"half-rpm" indication from Lightspeed

Hi everyone,
Any of you lightspeed drivers ever seen your *indicated* RPM drop by half? Apparently my lightspeed loses track of one of the two (flywheel) magnet sensor signals, stops firing two of the cylinders, and reports half RPM. When I turn off my mag, sure enough, it's only running on two cylinders.

Oddly enough, it usually happens at startup for the first minute or so then goes away. But it's now happened briefly in flight too, so I really want to get this figured out. Heat expansion is an obvious suspect, but it can happen when the engine is hot too.

I've been working with Klaus from Lightspeed on this, and have changed my flywheel-sensor gap many times with no change. He's running out of ideas and so am I.

I thought I'd check here to see if anyone has seen anything similar. Thanks for any leads!

Polarity

Paul,

I'm not a Lightspeed user, but I understand hall effect sensors, and I'm guessing that's what is used here.

Check to make sure that both magnets are installed with the same polarity. Unidirectional hall sensors really care about this. You could use another small magnet for this test.

Check to make sure that the gap for both magnets is in spec. You may have already done this.

Check that both magnets are roughly equal strength. A small screwdriver and calibrated fingers could perform this test.

Hope this helps.

Ted Johns
RV7 plans preview

Good ol' intermittent problems... How did you measure minimum (crankshaft aft) gap? Did you use clay or something like that? I was initially struggling to get my gap correct, and when I did, it would not operate when the crank was forward all the way. I finally mentioned to Klaus that the clear sealant on the hall effect chip seemed rather thick, and he said "oh, you should take a razor blade and shave it down to the chip" as if it was in the manual or something. In any case, the sealant was probably about .020 to .030" thick, and when removed, I found that I had margin.

I recommend finding the setting for the gap per the book (except do shave off the sealant). BUT, then I believe it is very important to establish that there remains some margin by shimming the sensor board aft by one thin washer from this nominal location. If it works, remove these washers and put it back to nominal. If it doesn't work, one needs to look for something else out of whack.

Thanks everyone for the ideas.

I have played with the magnet gap quite a bit, including taking it all the way down to touching the ring gear (I didn't run it like that), then going up from there. I've put clay in there as well to make sure I'm understanding the size of the gap and that the tracking is correct.

The weird, predictable yet intermittent nature of the problem has got me baffled. Klaus keeps trying to blame it on the fact that I had to cut his sensor connector off to run it through the firewall. Well, I didn't really want to drill a 1.5" hole in the firewall, sorry. I re-wired the connector very carefully using crimped pins. (open barrel pins, unfortunately -- hard to find a DB-15 that takes the good ones)

At the risk of stating the obvious, check the coil wiring on the coil that is not firing. Both connectors at the coil, and the BNC connector at the brain box. The only problem I ever had with my Lightspeed was my own doing...the coil connector wiring, at the coils.

Have you swapped the coils to rule out a coil itself?

dan said:

At the risk of stating the obvious, check the coil wiring on the coil that is not firing. Both connectors at the coil, and the BNC connector at the brain box. The only problem I ever had with my Lightspeed was my own doing...the coil connector wiring, at the coils.

Have you swapped the coils to rule out a coil itself?

Click to expand...

Well, I've looked at the coil wiring and it is in good shape, but have not swapped it. I don't think I would be getting the weird tachometer readings if there was a wiring problem there, but I guess anything is possible. Klaus hasn't gone down this road either, maybe we're both wrong. I'll give it a try...

Radial alignment of the magnets.

Paul, I agree with your assessment that the half rpm indication puts blame up stream of the coils. It sure appears that the magnet is marginally triggering one of the sensors, and with warm-up or whatever, it goes out of range.

This is a pain, but I'd recommend removing the prop, and using clay to determine not just the min/max gap, but also the radial alignment of the magnets. When the magnets are not flying directly over the chip, the margins are rapidly reduced. By rotating the crank until a magnet is over the chip (requires trial and error with clay), you should be able to measure both gap and alignment, as the magnet should leave a small ring impression in the clay. It may be that your pickup board is slightly high or low. Be sure to check all combinations - both magnets on both sensors.

Good luck!

Agreed, but just to make sure I created another BNC-to-coil wire and swapped it in -- but of course the problem wasn't happening this morning. Then I accidentally ran the engine with both coils *disconnected* and still got a normal RPM reading. So it seems like it's a problem at the sensors, in the sensor wiring, or in the brain box.

I also ran a string from the sensor connector at the brain box to see if there was some intermittent connection in there. Tried tugging on it with engine running -> no problem. I'll keep it there and will try again next time it acts up.

I have put clay in between the flywheel and sensors to see that the magnets are passing over the sensors. They look close enough for me, though one is about 1/32" off. I have asked Klaus if that's a problem, haven't heard back as I believe he's at OSH.

I just remembered another theory from a while ago...at least once, the problem occurred when I turned on the alternator, then went away once I turned it off. Electrical noise or coincidence? The sensor wire runs parallel to the alternator lines.

It is possible other instances of the problem have started when I turned on the alternator, as that's one of the things I do shortly after starting the engine.
However, it usually goes away on its own without turning off the alternator.

I need to experiment with turning off the alternator, next time it happens.

Klaus hasn't heard of alternator interference causing problems with his product and was skeptical of the idea -- nonetheless I tried improving the shielding at the brain box, and moving the signal cable a little further from the alternator leads, to no effect.

Paul, good luck finding the problem

Paul Eastham said:

The sensor wire runs parallel to the alternator lines.

Klaus hasn't heard of alternator interference causing problems with his product and was skeptical of the idea -- nonetheless I tried improving the shielding at the brain box, and moving the signal cable a little further from the alternator leads, to no effect.

Click to expand...

I haven't had any problems yet. Mine run next the alt line also.

Kent

Here's something to consider about alternator interference getting into low-level signals. The alternator is a three-phase device which uses diodes to commutate and rectify the outputs of the three windings. The crossover point where one winding's voltage is dropping off while the next one is increasing is at 87% of peak at 60 degrees of the sinusoid. There is a period where the voltage on the windings is below the battery voltage, and during that period of time the battery is supplying current to the load. Let's say that during the peak of conduction the battery is charged to a peak of 14.4V, then when the AC is decreasing the diodes turn off at 13.6V. Now the average voltage being seen by a voltmeter on the battery will be 14V; an o'scope will show the variations. If the diodes have a 0.5V drop each, then the peak voltage on the windings will be 14.4V + 2 X 0.5V or 15.4V. Then the diodes conduction will cease at 13.6V + 2 X 0.5V or 14.6V. This, of course, assumes that the altrernator windings have zero resistance. Now on a sinusoid the 14.6V level would be at 71.5 degrees, 18.5 deg. after the peak. The next diode will start to conduct 23 deg. later when its sinusoid is 18.5 deg. before the peak. So the diodes will pump current into the load and the battery for 37 deg. then be off for 23 deg. during each 60 deg. conduction cycle. Those are 62% duty cycle on and 38% duty cycle off. If the load is 20 A, the alternator will put out a little over 32 A avg. to supply the load and recharge the battery. That current flows from the alternator TO the battery. During the time the battery supplies the load current, the current flows FROM the battery. So what we have is a very high back-and-forth current in the conductors from the alternator to the buss and to the battery; it's not a nice, smooth DC flowing in one direction. Those high, alternating currents can electro-magnetically induce voltages in nearby conductors carrying low-level signals. That's where you get the "alternator whine" in your radios! So my advice is to segregate low-level signals from high current signals as much as possible. Don't run them parallel for any distance. Strobes, too, have high transient currents. And please, please, attach a heavy ground conductor from the alternator ground-lug to the engine; do not rely on the mounting bolts!

Update: it is not the alternator, nor seemingly any interference. The problem still happened today with alternator off, even with everything electrical off.

I also tried another sensor spacing (this time increasing more toward .06"), no change.

Also tried power cycling the lightspeed, no luck.

Twice today I started the engine ("cold" at 70 degrees) and it happened for about 1.5 minutes. I then shutdown, turned everything off and went through the start checklist again -- didn't recur. The same sequence happened again again a few hours later in the day. Sure is tempting to think it's some sort of thermal issue, although things don't get that hot in 60 seconds, especially with the cowl off. The front of the crankcase is only warm after a 3 minute run like that.

Paul

Swapping the coils is quick & easy. Why not rule that out?

Connectors would be my first guess...

Paul Eastham said:

I re-wired the connector very carefully using crimped pins. (open barrel pins, unfortunately -- hard to find a DB-15 that takes the good ones)

Click to expand...

I would go back and get rid of any cheap DB connector you installed. A soldered butt splice for the signal lines would be the best option. If you use DB connectors in aircraft they should always be the machined pins.

Have you swapped the computer control boxes yet to see if the trouble follows the box? If it does you will know where the problem is.

Paul,

Did you check the magnet polarity?


Ted Johns
RV7 plans preview

Ted Johns said:

Paul,

Did you check the magnet polarity?

Click to expand...

Yes. Also magnet strength seemed similar.
After a two-hour flight without any problems yesterday, running hard and high OAT, I taxied to the fuel island and filled up. Restarted the engine 5 minutes later, and there's the problem again. I'm really starting to doubt the thermal theory.

Hopefully when Klaus is back from Oshkosh I can convince him to send me a loaner box to try swapping out. He has been reluctant in the past but now I have a lot more data that seems to rule out most other parts of the system.

I haven't tried swapping the coils yet Dan, though I will soon. Anything is possible, but the RPM reads correct even with both coils disconnected, so it's hard to imagine how a bad coil would cause this particular symptom.

Thanks for all the ideas, everyone! We'll get it figured out soon...

Paul Eastham said:

Yes. Also magnet strength seemed similar.
After a two-hour flight without any problems yesterday, running hard and high OAT, I taxied to the fuel island and filled up. Restarted the engine 5 minutes later, and there's the problem again. I'm really starting to doubt the thermal theory.

Hopefully when Klaus is back from Oshkosh I can convince him to send me a loaner box to try swapping out. He has been reluctant in the past but now I have a lot more data that seems to rule out most other parts of the system.

I haven't tried swapping the coils yet Dan, though I will soon. Anything is possible, but the RPM reads correct even with both coils disconnected, so it's hard to imagine how a bad coil would cause this particular symptom.

Thanks for all the ideas, everyone! We'll get it figured out soon...

Click to expand...

Wow. I suppose at this point I would wait for the system to go into this half speed mode and then hook up a signal generator to the sensor input. If the ignition still was in half_X mode, you would know the box was at fault. Intermittents like this are really a PITA.

It seems like you have debugged this enough though. Klaus really ought to send you a loaner.

Ted Johns
RV7 plans preview

Paul, just to be clear, what are your minimum and maximum gaps? I.e., when the crankshaft is pushed fore and aft. You mentioned you increased it to .060". Is that with the crank fore or aft?

If the crank is getting a little to far away from the sensors, perhaps a dive with no power to push the crank back might tell you something.

Also, do you have two systems?

Sure seems like a flakey top or bottom hall effect sensor or a marginal magnet proximity problem.

Did you do the clay thing on both the top and bottom sensor? I know that I installed my pickup board using the little metal tabs, but that the top and bottom sensors were at just a little different radius.

For one ignition, each one turn of the crank should cause 4 Hall effect events (two magnets, two sensors). It really appears that one of the sensors is not reliably picking up.

Thanks for everyone who chimed in with ideas. To follow up (and hopefully close this thread forever : ), Klaus sent me a loaner controller box. After about 4 hours, 3 flights, and at least 8 starts, the problem has not returned. Fingers crossed!

to set the gap

Paul Eastham said:

Thanks for everyone who chimed in with ideas. To follow up (and hopefully close this thread forever : ), Klaus sent me a loaner controller box. After about 4 hours, 3 flights, and at least 8 starts, the problem has not returned. Fingers crossed!

Click to expand...

Paul,
I truly hope the box was your problem. I though I might offer a FYI for everyone. Most automotive stores carry or can get brass feeler guages to do magnet to coil distance setting. Perhaps everyone here knows this already, as I have a 20-year-old set in my tool box. I got them from Snap-on but only because they were there that day. I did motorcycle ignitions back then. Many types of hall effect sensors. moving magnet, coil and magnet with a reluctor, they all have to have the gap set. I even got a brass wire guage set by special order. FWIW
Bill Jepson

Bill, there is no access at all to the gap when the ring gear is in place - that is why the clay is the best route. Clay is put on the sensor, the ring gear is very carefully put in place (with the crank forward or aft, depending on which measurement you are after). Then it is removed, and the clay measured with a digital caliper or equivalent.

AlexPeterson said:

Bill, there is no access at all to the gap when the ring gear is in place - that is why the clay is the best route. Clay is put on the sensor, the ring gear is very carefully put in place (with the crank forward or aft, depending on which measurement you are after). Then it is removed, and the clay measured with a digital caliper or equivalent.

Click to expand...

Sorry to say, but YUK, what a deficient design. Needs a "windowed" ring gear.
Perhaps that isn't practical though. Too bad.
Bill

Rotary10-RV said:

Sorry to say, but YUK, what a deficient design. Needs a "windowed" ring gear.
Perhaps that isn't practical though. Too bad.
Bill

Click to expand...

Yes, it is a bit of a hassle. I would like to see more margin on the upper clearance as well. I'm sure that there are classic trade-offs involved.

AlexPeterson said:

Yes, it is a bit of a hassle. I would like to see more margin on the upper clearance as well. I'm sure that there are classic trade-offs involved.

Click to expand...

Alex, This is the reason most of the auto and motorcycle ignitions went to a small reluctor wheel with a radially spaced hall-effect pick up. Most went to a fixed permenent magnet mounted behind the coil and a reluctor. Just a toothed wheel to cut the flux lines. Much easer than mounting the moving magnets. It must cost more or everybody would be using it!
Fly safe,
Bill Jepson

Rotary10-RV said:

Alex, This is the reason most of the auto and motorcycle ignitions went to a small reluctor wheel with a radially spaced hall-effect pick up. Most went to a fixed permenent magnet mounted behind the coil and a reluctor. Just a toothed wheel to cut the flux lines. Much easer than mounting the moving magnets. It must cost more or everybody would be using it!
Fly safe,
Bill Jepson

Click to expand...

Bill, I believe that what you described is exactly what is inside the mag hole mounted version of Lightspeed, called "Hall effect" to erroneously differentiate it from the direct crank pickup (which is also hall effect). As with everything, there is a tradeoff - gears, moving parts, etc.