also if you have a normal 4 cylinder engine then you don't need any 2nd channel offset.

It is set to zero.

Also what value resistor (if any) do you have on your VR input. You might need one. Try 10K and move up from there, probabily 40K will be a maximum.

I have 22K resistors on my inputs. That worked fine with dual tachs, with exactly the same tooth profile, but with one tooth instead of seven. Would changing to single crank wheel from dual tachs change the VR input gain?

Thanks,
Paul

[grippo]

Is the top sensor or the bottom sensor connected to cyl 1 ?

[SQLGUY]

The bottom sensor is for 1/4 but it is connected to VR2 (pin 4). The top one is for 2/3, and it is connected to VR1 (ping 32).

[grippo]

Then as far as the processor is concerned, cylinder 2 is what it considers cyl no. 1 and it(your cyl 2) will be fired from Ign 1, assuming Ign 1 is connected to cyl 2. The processor considers tooth 1 (T1) to be the first tooth after the missing one, so T1 is the tooth pointing West. When the wheel turns 180 deg CCW, cyl2 (processor cyl 1) will be at TDC and T1 will be pointing East. It will reach the VR1 sensor 90 deg later, so ithe tach signal will arrive 90 deg ATDC of cyl 2. So -90 deg should be the proper trigger offset. Regardless, you should get synch on any reasonable offset and skip teeth=4 is also correct.

The thing to focus on is that you are not getting synch. Another thing to note is your experience with the plug out and it ran. This also happened to Bruce a few months ago trying to get an FSAE engine going. I believe it turned out to be a problem in the coil wiring, but I don't know if we ever figured out the mechanism for the failure. However, what he saw was consistent, it worked when he held the spark plug in the air but of course with the threads grounded, but failed with the plug screwed in.

Another thing - don't waste time on the pmask and tmask. They have no effect with a toothed wheel.

You may be a candidate for a device Bruce and I are working on. Its a program you load on microsquirt in place of the real program and run a graphing program on the pc. You connect with a serial cable and crank for a few secs. The edge timing data are then dumped to the pc program and you can see a square wave exactly as the processor sees it. And you can scroll through it. I have a rough working version. I will send this to you tomorrow if Bruce hasn't finished adding the options he wants to it.