Hello, I just recently got my car running with an EMU Pro 8, first post here! It’s a V8 that I converted from Bosch CIS. I am using the stock flywheel with a VR sensor as the primary trigger (multitooth, divider = 3 and tooth count = 48 for a 144 tooth flywheel) and I am using a single tooth cam wheel with a Hall effect sensor. I randomly am getting “Unexpected Missing Teeth” on the primary trigger which cause big, undriveable lean spikes. Eliminating the pulldown and setting the filter to high helped but it is still happening frequently. The VR sensor is a replacement for a stock sensor that was used for the electronic ignition originally on the car. There is no way to adjust the air gap so hopefully this isn’t the root cause!
Here is a scope trace during the issue. It shows an extra tooth on the primary trigger and then all the injections are missing. There are other cases where all the injections end up stacked on top of each other. This makes me think that maybe there is some interaction with the cam sensor that is causing an issue.
From this scope trace it also looks like my cam sensor is inverted, maybe I have it wired wrong… I don’t seem to have a picture of the cam wheel but this is the design, it is just laser cut mild steel.
Here is a screenshot of the issue happening very frequently at idle (Lambda 1 is disconnected). Sometimes it happens much less (even not at all) and other times it is like this making the car undriveable. The missing teeth happen most at idle but can also happen at higher speeds, there doesn’t seem to be a real pattern.
I haven’t played with the “Threshold” value at all yet (currently set to default 0), thought it would be good to get an actual scope trace to understand better what the signal looks like. Before I make some adapter harnesses to do that, is there a way to record the raw analog signal in the ECUMaster client? Or has anyone seen similar issues and have other suggestions to try? I might try to disable sync above a low engine seed like 1,000 rpm, I assume this would eliminate the cam sensor as an issue so if it still occurs then it must be just the primary trigger that has a problem.
need to get this fixed so I can start to actually tune the calibration consistently!
thanks,
Chris
I can only add 1 image at a time, here is the cam sensor:
and the trace showing the lean spikes and missing tooth count
Hi,
The scope shows an additional tooth on the primary trigger that shouldn’t be there.
We have two possible explanations for this:
- Electrical noise - the most common source of noise is the ignition system. Since this is an older engine, make sure you have spark plugs with resistors. They should have an ‘R’ in the name. Supply wires for the coils can also create interference, especially if the wire routing is not optimal. And as usual, a bad grounding strategy can also be the culprit.
- Mechanical play between the camshaft and crankshaft triggers - there will be an issue if the secondary signal moves too much in relation to the primary signal. If the secondary edge moves over a single tooth of the primary trigger, the primary decoding will be incorrect. And with a 144-tooth wheel, you probably don’t need much play for that to happen.
No, there isn’t. The scope tool is more of a digital analyzer. You need an actual oscilloscope to dig deeper. However, with that, you can examine the primary and secondary signals while checking for noise and mechanical play simultaneously.
A solution for the mechanical play may be as simple as changing the edge for one or both signals. There are four combinations of falling and rising edges, all of which are valid in this setup. The only difference between them will be the TDC offset for the first cylinder. A different combination of edges might allow the signals to shift slightly without jumping over a primary tooth.
1 Like
Thanks for the feedback, I’ll get something setup so that I can look at the signals on a scope. I am using VAG active coils but there is a ~1 foot section of the harness that has the coil power and the crank signal in the same bundle. The crank and cam signals are in separate shielded wires though within the bundle.
The mechanical variance is interesting and seems possible since I only see the issue right at the cam falling edge. So root cause would be the edge of the cam and crank signals being too close together? Changing rising and falling edge could move them apart enough to make it more robust against noise? I could make a new cam trigger but might be hard to align perfectly with 144 flywheel teeth! Or is it the divide by 3 = 48 teeth that need to not align?
Chris
It’s not about the noise. If there is enough play between crank and cam, and the edges are very close together, they can swap places. Instead of the cam edge being recorded first and then the crank edge, the cam was moved so that the crank edge is recorded before the cam edge. That will break the synchronization.
Try all the edge combinations and see if it helps.
It’s the easiest thing to do now.
Set primary as rising, set secondary as rising, and test.
Set primary as rising, set secondary as falling, and test.
And so on.
It will change the distance between those edges, and maybe one of the combinations will be good enough. Or try to eliminate unnecessary play if it’s significant.
1 Like
Thanks! I tried disabling sync above 1000rpm and this made the engine run perfect, got the longest drive in yet and got some VE table tuning in 
Need to try the different rising/falling edge tomorrow so I can turn sync back on. The cams are belt driven so there is potential for some play but not much I can do about it so hopefully I can get enough gap by changing the edge type.
1 Like
That’s good.
That mostly confirms it’s not a noise on the crankshaft signal.
Remember that when you find a working combination of edges, you must check the ignition angle with a timing light and adjust the TDC offset, as it will shift after making these changes.
took a while but I finally ran through all the different rising/falling combos and was able to find one that works, thanks! Have some new questions I will start a new thread on.
