Still struggling with ignition issues [mostly solved]

[SQLGUY]

Hi folks,

I am still having problems setting up ignition on my 1982 Yamaha SECA 750. Basically what I am seeing is rather intermittant spark on the 1/4 cylinders, which are the cylinders picked up through the VR2 input with dual tachs selected. (This is a wasted spark setup, so the secondary coil must discharge through both 1 and 4 spark plugs.) I still can't get the bike to start or run for more than a second or two.

I have a timing light on the number one wire; sometimes I can crank and I get no spark/flash at all, sometimes it's intermittant. I hooked up my scope to the driven side of the coil. What I see there is an occasional 20msec or so low-going pulse that corresponds to a spark flash, and more frequent very narrow pulses that don't trigger the timing light. The pulses that flash the timing light also give a good spark, while the narrower, spiky, pulses do not produce a spark. I've tried new plugs and I have also verified the resistance of the wires and secondary coils.

I currently have the dwell set for the default of 3.1msec, with 2.0msec max spark duration. Given that the red line is 9500 RPM, and the coils each fire once per rev, there is only a total of 6.3msec between triggers at red line, so I don't think correct dwell could be much longer, unless it should be higher for starting and get crowded down at high RPM.

What is the code trying to do with dwell during cranking/starting?
Any ideas why I am seeing those occasional long pulses?
Please advise what I could check or try next to make some progress here.

Thanks,
Paul

[grippo]

It would have been better to have kept on your first thread. There you said you were getting spark, but you couldn't control advance. Now you are only getting intermittent spark at best. We need to start from the beginning. Since you have a scope, check that you are getting good solid tach inputs on both channels. This can be determined by a datalog that shows solid synch with no rpms jumping to 0. If this is not the case you need to scope the inputs. Once we know this works, we can move on to your output configuration.

[SQLGUY]

Since my pickups were causing a trigger at about 30BTDC, they weren't suitable for use with the trigger rise option (too much chance of kickback, etc). Somewhere I had read that the trigger offset was taken into account when using trigger rise; you corrected my understanding on that, which helped things fall into place.

So, I ended up going Calculated, with a -330 degrees trigger offset, and 000 in the two squares normally hit during cranking. This gets me a TDC cranking spark which switches over to 7 degrees of advance after starting.

This setup starts and idles relatively well. I am not sure why the trigger rise setup wasn't consistently firing the timing light, but with this setup it fires just fine.

Now on to fuel issue.....

Thanks for the help,
Paul

[SQLGUY]

After fighting with fuel for a while, to narrow down variables, I tacked my old TCI ignition control module back in to the harness. Guess what... the bike revs very well now. With uS controlling ignition, I could eventually get it to start, but had almost no throttle reponse. It could not be revved over 1800 RPM or so.

I've attached my msq. The ignition settings are still where they were before I put the TCI module back in. To re-add the TCI module, I opened the uS case and ran jumper wires from pins 4, 32, and 33 to the corresponding pins on the TCI module; I also removed the IGN out pins from the uS connector and tucked them in to the TCI connector.

My stock coils are 2.5 Ohms primary. From uS I was trying to run them with 3.1msec max dwell and spark table per the attachment. Timing and advance looked OK, though I am going to do more analysis with the timing light to see how advance is acting with TCI. Coil charging was set to standard. As far as I can tell reading only, TCI is simply a solid state contactor system. The coils are still fed 12V at one end, and the TCI box switches ground to them.

I also hooked up my scope to the TCI output. I see two pulses per cycle. One is a very narrow, inverted spike, the other is a cleaner 8mS or so inverted pulse. This gets as short as 5mS at 4000 RPM. I don't know what it gets to at higher RPM. This would seem to indicate that a max dwell of 3.1mS would be much too low. What do you think?

I'd really like to use uS for spark and fuel, not least because this will fit much better in the space avaialble. Where should I go from here?

Thanks,
Paul

[grippo]

Interpiolating between your last 2 posts it appears you got ignition working relatively well at starting and idle. Then you did some fuel tuning and things deteriorated, so you hooked up the TCI and confirmed it wasn't the fuel that was causing the problem. You need to go back to where you were at the first post and check the timing with a light and make sure it matches MT and both light and MT gauge change as expected when you rev the engine. Then you need to get a datalog running from idle to the rpm it starts dying at and check to see if the rpm ever goes momentarily to 0. You can't proceed to do anything until you get to this point. If you have done this for idle, then you need to go from idle to whatever rpm it starts to break up. If the rpm stays non-zero, then we know the tach input is good and we can look at the output side. You can increse dwell by .5 ms at a time and see if it improves things, but I doubt it will.

[SQLGUY]

I've done some fuel tuning with the stock ignition in place and have been making some progress there. Since things were running consistently with that setup, I decided to try again experimenting with using uS to drive the coils.

I started by making a spark map to match the factory advance curve. I also set the max dwell to 8ms and the max spark to 1ms to match what I was seeing with my scope when the stock module is driving the coils.

The difference on swtiching the coil drive back to the uS was obvious. The bike can be started but idles rough and again has no throttle response. Combustion seems to be degraded as well, as idle EGO reads higher. The timing light shows timing to be correct but not always strong enough to fire the light. I have run datalogs with both ignition setups and in neither case did I see RPM reading zero when the engine was running.

So, I would think that my current setup should match the behavior of the TCI module, but obviously it doesn't. What else could account for this difference?

Thanks,
Paul

[SQLGUY]

Everything I'm seeing in the manuals and on the forums deals with missing tooth pickups. Is the VR zero crossing setup in uS going to get confused by such a wide single tooth?

I'm going to play around with rising and falling edge settings today and also measure the inductance of my coils with a resistor and function generator (so I can get a solid value for max dwell), but I am concered about whether this trigger system just doesn't make sense to uS.

Where can I pick up the processed trigger to the CPU for VR1 and VR2? I'd like to check those on the scope versus the VR signal to see what's actually happening during tach acquisition.

Also, I've posted the data capture I did with the uS running ignition (msq above). I restarted the engine a couple of times during this trace. If you look at the time past 17128, you'll see the most stable part of the capture. The thing that concerns me here is that, with max dwell set to 8ms, I am still seeing dwell jumping all over the place and falling as low as 1.6ms, even though engine speed never exceeds 1400 RPM. I would have expected dwell to stay pretty consistent relative to RPM. Any ideas what's going on here?

Thanks,
Paul

[SQLGUY]

I tried switching falling edge, which I was using, to rising edge, and adjusting the trigger offset. I didn't notice any improvement.

This is what my VR input looks like on the scope (at least within the limits of my drawing ability - there is a decent amount of flat space between the two pulses):



I have set the ignition mask time to 5ms and the mask percentage to 50, but that hasn't helped either.

I am going to see if I can find suitable parts at the hardware store to fabricate an alternate rotor with a narrower tooth. I don't want to alter the original one, in case I do end up having to use the original ignition system instead of uS.

[SQLGUY]

Didn't get to measuring the coil inductance yet, but I did fabricate a new VR rotor. Sorry to say that, while it gives a much narrower (and, surprisingly, taller) VR pulse, it didn't seem to help the ignition issue.

I put the tooth of this one so it sits in the middle of the VR pickup at TDC (to help simplify trigger offset and cranking settings).

[grippo]

I ran your msq and I think I know the problem. The tipoff was the small dwell. The only way such a large drop in dwell could happen is if you try to start dwell before the tach pulse comes in. If this happens, in dual spark mode, the dwell will not be started until the tach pulse comes in, thereby reducing dwell to nothing if that is what it takes. In your case you had the following setup:

--------------tach1-----------------x----------------------------------------------------------------------tach1
----------------|<-------30deg-----x--------------------trig offset=-330deg----------------------------->|

where the x is the spark at 0 advance, so it is also TDC. What trig offset = -330 deg means is that the next tach pulse occurs 330 ATDC (we are only considering the first channel here, the other is the same but shifted).

So in your case you only had 30 deg to play with, but your advance was around 10-20 deg and your 8 ms dwell took up a lot of deg also, so you would get a decent dwell only very occasionally when you had a very small advance. I only ran this at 300 rpm because that is what I used the last time, but I was able to reproduce decreasing dwell by changing the advance to > 20 deg.

If you really have to have such a long dwell and be able to rev really high, you need to move the sensor or the trigger tooth so that you can get a trigger offset of about -60 deg, so you will have about 300 - max adv or about 260 deg of dwell to play with. I think you can just grind off the nub you welded on and re-position it and this should work. But maybe first see if you can get it to run with 0-5 deg of advance - I know this is too little, but you should get a better spark even though it is too retarded.