5
\$\begingroup\$

I am developing an autonomous heli drone, powered by a petrol engine. Servos and radio receiver use power from a nominal 4.8V NiMH battery pack. Autopilot takes power from a nominal 10.8V NiMH battery pack. (These battery packs with full charge are typically 5.3 -5.5V and 11.7-12.0V)

Petrol engine ignition system is magneto. The low voltage coil produces +/- 150V, this is fed to the high voltage coil connected to the spark plug. When engine is running, electrical noise is present on the airframe, and measurement of the power to the servos and autopilot shows voltage spikes of up to 500mV. This adversely affects the flight electronics.

Ideally, I would like to prevent the ignition system from affecting the other aircraft systems, but cannot find anything on the web regarding suppression of the low voltage side of the magneto ignition. I have tried suppression of the high voltage side using additional resistance in the form of a 5K resistance plug cap - this did not significantly change the noise being induced elsewhere. This makes me suspect the low voltage side.

If I cannot eliminate the source of the noise, then as an alternative I need to clean up both power supplies, using some type of regulation/stabilization immediately upstream of the flight electronics.

Any suggestions?

\$\endgroup\$
7
  • \$\begingroup\$ Please give more details on the magneto system. What I find on google, is the low voltage coil and the high are both wound on the same core. Meanwhile, check ground paths. If the return current has to go through a big loop, there will be a bit rfi. \$\endgroup\$ Feb 24, 2013 at 15:49
  • \$\begingroup\$ Magneto has separate low voltage and high voltage coils. Low voltage coil is positioned adjacent to engine flywheel, approx 150V is generated when magnet passes. This low voltage is routed directl to the high voltage coil, where it is transformd into HT for the spark. Magneto is not the same as that used in piston engine aircraft - it has no breaker points. It is the same as commonly used in single clinder petrol engine garden appliances such as weed wacker and hedge trimmers.... details can be found by googling "Zenoah G26 manual". Ground path is short - straight to engine body. \$\endgroup\$ Feb 24, 2013 at 17:03
  • \$\begingroup\$ Do your low voltage electronics connect electrically to the airframe, or are they isolated? \$\endgroup\$
    – HikeOnPast
    Feb 24, 2013 at 17:59
  • \$\begingroup\$ Low voltage electronics have common ground wire with each other, but are isolated from the airframe metal parts. Airframe is made of G10 material which is excellent dielectric. Metal parts consist of three assemblies - engine, main rotor shaft and head, and tail boom/tail rotor case and mechanics. Each is electrically isolated from the other. When braided ground straps are attached to electrically connect them, the electronics go crazy. \$\endgroup\$ Feb 24, 2013 at 21:43
  • 1
    \$\begingroup\$ Can you deconflict the following statements?: "When engine is running, electrical noise is present on the airframe.."; "Airframe is made of G10 material which is excellent dielectric." \$\endgroup\$
    – HikeOnPast
    Feb 25, 2013 at 3:33

3 Answers 3

2
\$\begingroup\$

Basis for my unfounded assumptions: I think the magneto is flooding the engine assembly with noise. It isn't dissipative, and there is no connection to a larger ground. Given the insulating nature of design, not even earthed when landed. It's just one side of a circuit with big spikes on it.

So: Don't earth your battery powered modules to chassis. You know it makes it worse.

Float the modules, make Faraday cages for the electronics modules and use screened cables for your interconnects.
Use the outer screen to connect the Faraday cages.
Connect the screen at both ends. ( contrary to normal practice.)

Float the electronics from the Faraday cages. Currents from the magneto and other RF sources in the environment are not wanted. (you will pick up plenty of RF once you get airborne)

Lightweight Faraday cages can be made from plastic enclosures and much conductive screening spray. ( generally on the inside, where it doesn't get rubbed off)

Where possible use transformer isolated differential signals between modules. Single ended or ground referenced signals will be problematic.

Shared power supply will cause problems, assume the power supply is a noise source. Try at least common mode chokes at the boards. In noisy mobile systems there isn't a convenient chassis or ground to bypass power line noise to.

Antenna feedthroughs are problematic.

Military solutions to this kind of problem involve screening the noise sources as well. ( see the comment about Fitted For Radio (FFR) vehicles).

\$\endgroup\$
0
\$\begingroup\$

If it were a spark ignition system, put a capacitor across the B+ side and ground.

Here is the first hit I get on 'magneto filter capacitor', and here is an article about maintaining them: Mag Check .

Route ignition wires close along their ground return paths, to reduce the size of the loop they make, and be sure the ground return path is good. For instance, is the high voltage coil's ground direct to the engine?

Ground the electronics to the engine block and Heli frame. That might help. The kill switch wire might be radiating, be sure it is shielded. And as a final resort, run the spark wires through coax shield.

The coils, both the hv and the magnet coil probably radiate pretty badly too. A loop of copper tape 'round the outside of the coil, like you see on smps transformers, can suppress that. You can explore the emi with a 'scope probe with it's ground wire clipped to the tip. Wave it around the engine to see where the emi is coming from.

\$\endgroup\$
2
  • \$\begingroup\$ I already reserached magneto capacitors. Unfortunately, these are used on magnetos with contact breaker points. The magneto on my single cylinder petrol engine has no points. Its exactly the same as the magneto on many small petrol engines used in weed wackers, hedge trimmers and so on. there is simply a magnet on the flywheel, this passes the primary coil, creating and collapsing the magnetic field. \$\endgroup\$ Feb 24, 2013 at 16:57
  • \$\begingroup\$ I will try some of your more recent additions. Grounding the electronics to the engine block made things much worse. I made a shield for the HT cable from secondary coil to spark plug - no effect. I'll make two copper loops to go around the LV and HV coils, and use the scope probe as suggested to see if it helps. Neither the kill wire nor the wire joining LV and HV coils are shielded - I'll also try that. \$\endgroup\$ Feb 24, 2013 at 21:47
0
\$\begingroup\$

Isolated DC to DC converters with isloated PWM. Have an analog 5 volt rail from your batteries for servos and other power devices with a simple BEC and put all of your sensitive electronics on an isolated ground and keep it that way with optoisolators, sparkfun.com sells simple PWM isolation boards. I have seen this exact problem before and solved it this exact way.

\$\endgroup\$
1
  • \$\begingroup\$ P.S. I would get rid of the nickle metal hydride batteries and switch a single lithium ion polymer pack 3 cells for 11.1 volts nominally. Use a BEC for the servos and a 9 to 36 volt to 5 volt isolated DC to DC converter. \$\endgroup\$
    – minimer
    Aug 23, 2013 at 8:59

Not the answer you're looking for? Browse other questions tagged or ask your own question.