When driving your car, youve come to expect it:
Turn the key engine starts up, regardless of outside air temperature, humidity and density altitude.
Unfortunately, this cannot be realized quite as easily in our glider engines, even though we did give it a try. Specifically, there have been a few notable incidents of start up troubles leading to a complete failure of the drive belt.
The following paragraphs are mean to capture accurately just what happens during the start up phase and we hope it will enable pilots to adjust their procedures to avoid troubles.
Instead of choke, glider engines use a primer, a valve that, when open, directly injects fuel into the air stream in the intake. The most primitive primers consist of a simple rubber ball. Our primers are electrical and controlled by the DG DEI engine control unit. The DEI uses the temperature of the engine coolant to control the primer. If the primer switch is on Auto, you dont have to do anything as two things happen sequentially:
- For as long the engine starter is engaged the primer injects a defined amount of fuel until the engine start up
- After the engine has been found to be running, the primer continues to inject fuel for a short amount of time to avoid the engines cutting out again. The pilot can extend this period of primer-induced richer fuel-air mixture by pressing the start button again during this phase
Both the primer fuel flow and the time window of post-start up primer operation can be preset by the pilot using the DEI. Of course, we deliver our gliders with an optimized setting for these parameters except
….and this gets us back to the drive belt failures which are linked to the primer function. Virtually all of these incidents occurred in the US. Why?
We were initially convinced that our drive belts were absolutely fail safe as they are designed for a failure load of 4 metric tons! Who or what might tear at the drive belt with more than 9,000 pounds? Our initial hypothesis was a materials defect. Further research ruled out this hypothesis and we went through extensive brainstorming sessions and pursued a few other theories that equally turned out to be dead-end streets.
Our current thinking is as follows:
- Failure invariably only occurs before the engine fires up and on the ground.
- Failure never occurs during an inflight start-up or while the engine is running
- Failure only occurs after the engine has already coughed a few times on the ground but ultimately didnt start up.
- We suspect failure occurs after the engine during start up experiences a premature ignition in the combustion chamber before the piston dead center is reached. The resulting rapid decelerations combined with the inertia of the propeller and other moving parts can lead to very high peak loads on the drive belt leading to failure.
Prior research during the design phase of the engine had actually identified this problem and it seemed that a torque-limiting clutch would be the way to go. We designed it, built it, and tested it and in fact, we still offer it as an option. The testing however revealed that the torque-limiting clutch was showing signs of slippage after every single flight. At first, we thought this wouldnt be a problem:
Since there was no measurable reduction in power, we figured the device was working as intended just reducing or avoiding peak loads on the drive belt. Soon enough, however, the clutches started to come apart. On problem had been replaced by another!
While we were unable to fully explain the failure of the torque-limiting clutch, it is clear that the addition of a heavy component to the drive train alters the resonance frequencies significantly.
One thing, however, did become clear and now it is your turn:
The drive belt only fails after the engine is operating in a seriously over-rich condition and then misfires. Such a combustion can have so much power that if it occurs before piston dead center, the engine is immediately yanked into reverse motion and the formerly slack side of the drive belt is pulled at and fails.
And with this we are back to the start up phase:
You know that you should adjust the carburetor mixture to a leaner setting if you operate in high density altitude conditions. The adjustment of the main circuit is very easy and and can be accomplished by means of the adjustment bolt and this makes sure the engine receives a slightly leaner mixture at density altitudes above 1,700 ft MSL. However, the main circuit is not what is causing our problem here. The same goes for the idle mixture adjusting screw.
The only device that controls the mixture during the start-up phase is the primer and who the heck adjusts the primer???
Well, people should adjust it because it is so easy. The maintenance manual for the DG-800B describes in paragraph 4.32.2 (page 82) how to adjust the primer. The standard factory settings for the primer are 8 for the fuel flow and 3 for the duration of post start-up operation of the primer. Do not adjust the duration of the post start-up operation of the primer, but do reduce the fuel flow to 6 or 7 when starting up the engine in a high density altitude environment.
For the DG-808C such as the one I own it is captured in the flight manual: Paragraph 7.4.3, setup on page 7.13 how to adjust the primer dosage.
The standard setting is 99% – change it to 80% just to try it out!
The result of our analysis is thus as follows:
- In order to avoid the risk of a drive belt failure you should always err on the lean side for the primer mixture
- Try to set it to 6 (or 80%) and determine whether the engine start up is just as quick as before. If the start up quality of the engine is impaired, revert back to the original setting, if not you may be able to lean the mixture even further.
- Be especially sure to adjust your primer mixture when starting up the engine in high density altitude environments
And why did the failures all occur in the US?
Simply because most of our DG powered gliders fly in the western regions of the United States. It is hot in Arizona and you are 5,000 ft MSL in Minden. Both are high density altitude situations requiring a leaner primer mixture. Of course the pilots didnt know better then. We hope that has changed now and encourage you to experiment!
– friedel weber –
Translation: Kai Talarek