rough mid-range and carb questions

We've probably covered most of this already. 

Compression is good 78/80. 

I have 4 EGTs and they don't show any meaningful difference.  Plug color has also not been useful.  Reading plug colors usually requires running at the RPM you want to test, then shutting off the engine, but the shaking it too harsh to even consider that. 

Coils, plug wires, and plugs have all been replaced multiple times with no change.  Ignition modules have been swapped A to B with no change.  Trigger coils have not been replaced of course, but the air gap was corrected with no change.  Yesterday I swapped the red stator power inputs to the ignition modules, and that didn't have any noticeable effect either.  When my 8ch USB scope arrives, I should be able to prove whether this has anything to do with ignition.    

Mag checks are generally even, but higher than I'd like, around 250 RPM.  Switching ignitions in flight drops a couple hundred RPM, but the engine runs smoothly at cruise power on either or both ignition.  Idle is also fine on either or both ignition. 

The engine mount is a custom design that I made for the Onex since Sonex doesn't make a Onex mount for the 912.  They do make 912 mounts for some of their other planes, but they're bed mounts, which I don't really care for.  I made mine using the ring mount, and Lycoming sized conical dampers.  I do wish they were softer, but they're still essentially new, so not deteriorated in any way. 

The prop is dynamically balanced at 5000 rpm, and is smooth in flight.  It is certainly possible that there is some resonance going on between the prop, engine mount, and airframe.  It's definitely suspicious that the vibrations are at certain RPM ranges, and that nothing I've tried seems to make a difference.  This is probably my leading suspicion right now.

Rusty

did you try different ignition modules altogether. To eliminate that at least

If it behaves identical with other modules then that is firmly off the list.

You say propeller is smooth at 5000 RPM  in cruise. Meaning no vibration there ?

I have not tried different ignition modules.  These are about 2 years and 60 hrs old, since I replaced them both ($1800).  I have two on my ancient engine, but that one has the old trigger coils, and the modules look somewhat different as well, so I'm not 100% sure they're compatible.  I do have a pair of these https://www.ignitech.cz/en/vyrobky/bikes/rotax/rotax%20912%20and%20914/rotax%20912%20and%20914.htm  on the way as well.  I ordered them a couple weeks ago, mostly to play with on the test stand engine, but they might get pushed into service if needed.  Aside from being a fraction of the price, you can also tune the ignition advance rather than have it be fixed like Rotax does with their analog modules.  They also have the ability to run off 12V or the stator power, so if you had a bad stator coil, you could use one of these modules. 

The engine has no vibration at idle, or 5000+ RPM, which makes it hard to believe there could be anything wrong with the ignition.  The vibration peaks around 3500 RPM.  I captured a bit of vibration data with a phone app, and the only peak you see corresponds to the pistons firing.  There's no sign of prop frequency, so I don't think the prop has anything to do with this.  I could have believed the carbs were responsible, but that's pretty unlikely now that they were  needlessly overhauled. 

I can fairly easily reduce the amount of compression on my conical mounts by adding a small shim to the sleeve tube.  That will likely be a good thing even if it doesn't solve the vibration issue, since the mounts are really firm.  When I made the mounts, I basically guessed at how much compression I should have.  I intentionally went on the short side for the sleeve, knowing it would be easier to shim it later than cut it off.   I also ordered a set of "soft" conical dampers, which may or may not fit.  

We may have some better weather Mon and Tue, and I should have my USB scope, so I'll try to prove once and for all if this is an ignition issue.  The AFR unit is here as well, but I'll have to weld some bungs into short extension pipes to test on the ground. 

It's a good thing I enjoy this sort of troubleshooting and testing, or this thing would have been on Ebay already :-)

Rusty

i got the same Ignitech modules, as part of my own troubleshooting. And of course the original ones were almost 20 years and good to have spare parts.... Did not make any difference in my case on how the engine runs that I can notice but like the appearance of the new modules

I tried to scope the plug wires using inductive probes, but that was more of a mess than anything else,  There's so much ignition noise that you pick up pretty much every spark pulse even when the probe isn't connected.  Best I could tell, there was no change in the behavior in the rough running area.  I also tried to scope the stator power to the ignition modules.  Those are ugly, but that's pretty much to be expected.  They showed no sign of change at the rough RPMs.  

There's still a definite change when switching ignitions, but not something you can apply any logic to.  At some RPMs, A is smooth, and B is rough, but at other RPMs, it's just the opposite.  At 4000 RPM, the RPM drop is under 300, but not by much, maybe 200 and 250.  Oddly, the rough ignition often has the lower drop. 

As a final (I hope) ignition test, I made up some temporary switches to use in place of my Aircraft Spruce rotary key switch.  Made no difference of course, but it was easy to test.  The last thing I did was temporarily install the Ignitech ignition modules using 12V power rather than the stator.  No change in the roughness, but they started right up and ran every bit as well as the stock units, so we'll call that good to know for now. 

A couple people who've had to endure all my noisemaking at the airport commented that the engine looked smooth to them.  There is definitely extra vibration centered around 3500, but I started to wonder if it seemed worse than it really was due to all the noise you get inside an aluminum plane, with the canopy shaking around on top (literally <g>) of that.  I had someone else run the engine while I walked around and watched.  I also felt the airframe.  I have to say the problem isn't nearly as obvious from outside. 

100% chance of rain tomorrow, so I'll take the time to weld bungs on the temporary extension pipes, and I'll construct a temporary wiring harness.  Thursday I'll try to measure the AFR to see if anything is grossly out of range.  I'll probably go ahead and install a larger crossover tube in a spare set of intakes, and plan to see if that reduced the problem any.  I suspect it will.

Some day, I might even fly the plane again, if I can remember how :-)

Rusty