need help diagnosing engine performance - ML Midge - compression ignition diesel

[jimbo7636]

Today was my second day of break-in on a compression ignition engine I machined and assembled. Its the ML Midge engine. I noticed some exhaust gunge coming out where the compression screw enters the cylinder head (see the two photos). I was not expecting that. It must mean there is some exhaust leaking past the contra-piston. Is this normal, or do I need to go make another contra-piston? I did have one good several-minute long run where the RPM seemed correct, but other times the engine is just chugging along (see video too). Another run and the engine was running slow but as the fuel was fully consumed it seemed to run faster - so that run was several minutes long.

I feel like I have the fuel recipe down.

There was a prior run where I felt like I possibly turned the compression screw too far and the piston was hitting or close to hitting the contra-piston. Is that a big deal?

 

[f2cf1g]

The leakage past the contrapiston suggests it's much too loose, it should be a firm press fit in the cylinder. A loose contrapiston is a no-no. You shouldn't need a locknut on Midge if the fit is OK. Contra hitting piston is a definite big deal.

Before you make a new contra, try expanding what you have. If the inside is flat, give it a good tap with a hammer and a flat ended drift to expand the crown. Alternatively you could try to expand the skirt instead. Hit it with something round and gently tapered. If you have to make a new one the most reliable method of getting a good fit is to machine it with a thin tapered wall which will compress into the cylinder. I think this method is described on the late Ron Chernich's site if you're not familiar with it.

In the run shown in the video the motor is probably oscillating not rotating. This can happen with a combination of over-compression and too much fuel, i.e. too rich a setting or flooded crankcase. Back the comp off and check your needle opening. The amount of needle opening in turns is obviously dependent on the taper on the needle but 1 1/4 to 1 1/2 turns is often a good guess. Hope this helps. I've built a Midge, it runs extremely well.

 

[jimbo7636]

The leakage past the contrapiston suggests it's much too loose, it should be a firm press fit in the cylinder. A loose contrapiston is a no-no. You shouldn't need a locknut on Midge if the fit is OK. Contra hitting piston is a definite big deal.

Before you make a new contra, try expanding what you have. If the inside is flat, give it a good tap with a hammer and a flat ended drift to expand the crown. Alternatively you could try to expand the skirt instead. Hit it with something round and gently tapered. If you have to make a new one the most reliable method of getting a good fit is to machine it with a thin tapered wall which will compress into the cylinder. I think this method is described on the late Ron Chernich's site if you're not familiar with it.

In the run shown in the video the motor is probably oscillating not rotating. This can happen with a combination of over-compression and too much fuel, i.e. too rich a setting or flooded crankcase. Back the comp off and check your needle opening. The amount of needle opening in turns is obviously dependent on the taper on the needle but 1 1/4 to 1 1/2 turns is often a good guess. Hope this helps. I've built a Midge, it runs extremely well.

Thanks for sharing your knowledge of the ML Midge. Im really glad to hear you have one running well. I took the engine apart carefully and was happy to see the tapered contra piston I made per the Chernich site had no blackening on the upper part of its taper. so that makes me feel it was sealing. I re-installed the contra-piston and had to use force to tap it into the cylinder (like I did the first time).

I think I found the issue - the solder joint between the needle and thimble was broken! So Im guessing I was getting way too much fuel into the engine and thats why it was chugging. So Im going to practice soldering, fix that needle, and keep on testing. I might slit the thimble now that its free again to improve how the thread sits tight on the spraybar.