Rough Running Rotax 912/914 Engine: Ignition and Exhaust

We have had seven major causes for rough running of our Rotax 912 engine. 

Four of them are covered in this post. They are: fouled plugs, faulty ignition cable, improper grounding of ignition module, and broken exhaust baffle.

The remaining three causes are discussed in separate posts: carburetors not synchronized, imprecise propeller pitch, and worn Belleville washers in the gearbox.

Fouled Plugs

During most years, we fly almost exclusively using 100 octane low-lead avgas (100LL).

Our use of 100LL avgas is not a choice but forced on us: Almost all airports that used to offer automotive gasoline (autogas) have stopped that service. 

The term low-lead  for 100LL is an industry ploy to make us feel good. One gallon of fuel contains about 2 grams of lead. That is a large amount compared with the quantity in the leaded automotive gasolines of the 1960s. 

The lead problem stems from the fact that in recent decades Continental and Lycoming, the traditional manufacturers of aircraft engines, raised engine compression for higher performance. 

If this had not been done, the addition of lead would have been unnecessary, and everybody would have had a perfect fuel. 

As an aside, even though Rotax engines have higher compression than the traditional Continental and Lycoming engines, they would run well on that unleaded fuel. That's due to the water cooling of the cylinder heads. 

Let's look at the effect of the lead in 100LL. 

Rotax recommends replacement of the spark plugs every 100 hours. This may not suffice under extensive use of 100LL avgas.

The symptom of deteriorating plugs is rough running right after you have started the engine. This gradually goes away after several minutes. 

Over time, the roughness after engine start increases. Eventually starting the engine becomes an unpleasant experience where you fear that the shaking might damage the engine.

Don't allow the situation to deteriorate to that point. Instead, remove the plugs from a cold engine and replace them with new ones. Use anti seize compound on the plug threads.

Never replace the plugs of a hot engine. If you do this anyway, the hot cylinder heads will shrink around the cold new plugs. Removal of the plugs months later will destroy the threads in the cylinder heads. 

We use NGK plug DPR7EA-9, which is a suitable replacement for the NGK plug that used to be recommended by Rotax. A good source is any Honda motorcycle dealer since it is used on their six-cylinder bikes, or even cheaper, Rock Auto.

We say "that used to be recommended" since in the late 2010s, Rotax introduced another spark plug, part number 296 656. It is about ten times more expensive than the original NGK plug. 

We have no experience with the new Rotax plug. Given that the NGK plug is so inexpensive and has proved to be reliable in our engine, we will continue using it.

If NGK ever discontinues production of the plug originally specified by Rotax, you may switch to DPR7EA-9. It will always be available since Honda uses it. We have relied on it for 20 years.

Faulty Ignition Cable

When we installed the new Rotax 912 in 1995, it exhibited poor idle when started, uneven running up to 3,000 rpm, and worse yet, would really shake in flight when the throttle was pulled back to idle. 

The latter problem was really disconcerting when power was reduced for landing.

When flying on 100LL avgas on long-distance trips, the engine would quickly show the effect of lead-fouled spark plugs and run rough during start-up and taxiing.

We tried all the suggested cures, such as balancing carburetors, adjusting mixture, replacing spark plugs, cleaning carburetors. 

These remedies improved the problems somewhat, but performance never was really satisfactory. 

Eventually, we assumed that this was just part of the Rotax performance. That is, until a recent annual inspection turned up the real culprit.

For the discussion, we need to look at the ignition system.

Each ignition cable of the Rotax engine is connected to the spark plug with NGK connector VD05FMH. The connector housing is made of rubber. Inside the connector is a threaded pin. 

The ignition cable is pushed into the connector until it touches the pin, then the connector is twisted so that the connecting pin is threaded into the copper strands of the cable. 

The process stops when the cable runs against the bottom of the connector. 

Once the cable has been installed into the connector, a tie wrap is pulled into the groove  of the neck of the connector and tightened. 

The tie wrap seals the connector around the cable and thus prevents moisture from getting into the connector.

The rubber of a new connector is quite soft and pliable, and the opening where the ignition cable is to be inserted is undersized relative to the diameter of the cable. 

When a new cable is pushed into a new connector and then the connector is twisted to thread the cable onto the pin, the dry surfaces of cable and connector induce quite a lot of friction. 

Indeed, the friction may become so great that it may seem that the cable has bottomed out in the connector when it has barely touched the pin.

When this happens, the cable does not properly connect with the pin and fires with a weakened spark or not at all. 

In the latter case, a significant rpm drop during the ignition check points out the problem. But in the case of a weakened spark, there may be enough spark that the problem is not detected by the ignition check.

Exactly this installation mistake was made at the factory on my engine. The connector of the spark plug for the top #4 cylinder was never threaded onto the pin. 

After 16 years and 1,400+ hours, the cable suddenly pulled out of the connector during the annual inspection when we removed the connector from the spark plug for the compression test. 

That's how we became aware of the faulty installation. Throughout the years, the ignition check never showed a significant rpm drop.

It is easy to avoid this problem. When installing a connector, measure the depth of the hole of the connector into which the cable is installed. 

This is done best with a thin wire, such as safety wire, so that the depth to the bottom of the hole and not just up to the threaded pin is measured. 

Mark that depth on the cable with masking tape. Now we know that the cable must be installed up to the masking-tape mark. 

For easier installation, apply a very, very thin coat of non-conducting silicon grease to the cable, and then push the cable into the connector and thread the pin onto the cable. 

Due to the coat of silicon grease, this can be done with reasonable effort, and twisting of the connector will become impossible exactly when the cable hits the bottom of the hole and the masking-tape mark is reached.

Finally, add the tie wrap to tighten the connector around the cable.

NGK plug connector

If you have a Rotax 912/914 engine where the above symptoms of roughness occur and all other remedies have been tried, you may want to test proper installation of each connector as follows: 

Place masking tape on the cable next to the connector, cut the tie wrap, unscrew the connector from the cable, and compare the masking-tape mark with the depth of the hole.

If the installation had been done correctly, reinstall the connector using the above procedure.

If you find a faulty installation where a connector was not sufficiently twisted onto the cable, repair this as follows: 

Cut off about 1/2'' of the cable so that you see pristine copper strands, and install a new connector. Do NOT try to use the old connector since the pin likely is damaged due to arcing.

When we did this repair for our Rotax 912 engine, the engine had smooth idle immediately after cold start, had more power on takeoff, and ran smoothly when power was reduced to idle in the air. 

It felt like a completely different engine. No wonder; for one of the cylinders, just one spark plug had been firing properly. 

Since that plug was at the bottom, it fouled easily with 100LL avgas use, producing a rough running engine with avgas use.

Of course, we also checked the remaining connectors for proper installation. They all turned out be correctly installed. But for good measure, we replace all of them, too.

The above fix worked well until 1,600 hrs total time, when another problem surfaced.

Improper Grounding of Ignition Module

After several years of flying, the engine sometimes would not shut down properly: When both ignition circuits were turned off, the engine would shake for several seconds as if the ignition was intermittently turned off and on again. 

At first we suspected that lead build-up in the engine caused dieseling. But an inspection of the cylinders showed that there was almost no lead deposit. 

This likely was due to the fact that for local flying we use premium unleaded gas, and on cross-country trips we add Decalin after each refueling stop.

The Zenith factory counseled that we contact Rotech in Canada since the factory had never seen this problem before.

Rotech said that most likely the problem came from faulty grounding of the Capacitor Discharge Ignition (CDI) module. That module sits on top of the engine. 

Ignition Module

Several grounding wires of the module are attached with one small bolt to the intake manifold of the #1 and #3 cylinders.

To establish good ground connection for those wires, Rotech recommended that the brass loops of the wires be cleaned and reattached with dielectric grease.  The latter step means that you first reattach, and when that is done, brush dielectric grease over the connection to prevent moisture from entering. 

We did that and also carried out cleaning and reinstallation with dielectric grease of the grounding cable that connects engine and firewall.
The results:

1. The engine shut off immediately when the ignition switches were turned off.

2. The engine started much better than before. Just a momentary touch of the start button fired up the engine, even in cold weather.

3. When one of the two ignition systems was turned off during runup, the rpm drop was so much reduced that it was almost unnoticeable.

4. Idle was much smoother than before.

5. At any cruise setting, which for us is anywhere between 4,000 and 5,000 rpm, the engine ran much smoother than  before.

After one year, the problem returned, though was not as severe as before. Cleaning of the brass loops of the CDI module and reinstallation with dielectric grease brought some, but not perfect, relief. 

From then on, the problem returned time and again.

Eventually, we decided to ground the CDI module with a separate wire that goes from the module to the point where the battery ground cable is attached at the firewall. 

Bolt with wires of CDI module and additional grounding wire

Engine performance during start and shutdown instantly became perfect again, and has remained that way.

Another Grounding Problem

Above we discussed how improper grounding of the ignition module may prevent proper engine shutdown: The engine may hesitate to stop, in extreme cases may act as if it was dieseling.

But there is a second explanation when the engine does not immediately shut down. 

The ignition is turned off in the cockpit by connecting two wires of the ignition module with ground. These two wires are connected at the ignition module with two bullet connectors. 

If the bullets or their counterparts are corroded, grounding of the two wires in the cockpit will not turn off the ignition. Here is the fix for this problem.

Pull out the bullets. Clean the inside of the connectors with a cloth dipped in lacquer thinner, using a  small Phillips head screwdriver to gently push the cloth into the connector. 

Then rotate the screwdriver to clean the inside surface the connector. Repeat this step until the cloth comes out clean.

Clean the bullets using a cloth dipped in lacquer thinner until you have a shiny surface. If there is some corrosion that you cannot remove that way, use 600 grid sandpaper to get a metallic surface. Do not use any coarser sandpaper; even 400 paper is not fine enough.

Add dielectric grease to the plastic sleeve at the end of each bullet, making sure that none of the grease ends up on the metal bullet. Reinsert the bullets. 

The grease prevents moisture from entering the connection and causing corrosion, so you will never have to clean the bullet connectors again.

Summary of Grounding Problems

Defective grounding of the CDI module may cause inferior engine starting, excessive rpm drop during runup, rough idling, reduced or uneven cruise performance, and continued running after shutdown.

The last of these problems may also be caused by corrosion of the bullet connectors where the two wires used to turn off the ignition attach at the ignition module. 

Installation of an extra grounding wire that goes from the CDI module to the battery ground and cleaning of the two bullet connectors are a reliable fix.

A Related, Very Important Question

Suppose the engine does not shut down when you turn off the ignition. Do you know how to turn off the engine?

Turning off the master switch will not help: Since the engine has a mechanical fuel pump, it will keep on running until the tanks are dry.

There is only one reliable way to do this: Turn off the supply of fuel. 

Is your airplane so configured that you can do this while you are sitting in the cockpit? If not, it is of the highest priority that you find some way to do this. 

In our plane, this is not a problem. Since we have an all-electric fuel pump system, we simply turn of the electric pumps, and the engine stops within one to two minutes. Nevertheless, we also have the option of turning of the center tank, which supplies all fuel for the engine.

Broken Exhaust Baffle

From one day to the next, engine performance of our Rotax 912 engine changed dramatically: 

It ran smoothly only above 4,800 rpm. During takeoff and cruise, this did not matter much, but during landing the shaking of the engine was disconcerting.

During a trip to the Rockies, the weld of one of the two tailpipes cracked badly. We had the pipe rewelded at the Brigham City airport and made it back home. 

Talking with Travis, the production manager of the Zenith factory in Mexico, MO, it became clear that the muffler should be completely overhauled. 

Muffler with tail pipe segment after multiple weld repairs

During that process, both tailpipes and a broad strip of the muffler were removed. 

It turned out that the baffle welded to the two tailpipes inside the muffler had disconnected, and the engine had been running without any baffle. 

That explained why the engine would run rough below 4,800 rpm.

Travis proposed a single tailpipe replacement that goes through the muffler. 

It is welded shut on one side, and has holes of suitable size and quantity that constitute the baffle. 

The work was done at the Zenith factory, which also supplied ball joints, springs, and hooks for the exhausts pipes going from the cylinders to the muffler.

New tailpipe, capped on top

Back at our airport, our friend Chuck welded the ball joints and hooks to the muffler and pipes after careful fitting. 

In the installation, the springs are secured by safety wire so that in case of a breaking spring, the remnants of the spring are retained and not dropped onto a runway or taxiway. 

Safety-wired springs

The springs are coated with a strip of high-temperature RTV to reduce harmonic vibration, which causes springs to break and produces wear of the hooks.

Strip of high-temperature RTV

After extensive on-the-ground testing, including a high speed test on the runway, we took to the air.

The performance was amazing: 

No vibration whatsoever above 3,800 rpm. Below that rpm, there was some vibration that turned out to be lack of carburetor synchronization. The remedy is described in a separate post.

During a recent trip to the East Coast, we discovered that oil temperature, coolant temperature, and fuel consumption are significantly lower: oil temperature by 10 deg F, coolant temperature by 5 deg F, and fuel consumption by 10%. 

The elimination of vibration and decreases in temperatures and fuel consumption evidently are due to the increased back pressure inside the exhaust system. 

That effect is produced by the baffle designed by Travis of the Zenith factory. Excellent work indeed!

Have any questions or feedback about rough running of the Rotax engine? Please share your thoughts in the comments.