RPM Rollback

[Runtoeat]

Thanks Corey.  Looks like it might require running a new wire to the solenoid.

[Anticept]

It won't be just one wire. The power has to come from somewhere, which means if the wire is the issue, you'll probably have two or more.

Make sure you look carefully for any place that might rub wires together or on a positive bus. Common areas: near ignition switch, near ignition connectors, or through the firewall.

The ignition connectors can be real subtle. Take them apart and check for cleanliness. Check the wires going into the ignition modules. I had a module failure because vibrations loosened a wire right at the entrance into the modules.

The switch can fail internally too.

Finally, the modules themselves could have an internal failure. If one fails and lets power on the wire, it will charge the other module's retarded timing mode.

[ct9000]

Reading this whole story in total just now, I think that you are getting electromagnetic interference. The sense wire (input) to the soft start module is very high impedance. What can happen is that when the solenoid is not engaged the now open end wire can act as an aerial depending on where it is run. to prove this theory you can provide a discharge path for any stray pickup by using a small capacitor to a good ground. Try a poly cap of about 0.47uf. rated at say 350v.  

[Mike Koerner]

Just as a general comment about CT wiring:

I had a short in the landing light circuit where the wires pass through the fire wall. The fire wall is carbon fiber. There is no rub strip on the edge or protective sleeve on the wire bundle. The wires chafed against the sharp edge of the pass-through hole until it worn through the insulation, shorted it and burnt the edge of the hole.

We could smell the burning when we tried to taxi for takeoff one night, but I couldn't find the problem. We postponed the flight until the next morning. The smell was gone because the landing light wasn't on, but I didn't realize this until several weeks later.

There should be a rub strip around the hole where the wire bundle goes through the firewall or the wires should be protected with a sleeve. Maybe other planes have this and it's just mine that didn't.

Mike Koerner

[Roger Lee]

Wires through the fire wall should be protcted. The CT should have the wires in fire sleeve or a flex plastic tube. I haven't seen a CT with bare wires through the fire wall.

[FlyingMonkey]

31 minutes ago, Roger Lee said:

Wires through the fire wall should be protcted. The CT should have the wires in fire sleeve or a flex plastic tube. I haven't seen a CT with bare wires through the fire wall.

Mine was in fire sleeve, but the sleeve ended right at the firewall and the bundle was poorly protected.  I adjusted the fire sleeve to protrude through the FW, then used high-temp RTV to bond it in place inside the FW by the bus block.  I could see daylight through the hole before that, not a very good firewall in that condition...

[Anticept]

17 hours ago, ct9000 said:

Reading this whole story in total just now, I think that you are getting electromagnetic interference. The sense wire (input) to the soft start module is very high impedance. What can happen is that when the solenoid is not engaged the now open end wire can act as an aerial depending on where it is run. to prove this theory you can provide a discharge path for any stray pickup by using a small capacitor to a good ground. Try a poly cap of about 0.47uf. rated at say 350v.  

The soft start system is connected to the solenoid control terminal. I suppose that control coil would be the cause of your high impedence, but I don't think it's true to say it's a completely open wire when the solenoid is not engaged... by nature of solenoid design, the control coil is always connected.

I don't know enough about radio systems to comment as an authority here though (it's my weakest electrical branch), but I do want to make that point and let people better than I determine if that's a factor.

[ct9000]

Although the solenoid is still connected it is a coil which acts like a very high impedance device at high frequencies, electromagnetic interference is normally high frequency or at least sharp rise time. Not unlike the back EMF you can get from any coil, after all that's how ignition coils work. I can't say for sure this is your problem but it worth thinking about. 

[Anticept]

That's what I figured. I saw was "when the solenoid is not engaged the now open end wire", and wanted to point that out without sounding like I disagree with you, since you were talking about impedance (it might as well be an open wire, that much I do know!).

I believe though, the softstart modules use a capacitor, maybe a FET, and a pull down resistor as part of the design. I base this on the fact that it takes 5-8 seconds (quite a large range) before the softstart switches timing. Their modules are very simple in design, so I doubt they incorporated any kind of timing control involving a clock circuit.

As I was typing this, I walked out to an airplane, took the soft start wire off the starter solenoid, hooked up my voltmeter, and hit it with +12v. It drained to 5 almost immediately upon removing +12v, but it took several seconds with a slowing curve to drop below 1 volt. Definitely a capacitor and drain in each unit.

[Runtoeat]

ct9000, thanks for the thought on a current being induced into the soft start feed.  This is a possible source for current we hadn't considered. As you point out, there is the relay winding inside the solenoid which could act as a induction coil and this winding connects directly to the soft start wire by way of it's connection to the blade connector. Corey suggests that the soft start can be activated with a low current.  Addition of a small capacitor may be indicated if this is required (see comments below)

Corey, an internal voltage leak is a possible source I hadn't considered and this would be a hard one to diagnose.  Our rpm drop was intermittent and perhaps affected by temperature - as the engine heated up, this problem seemed to be more prevalent. Internal breakages are sometimes the result of unequal expansions of potting material and the components.  I have a set of new soft start modules that could be swapped to trouble shoot this.  We did take a good look at the module wiring at the modules and also checked the connectors and wires quite well.  Didn't see anything amiss here.  I strongly suspect that installing a new switch and then examining and cleaning up the wiring at the switch eliminated the rpm rollback.  After replacing the switch and insuring the wires were installed correctly here we have not experienced the rpm problem we were having.

The possibility of the firewall abrading the wires passing thru is a real issue on our CT's.  We did check the pass-thru's and they seemed to have the plastic protection Roger mentions and we found no suspect wires but the pass-thru's aren't very well done as Andy indicates.  Except for a internal voltage leak in the modules, the other possible voltage sources brought forth here could be identified by providing a temporary "clean" interruptible 12v. "stand alone" 12v. supply to the soft start circuit. 

We flew the CT today and it performed without incident.  RPM's were normal with no roll backs. Thank you.

UPDATE: Corey,  I was just going to post this but see your additional comment about applying the voltmeter to the module.  Do I read you correctly that you see a voltage decay from 12+ down to 5 volts after a few seconds and that you feel that the module uses a simple capacitor that is wired to drain off the current, which in turn shuts off the soft start cycle?

[Anticept]

I did not mean to convey that they are switched by current. FETs are voltage controlled, not current controlled. So you could have next to no current, and still activate a FET. However, the fact there is a capacitor and drain that I just now tested for, it may still effectively counter minor interference. It's possible to test for the size of the drain and the size of the capacitor, but that's not really needed here.

The mention of impedance is due to to the fact that a coil introduces a back EMF while the magnetic field builds. The more rapidly the current fluctuates, the higher the "apparent resistance" of the coil. Rapid fluctuating interference can actually cause voltage to rise high enough to interfere with circuit logic, even though there is next to no current backing it up... but it doesn't take hardly any current at all in modern electronics anyways.

However, the moment people start talking about antennas and EMI, I get very reserved about commenting. I know how radio circuits work in theory, but radio is freaking voodoo to me still, it's extremely complex and a whole huge branch of electrical theory of its own.

EDIT in response to your edit, @Runtoeat : Yes, highly likely there's capacitors and drains in it. Possibly hooked up to a FET or something to control which parts of the circuit are active.

It would be interesting to see how RPM dictates advanced vs retarded timing. If we knew that, I could take a reasonable guess as to what the capacitor drains into in the soft start component. Keep in mind, RPM speed below 600 will ALWAYS have a retarded timing, soft start activated or not. Soft start just seems to extend the retarded timing mode to give the engine a bit more time to get up to speed and warm up.

[Runtoeat]

Corey, I don't like to hear that there are capacitors controlling this.  Do I recall correctly that it was defective capacitors in the old modules which resulted in the failure of the engine getting spark for "start" cycle?

[Anticept]

The modules are CDI, or Capacitor Discharge Ignition. Soft start or not, it has capacitors regardless.

For an engine to fail, BOTH modules would have to fail. There's enough redundancy not to worry, but just like aircraft magnetos, don't ignore a large mag drop.

Many many years ago, a guy flying an aircraft became one of those famous crashes. He did a run up and determined there was a mag failure. He decided to continue the flight because "he still had one mag"... the other failed shortly after takeoff and crashed. Maybe that's hearsay though, this industry sure does love its stories and I haven't found an article on it (haven't really tried), but the point stands: if you detect a problem, get it fixed!

In fact, CDI failure is often considerably less damaging to an engine than a magneto failure. Strip a few teeth on a magneto, and you can end up with a spark at a really bad time and blow engine components through the cowling, thus only requiring one failure to cause catastrophe. If a CDI timing fails, it means you aren't getting a spark at all, but you have a second module that will still spark.

[Tom Baker]

14 minutes ago, Runtoeat said:

Corey, I don't like to hear that there are capacitors controlling this.  Do I recall correctly that it was defective capacitors in the old modules which resulted in the failure of the engine getting spark for "start" cycle?

I believe it was bad solder joints due to the switch from leaded solder to unleaded solder.

[Runtoeat]

Corey, point well taken.

[Runtoeat]

Tom, thanks for jogging my feeble memory.