What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Tach reads 300-400 RPM high only when Left Lightspeed is on?

Narwhal

Member
Hello,

75 hours on this airplane since new. Everything worked perfectly for 75 hours then all of a sudden this started.

It's a dual lightspeed plasma III on an IO-360. At idle, and up to about 2000 RPM, the tach on the G3X reads accurately, then it starts to read 200-400 RPM too high by full power, as compared to an external tach (tru tach ii on the glareshield). Redline is 2700 and the tach will go to 3000+ if left alone. If I pull the circuit breaker for the left ignition the problem goes away and the tach is accurate, immediate return to normal RPM indications, only to return to abnormal indications if the circuit breaker is closed.

The system uses a magnet ring mounted to the flywheel and two mini sensors mounted behind the magnet ring, one sensor for each ignition. The G3X tach displays the higher of the two RPM indications, thus when the left system tach reading goes bonkers that's what shows up on the GDU-465/GEA-24 tach.

Additionally, I have encountered engine stumbling/roughness at full power during run ups if I allow the engine to run with the Tach showing 2900+ (while verifying it's not actually spinning that fast with an external tach). I assume this is the electronic ignition doing something to the timing/spark when it thinks the engine is overspeeding, even though it's not? The high RPM on the tach sets off a master warning on the G3X of course.

I've attached pictures of the mini sensors looking from above. Right sensor (functioning correctly) in first pic, left sensor (malfunctioning possibly?) in second pic. I know that's not exact but does the left-side sensor in the second picture look askew? I didn't install this, purchased a previously completed airplane. Could it be wiring or soemthing else? Could a misaligned sensor cause these problems? Could the sensor be contacting the flywheel at high RPM due to vibrations? I have a professional avionics shop looking at this, just trying to help get the bird back in the air.

Thanks.
 

Attachments

  • 20220709_200706.jpg
    20220709_200706.jpg
    194.6 KB · Views: 144
  • 20220709_200550.jpg
    20220709_200550.jpg
    200.8 KB · Views: 145
Don't know the light speed, but do know VR and hall effect sensors. All of them have a maximum gap between the magnet and the hall sensor or the steel and the VR sensor. Once this gap gets too big, the processor has a hard time accurately interpreting the pulses and therefore the revolution counts. Generally speaking, the faster it is spinning, the harder it is for processor to read the pulses, when the wave form is flawed.

Your pics show a clear problem. Make the bad one look like the good one and problems will likely disappear. The canted sensor is too far from the magnet. Don't know if that chunk of metal has two sensors or just one in the middle. Notice how far the sensor is at the middle, compared to the good one. This will change the size/shape of the pulse or simply miss pulses and not surprising that the processor is struggling to make it match what it was taught to see as a good pulse.

I would guess the LS has a rev limiter and it is dropping occasional spark events once it passes whatever limit it has programmed and this accounts for the rough running at 2900. S/W rev limiters will not send every 3rd 4th, etc. spark event to force the RPM back under it's limit. Several ways to do it but have know idea how LS does it.

Larry
 
Last edited:
Thanks, I think the chunks of metal have 3 sensors each according to the diagram in the lightspeed manual.

I did notice there is an optional rev limiter on the lightspeed plasma iii units, apparently mine has it. It doesn't seem like it's great for the engine when the rev limiter kicks in, but I guess if it was a real overspeed and not a tachometer error it would be a nice feature to have to prevent engine damage.

We'll focus our attention on rectifying this sensor situation, which sadly doesn't seem easy because of the confined areas in which you must position tools to reach the fasteners. Hopefully it is not a faulty sensor, because that seems like major surgery. I wonder how/why it got canted like that?

lightspeed engineering said:
Light Speed Engineering is now offering a rev limiter for the popular Plasma III Ignition systems. The Rev Limit occurs at 2975 - 3000 RPM and the ignition resumes normal operation when the RPM drops below 2700 - 2750 RPM. The tach signal is continued during the ignition cut-out so that RPM history can be evaluated on data recording systems.
This is a great feature for those using a constant speed propeller that does not have counterweights to keep the engine from over-speeding in case of a momentary loss of oil pressure to the prop or any other governor failure.
The circuitry can be ordered with the Plasma III as an optional feature or it can be retrofitted to existing Plasma III ignition systems at Light Speed Engineering.
 

Attachments

  • lightspeed.jpg
    lightspeed.jpg
    96.3 KB · Views: 142
Last edited:
Hopefully it is not a faulty sensor, because that seems like major surgery. I wonder how/why it got canted like that?

Given the gap shown in your pic, it is absolutely not surprising that the sensor is struggling to get a pulse. Hall Effect sensors are pretty robust and I doubt it has gone bad. Generally speaking, they either work or they don't.

Larry
 
If you call Klaus and provide him the S/N of your box he can advise if yours has the rev limiter feature.

Suggest you first read the online Plasma manual in it's entirety before calling him.
 
I had a 2007 vintage Plasma 3 giving wild rpm readings from time to time over the course of a year. In my case, the rpm would double for a second or two. Lightspeed thought software, Garmin thought ignition... Garmin was right. Ignition totally died when on downwind, I believe, as the rpm went really erratic. I was too dumb to test the ignitions after I landed, but the next time I went to fly, one was totally dead.
 
I think I understand what your'e saying is that the engine is running correctly on
the questionable system, it's just the g3x indicating ~300 rpm high, right?

The pulse of the magnet on the flywheel as it goes past the sensor is what triggers the spark, so if there was something off of faulty with the sensor itself, you would either experience it as the system not firing on that cylinder at all (missing), or firing at a different time than the second system. That doesn't seem to be the case.

The light speed box has a different output signal for the RPM pulses that feeds to the G3X (see manual 2.7 section 1). These signals require a pull down resistor to feed into the G3x (G3x installation manual section 31 note #15). I'd start with looking at the signal on these pins, either with a voltmeter on a frequency setting or better an oscilloscope. My hunch is something is amiss on this side of the circuit.
 
If you call Klaus and provide him the S/N of your box he can advise if yours has the rev limiter feature.

Suggest you first read the online Plasma manual in it's entirety before calling him.

That is the best advice you could give anyone that is contemplating calling Klaus......... he has taught us well.
 
I think I understand what your'e saying is that the engine is running correctly on
the questionable system, it's just the g3x indicating ~300 rpm high, right?

The pulse of the magnet on the flywheel as it goes past the sensor is what triggers the spark, so if there was something off of faulty with the sensor itself, you would either experience it as the system not firing on that cylinder at all (missing), or firing at a different time than the second system. That doesn't seem to be the case.

The light speed box has a different output signal for the RPM pulses that feeds to the G3X (see manual 2.7 section 1). These signals require a pull down resistor to feed into the G3x (G3x installation manual section 31 note #15). I'd start with looking at the signal on these pins, either with a voltmeter on a frequency setting or better an oscilloscope. My hunch is something is amiss on this side of the circuit.

Yeah, the engine runs fine on both ignitions up to the point where the extra pulses start tripping the rev limiter I guess, because if the tachometer reads more than about 2900 the engine gets real rough. This happens on a normal takeoff if I don't pull the prop back early and/or used reduced power. Under 2900 indicated rpm on the GDU 465, the engine operates normally, and actually spins at correct rpm as verified by the external tach, its just that tachometer on the GDU 465 reads 300 rpm or so too high.

We hooked up an oscilloscope to the connectors between the lightspeed ignitions and the GEA-24 today then did engine runs with both ignitions on, just the left, and just the right. With both ignitions on, or just the left on, there are extra pulses being sensed. It looked like a few tenths of a millisecond after the initial pulse but equal in amplitude (not sure about the tenths of a millisecond timing, can't remember the scale on the y axis of the scope). Everything looks normal on the right. Anyway, not something we didn't already know, but the shop is working with Lightspeed to figure out why. The avionics shop repairman doesn't think it's the sensors, I suggested measuring sensor gap but the tech said the torque seal hasn't moved on the sensors' fasteners so it doesn't make sense that it would've run fine for 75 hours then have this problem due to sensor alignment if they haven't moved? I don't know, I'm just the dumb stick puller/pusher and make no claims of being an avionics repair tech, mechanic or even a builder; like I said I bought the airplane already completed. I will post up videos of the oscilloscope if I get them from the shop. The shop workers' off the cuff preliminary theory is that maybe something wrong the ignition timing coming from the lightspeed (it varies timing based on manifold pressure).
 
Last edited:
Y The avionics shop repairman doesn't think it's the sensors, I suggested measuring sensor gap but the tech said the torque seal hasn't moved on the sensors' fasteners so it doesn't make sense that it would've run fine for 75 hours then have this problem due to sensor alignment if they haven't moved? I don't know,

FInd a new mechanic! You posted the critical data a few days ago. "Sensor gap = .030 - .060" No way that sensor in the second pic has a clearance less than .060" I don't care what the torque seal looks like, that sensor is too far away from the magnet to work reliably. Why expect it to work reliably when it is not installed according to the specifications. Ask the avionics tech if he would expect a GPS navigator to work reliably on 9 volts, just because it did so for the last 75 hours. Specifications exist for a reason. No surprise it works most of the time. The hall effect sensors on my custom EI will read pretty well at 1/4". However, the wave form is not the same as when they are closer to .040 - .050" where they belong. Don';t forget that there is a bunch of software that is trying to decide whether any given signal is valid or not and the s/w designer picked .030 - .060 for a reason and likley coded the error handling around the wave form created at this clearance. Magnet strength also comes into play and if one magnet is a bit weak that can decrease the margin for sensor clearance.

Larry
 
Last edited:
Oscilloscope connected between the lightspeeds and GEA 24.

Just the right igntion firing: (the good one).

https://www.youtube.com/watch?v=s5A3tdiCOIc

Just the left ignition firing (bad one)

https://www.youtube.com/watch?v=8aM_TpO7iQ4

Both ignitions firing (bad one, notice the rev limiter from the sound):

https://www.youtube.com/watch?v=W-QKYiL5Rp0

That is just the computed RPM that has already been derived by the software, based upon the interpreted sensor pulses. You already know that this is not accurate at high RPMs. You really want the scope on the hall effect signal to understand why you are having a problem.
 
That is just the computed RPM that has already been derived by the software, based upon the interpreted sensor pulses. You already know that this is not accurate at high RPMs. You really want the scope on the hall effect signal to understand why you are having a problem.

OK, thanks for the help. The Oscilloscope exercise was requested by Lightspeed according to the avionics tech. I'm just trying to walk that fine line between being a helpful customer and obnoxious customer at this point. I'll diplomatically suggest again that we look at the sensor gap.
 
FInd a new mechanic! You posted the critical data a few days ago. "Sensor gap = .030 - .060" No way that sensor in the second pic has a clearance less than .060"

Good catch Larry. On my particular engine, I can reach behind the flywheel and loosen the sensor screws a bit, then insert a feeler gauge between the sensor and the magnet ring. I usually set mine for .045". If you can't do that, below is a jig made after careful measurements of the flywheel magnet ring. The dimensions may not work for you but you get the idea.
 

Attachments

  • IMG_2442.thumb.JPG.4c3cbb05fa3090059c2a5bd8b26948e7.JPG
    IMG_2442.thumb.JPG.4c3cbb05fa3090059c2a5bd8b26948e7.JPG
    67.9 KB · Views: 67
I have ordered some feeler gauges so that I can do this properly. Unfortunately .045" gauges are difficult to find in person so I had to order some, and it takes a while for them to get up here. I also got 0.030" and 0.060" feeler gauges for reference. Auto parts store feelers all seemed to top out around 0.02-0.03" max. I think I will be able to adjust using these shims without removing the propeller. Getting any kind of special tool like this to Alaska just takes some time though.

I did a quick test with a penny which is 0.0598" thick according to my calipers and it fit between the gap on the left side, but just barely; it got stuck without sliding all the way down, at least in the inboard part; it wouldn't fit into the outboard side of the left sensor gap. Obviously I need more precision than that if I'm going to adjust the gap.
 
Last edited:
I have ordered some feeler gauges so that I can do this properly. Unfortunately .045" gauges are difficult to find in person so I had to order some, and it takes a while for them to get up here. I also got 0.030" and 0.060" feeler gauges for reference. Auto parts store feelers all seemed to top out around 0.02-0.03" max. I think I will be able to adjust using these shims without removing the propeller. Getting any kind of special tool like this to Alaska just takes some time though. My professional assitance through the shop really isn't working out too well, I guess I just need to embrace the DIY nature of experimental nature and get used to doing things myself. They've worked on it for maybe 4-5 hours in the past 2 weeks....maybe they're just busy, or Klaus scared them, or I'm an A-hole. I'm not sure.

I did a quick test with a penny which is 0.00598" thick according to my calipers and it fit between the gap on the left side, but just barely; it got stuck without sliding all the way down, at least in the inboard part; it wouldn't fit into the outboard side of the left sensor gap. Obviously I need more precision than that if I'm going to adjust the gap.

Just go to the parts store and get the classic feeler gauge setup. IT goes from .0025 to .025. You then take the little nut / bolt out and pull off the different sheets to get the thickness you want and hold them together in your fingers. If you want .045', just grab the 020 and the 025 and you are good to go. No need for precision here. Just aim for a gap between .040 and .050.

FYI, a penny should be in the neighborhood of .060, NOT .006. Greeting card paper stock is around .006.

Larry
 
Last edited:
Just go to the parts store and get the classic feeler gauge setup. IT goes from .0025 to .025. You then take the little nut / bolt out and pull off the different sheets to get the thickness you want and hold them together in your fingers. If you want .045', just grab the 020 and the 025 and you are good to go. No need for precision here. Just aim for a gap between .040 and .050.

FYI, a penny should be in the neighborhood of .060, NOT .006. Greeting card paper stock is around .006.

Larry

Yep, that was a typo, meant .06 on the penny.
 
I had a similar issue when building my own EI. Had one inductive pickup that was not happy. Was find at low RPM, but up above 2600 it would get jumpy and the tach would start to spike. Could be the gap or the sensor. Your AV tech is obviously not an expert on EI systems, so probably not the best person to ask. As noted before, get the gap right. If the problem persists, you can replace the sensor or check it with an oscilloscope,
 
Welp, I finally got my feeler gauges in. I tweaked the left sensor so that the 0.03" barely fit and the 0.45" wouldn't fit through the gap. This actually made the problem worse. Now the tach redlines and the rev limiter (or engine popping/shaking) kicks in at about 1500 rpm instead of 2300. Still having the issue only with the left ignition on. Right ignition only with the left breaker pulled works perfectly. I'm tempted just to ask for a new sensor but I guess it's best to wait for the shop to put the oscilloscope between the sensor and the ignition box to be sure, due to the potential pain involved in sensor replacement. I've read that the mini sensors come with a sealed and potted wire though, so the entire wire would have to be re-run to the ignition box in the cockpit; painful. I hope I'm wrong on that.
 
Last edited:
Welp, I finally got my feeler gauges in. I tweaked the left sensor so that the 0.03" barely fit and the 0.45" wouldn't fit through the gap. This actually made the problem worse. Now the tach redlines and the rev limiter (or engine popping/shaking) kicks in at about 1500 rpm instead of 2300. Still having the issue only with the left ignition on. Right ignition only with the left breaker pulled works perfectly. I'm tempted just to ask for a new sensor but I guess it's best to wait for the shop to put the oscilloscope between the sensor and the ignition box to be sure, due to the potential pain involved in sensor replacement. I've read that the mini sensors come with a sealed and potted wire though, so the entire wire would have to be re-run to the ignition box in the cockpit; painful. I hope I'm wrong on that. The avionics shop is not responding to my further/calls and in-person please for continued help despite my best manners and even an attempt to bribe with some donuts. They haven't shown any interest in further assistance in 2 weeks.

It seems unlikely that the sensor moved on it's own and that implies something hit it, causing it to move. That means likelihood of damage is great. I recommend replacing the sensor. That said, the sensor produces the pulse and the computer must read and interpret that pulse. Therefore, problem could be in either spot. The fact that moving the sensor changed the behavior implies an issue with the sensor, but that is pure speculation.

One option is to swap the harnesses between the two boxes and see of the problem moves. If it does, the sensor is bad. If it doesn't, the brain box is bad. This will require resetting the timing before starting though. Magnets are fixed, but the two sensors are in different spots.

FYI, shielded cables CAN be spliced. My Hall effects have potted wires, but go to a weather pack connector near by and then on to the box. Hall effects use square wave with good current, therefore they should work well even without shielded wire. This can be confusing for techs, as many sytsems use VR (variable reluctance) sensors and those are highly susceptible to outside noise and shielded wire is critical.

Larry
 
Last edited:
At this point I would try swapping the 2 Plasma III modules (leaving coils/sensors alone).

If the problem follows the Plasma III module, then you could rule out a sensor problem.

Steve
 
At this point I would try swapping the 2 Plasma III modules (leaving coils/sensors alone).

If the problem follows the Plasma III module, then you could rule out a sensor problem.

Steve

OP, be advised that each box has a programmed position for the sensors and magnets, relative to TDC, based upon the sensors and magnets locations. This is how the box determines TDC and can correctly activate sparks. Swapping a box will bring it's previous setting for sensor location / TDC, which WILL NOT be accurate with the new sensor location. Backfires and other problems could ensue, if it even starts.

Swapping boxes is a good test, BUT new timing adjustment must be performed before engine start.

Full disclosure, I have never seen an LS ignition, but they all pretty much operate the same way - computer MUST establish a fixed offset between magnet pulse and TDC. It cannot self-learn; It MUST be taught.

Larry
 
Last edited:
Full disclosure, I have never seen an LS ignition, but they all pretty much operate the same way - computer MUST establish a fixed offset between magnet pulse and TDC. It cannot self-learn; It MUST be taught.
Larry

From the LSE website:

The Plasma III is a Solid-State, Discrete logic (no microprocessor) for maximum reliability: No Software & No Software Updates​

Not aware of any "teaching" required for an LSE system, but I would follow the pre-operation testing procedure in the manual prior to engine start after swapping the position of the ignition modules.

Something mentioned in the manual that Narwhal might want to check:

Do not route the input wires from the sensors (DC Mini Sensor, Hall Effect Module or Direct Crank Sensor) near the output wires (RG-400 primary ignition wires) or the high-tension leads going to the spark plugs. A ½” or greater separation is recommended to avoid electronic interference.​

Steve
 
From the LSE website:

The Plasma III is a Solid-State, Discrete logic (no microprocessor) for maximum reliability: No Software & No Software Updates​



Not aware of any "teaching" required for an LSE system, but I would follow the pre-operation testing procedure in the manual prior to engine start after swapping the position of the ignition modules.

Something mentioned in the manual that Narwhal might want to check:

Do not route the input wires from the sensors (DC Mini Sensor, Hall Effect Module or Direct Crank Sensor) near the output wires (RG-400 primary ignition wires) or the high-tension leads going to the spark plugs. A ½” or greater separation is recommended to avoid electronic interference.​

Steve

Thanks to all for the continued input, much appreacited.

I checked as well as I could for damage to the sensor, it appeared to be untouched although I didn't remove it, just lossened the mounting bolts for adjustment so it was difficult to see the most important side that faces the magnet ring behind the flywheel.

The wires were all routed pretty well by the builder; there are standoffs for the sensor wires and they are shielded, they definitely do not come within 1/2" of spark plug wires our ignition output wires, they only get bundled into the input connectors for the ignition modules close to the boxes in the cockpit.

The swapping of the boxes has been suggested several times now, I think that is a good plan. Unfortunately the wires are all very tightly bundled in precise lengths with no extra slack so doing this would probably necessitate completely removing the ignition modules from their mounts and relocating them to the cross-side spot, or finding extensions for all the cabling; then there is the concern about making sure the timing is correct which I will admit I don't know how to configure although I'm scouring the manuals for more info about that. Most likely I will just have to be patient with the avionics shop and wait for them to put their oscilloscope between the sensors and the ignition modules to ascertain with certainity where the fault lies. The plane has been grounded for a month with this problem now, what's another week or three.
 

Attachments

  • 20220723_165725.jpg
    20220723_165725.jpg
    287.6 KB · Views: 164
  • 20220723_165745.jpg
    20220723_165745.jpg
    262.8 KB · Views: 80
  • 20220723_165845.jpg
    20220723_165845.jpg
    208.1 KB · Views: 99
  • 20220723_165852.jpg
    20220723_165852.jpg
    244.9 KB · Views: 117
  • 20220723_165916.jpg
    20220723_165916.jpg
    331.3 KB · Views: 134
  • 20220723_171047.jpg
    20220723_171047.jpg
    137.5 KB · Views: 91
Last edited:
The swapping of the boxes has been suggested several times now, I think that is a good plan. Unfortunately the wires are all very tightly bundled in precise lengths with no extra slack so doing this would probably necessitate completely removing the ignition modules from their mounts and relocating them to the cross-side spot, or finding extensions for all the cabling; then there is the concern about making sure the timing is correct which I will admit I don't know how to configure although I'm scouring the manuals for more info about that. Most likely I will just have to be patient with the avionics shop and wait for them to put their oscilloscope between the sensors and the ignition modules to ascertain with certainty where the fault lies. The plane has been grounded for a month with this problem now, what's another week or three.
Yes, I was suggesting you simply unplug the connectors and vacuum line at the Plasma III module, take out a few mounting screws, and swap modules. Nothing to change with the wiring.

The timing is fixed by the placement of the sensors and magnets.

Plasma III.jpg

Steve
 
LSE

OP, be advised that each box has a programmed position for the sensors and magnets, relative to TDC, based upon the sensors and magnets locations. This is how the box determines TDC and can correctly activate sparks. Swapping a box will bring it's previous setting for sensor location / TDC, which WILL NOT be accurate with the new sensor location. Backfires and other problems could ensue, if it even starts.

Swapping boxes is a good test, BUT new timing adjustment must be performed before engine start.

Full disclosure, I have never seen an LS ignition, but they all pretty much operate the same way - computer MUST establish a fixed offset between magnet pulse and TDC. It cannot self-learn; It MUST be taught.

Larry

I don’t fly behind electronic ignition but in reading the LSE documentation I can see that they have the Hall effect sensors mounted 180 degrees apart (for 4 cylinders) and the cylinder phasing is controlled by the spark pug wire routing. Swapping the two boxes shouldn’t make any difference in timing between the boxes. Swapping the leads for the two sensors between the boxes may affect the firing order/phasing thus the spark plug lead routing between the two sets of coils would have to be swapped in this case - i.e. the routing of the wires from the coils for box one need to be changed to match that of the coils for box two, and vice-versa for box two, otherwise the sparks for each cylinder will fire 180 degrees out of phase from the correct TDC timing, if that makes sense.

Skylor
 
EVERY ei system must know the offset between magnet sense and tdc period. If this ei uses a specifically designed piece of hardware with a known offset, great, you can hard code the offset, with no user involvement. Again, know nothing about the lse, but op should take steps to learn before just swapping box s and turning the key. Bad timing can equal bad outcomes. To my eye, the sensors are about 30* apart, not 180.
 
Last edited:
Nothing to hard-code

EVERY ei system must know the offset between magnet sense and tdc period. If this ei uses a specifically designed piece of hardware with a known offset, great, you can hard code the offset, with no user involvement. Again, know nothing about the lse, but op should take steps to learn before just swapping box s and turning the key. Bad timing can equal bad outcomes. To my eye, the sensors are about 30* apart, not 180.

The stock lightspeed Hall effect sensor location is lined up with the case split, one on top and one on the bottom (see attachment). I can’t see anything in the OP’s pictures to indicate his set up is any different.

FYI, lightspeed’s electronics AFAIK are all analog - there’s nothing to hardcode (one of Klaus’s selling points). I suggest reading the Plasma III install manual if this doesn’t make sense to you.

Skylor
 

Attachments

  • CCC6A775-73E0-4AE2-AB63-8F9E070BB32D.jpg
    CCC6A775-73E0-4AE2-AB63-8F9E070BB32D.jpg
    178.2 KB · Views: 121
The stock lightspeed Hall effect sensor location is lined up with the case split, one on top and one on the bottom (see attachment). I can’t see anything in the OP’s pictures to indicate his set up is any different.

FYI, lightspeed’s electronics AFAIK are all analog - there’s nothing to hardcode (one of Klaus’s selling points). I suggest reading the Plasma III install manual if this doesn’t make sense to you.

Skylor

My sensors are 180 degrees apart except the left sensor is at the 9 o'clock position and the right sensor is at the 3 o'clock position behind the flywheel as viewed from the cockpit.
 
My sensors are 180 degrees apart except the left sensor is at the 9 o'clock position and the right sensor is at the 3 o'clock position behind the flywheel as viewed from the cockpit.

Yes, and since the sensors are 180 degrees apart there is no inherent difference in the timing between the two electronics boxes because each ignition system fires every 180 degrees of crankshaft rotation. As I mentioned earlier, if you were to swap the sensor connections between the two electronics boxes without swapping the sensors positions themselves you’ll end up firing the spark plugs 180 degrees from the correct crankshaft position. To do this correctly, you’ll either have to swap the sensor positions or reroute the spark plug wires to different cylinders in order to achieve the correct “phasing” as lightspeed calls it. This part is covered in section 5.1 of the Plasma III manual.

Skylor
 
Last edited:
Thanks for the help everyone!

My avionics tech switched all the connectors from the right ignition box to the left ignition box and vice versa.
We had to make extensions for the BNC connectors going to the ignition coils to make all the cables fit.

The problem still went away when the left igntion Breaker was off/open, the same as before. With both ignitions breakers in or just the left ignition breaker in the problem remained. So the problem was not following the boxes. Interestingly the G3X said "right ignition off" when I had the left breaker pulled, but that shouldn't matter I guess, seeing as how the left ignition module would now be connected to the right inputs on the GEA 24 and vice versa.

Being that we've checked the wiring throughly and extensively, this seems to indicate a faulty DC Mini Sensor. Now the fun of replacing the sensor commences. It is very difficult to access, especially to try to torque it properly. My avionics tech says the prop and flywheel or the baffling might need to come off...I'm contacting the builder to see of they have a recommended procedure. In the mean time at least I have a part on the way.
 
Last edited:
Happy to hear that you confirmed the source of the problem. Good luck with getting the new part in.
 
Welp, turns out the new sensor had no effect. Back to the drawing board....no idea would be at this point. I guess I should try replacing the ignition module even though the problem did not follow the module when we swapped the left ignition for the right....maybe we are misunderstanding the way the boxes interact with the GEA-24 and tachometer....an expensive misunderstanding for sure.
 
Doesn't your data so far indicate a wiring problem between the crank sensor and ignition module? Something is giving the box false signals, in order to create those errant pulses you note on the oscilloscope.
 
Welp, turns out the new sensor had no effect. Back to the drawing board....no idea would be at this point. I guess I should try replacing the ignition module even though the problem did not follow the module when we swapped the left ignition for the right....maybe we are misunderstanding the way the boxes interact with the GEA-24 and tachometer....an expensive misunderstanding for sure.

If the box swap didn't move the problem it is hard to imagine how the box could be faulty. Further, this is NOT a problem with the interface to the garmin box as you are seeing the rev limiter take over and the rev limiter lives in the ignition box, not the garmin box. That proves that the ignition box is seeing the high RPMs and not just a communications issue with the garmin box.

The box just observes and counts electrical pulses from the sensor. That leaves the sensor, the magnet that the sensor observes/counts (This shouldn't be an issue as the other sensor is seeing it) and the wiring/connectors between the sensor and the box that could corrupt or marginalize the signal. That said, if a magnet is weaker than spec, it is possible that one sensors sees it 100% of the time and the other only 98% of the time - tolerance issue. The faster the crank spins, the shorter the pulse and therefore the difficulty in capturing it. If there was impact damage that moved the sensor, it would stand to reason that the magnet could have also been damaged - they are quite fragile and crack easilly.

Is it possible the box swap effort was flawed? If not, I would replace the wire between the sensor and the box. Also. look for high current wires, like plug wires or alternator output, that are in close contact with the sensor wire (parallel contact MUCH worse than perpendicular contact), as that could be introducing electrical pulses on the wire that the box is interpreting as magnet passes. With hall effect sensors, they send a simple 5v pulse as the magnet passes. spark energy is very high and can easily bleed to nearby wires via magnetic field. I would also get a magnetic energy app for your phone and measure the strength of the magnets, looking for a significant difference between them which would be a problem.

While I don't know the s/w the box uses, I speculate that the sensors are seeing extra pules and not missing pulses. Higher than actual RPM in theory should be too many pules and not missing pulses, but this depends upon the software's error handling protocol. This points to electrical interference over wiring or magnet issues. Any re-wiring efforts immediately preceding this problem?

Larry
 
Last edited:
Thanks for the continued assistance.

So there is no rev limiter on my boxes. The popping must be from something else. Probably mis-timed sparks because of the erratic rpm pulses.

The wire between the sensor and the ignition module was replaced when the sensor was replaced because the sensor has a sealed and potted wire so it came as one piece; it has standoffs and does not come within an inch of any other wiring until it gets aft of the firewall and there it only gets bundled with the other wires going to the d-sub connectors on the ignition module.

I called lightspeed and he recommended that we check the sensor lateral alignment (i.e. make sure it is aligned with the magnet ring squarely). I feel like we did that pretty well while the propeller was off and we were replacing the sensor. It looks and measures vey well aligned!

Additionally we inspected the magnet ring. We did notice one scratch on the ring, but it is not close to the magnets. I realize there could be tolerance stacking at play, but logically if the right ignition system is working OK, with near identical sensor alignment, gap depth, wire routing, and the same magnet ring, it seems that those components are operating properly?

Lightspeed also recommended, as my esteemed VAF friends here did, that we put the oscilloscope between the right sensor and the modules and look at the waveforms. We haven't done that (yet).

It is possible that our ignition module swapping test was flawed....but doesn't seem too likely. The only thing slightly confusing while conducting the test was the fact that when I pulled left ignition circuit breaker, G3X was telling me "right ignition off", and again, that configuration (left breaker open) resulted in good tach and smooth engine operation, the same as before the module swap.

Either way, thank you for all the help. A frustrating problem to be sure, but such is the nature of complex mechanical devices, I suppose.

Here is a very blown up view of the only flaw we found in the magnet ring. I have no idea if this was there since original install while the airplane was operating fine, or if it could be associated with the problem. Keep in mind that is a very zoomed in image, the magnet ring is only 3/4" wide or so I tink.
magnetring.jpg
 
Last edited:
No Software!

...
While I don't know the s/w the box uses, I speculate that the sensors are seeing extra pules and not missing pulses. Higher than actual RPM in theory should be too many pules and not missing pulses, but this depends upon the software's error handling protocol. This points to electrical interference over wiring or magnet issues. Any re-wiring efforts immediately preceding this problem?

Larry

As alluded to previously, the LSE boxes do not have microprocessors and do not have software - Klaus advertises this as a selling point. Please see the attachment from LSE's website:

"Solid-State, Discrete logic (no microprocessor) for maximum reliability: No Software & No Software Updates"

Skylor
 

Attachments

  • LSEFeatures.jpg
    LSEFeatures.jpg
    303.5 KB · Views: 86
Last edited:
As alluded to previously, the LSE boxes do not have microprocessors and do not have software - Klaus advertises this as a selling point. Please see the attachment from LSE's website:

"Solid-State, Discrete logic (no microprocessor) for maximum reliability: No Software & No Software Updates"

Skylor

Yep.

I've been reaching out to everyone I know with a lightspeed, and finally heard back from a large organization that operates a fleet of carbon cubs with dual lightspeeds.

Their head mechanic said he's seen my problem a lot, and is pretty sure it's a broken positive wire to one of the ignition coils. He advised me to run it up on the bad ignition for awhile with the engine running rough and identify the cold cylinder.

I did a runup the other night, and since the engine popping is now occurring at a much lower RPM, I ran it up for a few minutes with the popping going on, on just the bad ignition (left only, right ign off). Initially I didn't like running it for more than a few seconds with the popping, afraid I was damaging he engine, especially when it didn't start until 2300+/- rpm. Over time, the CHT and EGT on the #2 cylinder started to drop significantly compared to the other cylinders. This was at about 1700 RPM as indicated by the tach, which is of questionable accuracy. This would seem indicate that the #2 top spark plug is not firing, firing weakly, or firing intermittently. That plug is connected to the Left ignition A coil.

I inspected that coil's wires as closely as possible. I did the troubleshooting tests from the lightspeed flow charts (I think and hope my avionics tech alraeady did this) and everything passed. Resistance betwen the coil pack +/- input terminals was 1 ohm, and the resistance between the output towers and the positive input terminal was O.L. Additionally the resistance between the center conductor and ground on the wire was O.L. (the wire coming from left ignition module A output to the ignition coil, which is the one my contact said was broken).

I replaced the top spark plug for the #2 cylinder just in case but that didn't help either. I'm thinking maybe I will try replacing the coil pack for for left ignition A and/or testing that coil pack with a breakout cable between the ignition module and coil pack.

Just gotta keep trying....
 
Last edited:
Yep.

I've been reaching out to everyone I know with a lightspeed, and finally heard back from a large organization that operates a fleet of carbon cubs with dual lightspeeds.

Their head mechanic said he's seen my problem a lot, and is pretty sure it's a broken positive wire to one of the ignition coils. He advised me to run it up on the bad ignition for awhile with the engine running rough and identify the cold cylinder.

I did a runup the other night, and since the engine popping is now occurring at a much lower RPM, I ran it up for a few minutes with the popping going on , on just the bad ignition (left only, right ign off). Initially I didn't like running it fro more than a few seconds with the popping, afraid I was damaging he engine, especially when it didn't start until 2300-+ rpm. Over time, the CHT and EGT on the #2 cylinder started to drop significantly compared to the other cylinders. This was at about 1700 RPM as indicated by the tach, which is of questionable accuracy. This would seem indicate that the #2 top spark plug is not firing, firing weakly, or firing intermittently. That plug is connecedt to the Left ignition A coil.

I inspected that coils wires as closely as possible. I did the troubleshooting tests from the lightspeed flow charts (I think and hope my avionics tech alraeady did this) and everything passed. Resistance betwen the coil pack +/- input terminals was 1 ohm, and the resistance between the output towers and the positive input terminal was O.L. Additionally the resistance between the center conductor and ground on the wire was O.L. (the wire coming from left ignition module A output to the ignition coil, which is the one my contact said was broken).

I replaced the top spark plug for the #2 cylinder just in case but that didn't help either. I'm thinking maybe I will try replacing the coil pack for for left ignition A and/or testing that coil pack with a breakout cable between the ignition module and coil pack.

Just gotta keep trying....

Have you tried swapping coil pack positions to see if the issue follows the coil?

Skylor
 
Have you tried swapping coil pack positions to see if the issue follows the coil?

Skylor

No, that's a good idea though. I might try that tomorrow, thanks! Seems wild that a faulty coil/wire could cause the faulty erratic and high RPM indications the tachometer, but I guess they can produce strong EMI if the wire is broken somewhere or they're malfunctioning?
 
Last edited:
Ok, I couldn't wait until tomorrow, and swapping the coil packs for the affected #2 cylinder didn't look too difficult, so I went out tonight and tried the swap.

Sure enough, the problem moved from the left ignition to the right ignition when I swapped the coil pack powering the #2 top and #1 top spark plugs (off the left module) for the coil pack powering the #2 bottom and #1 bottom spark plugs(off the right module). Now, the problem only presents when the ignition is on both or only the right one is on, the opposite as before.

The #2 cylinder was still the one showing low/erratic temps during the engine runup/popping, which makes sense because the problem probably swapped from the #2 top spark plug to the #2 bottom spark plug.

So I now I'm hoping a new coil pack will do the trick, although at this point I'm not going to get my hopes up too high.
 
Progress?

Ok, I couldn't wait until tomorrow, and swapping the coil packs for the affected #2 cylinder didn't look too difficult, so I went out tonight and tried the swap.

Sure enough, the problem moved from the left ignition to the right ignition when I swapped the coil pack powering the #2 top and #1 top spark plugs (off the left module) for the coil pack powering the #2 bottom and #1 bottom spark plugs(off the right module). Now, the problem only presents when the ignition is on both or only the right one is on, the opposite as before.

The #2 cylinder was still the one showing low/erratic temps during the engine runup/popping, which makes sense because the problem probably swapped from the #2 top spark plug to the #2 bottom spark plug.

So I now I'm hoping a new coil pack will do the trick, although at this point I'm not going to get my hopes up too high.

This sounds like progress!

Skylor
 
Y

I inspected that coil's wires as closely as possible. I did the troubleshooting tests from the lightspeed flow charts (I think and hope my avionics tech alraeady did this) and everything passed. Resistance betwen the coil pack +/- input terminals was 1 ohm, and the resistance between the output towers and the positive input terminal was O.L. Additionally the resistance between the center conductor and ground on the wire was O.L. (the wire coming from left ignition module A output to the ignition coil, which is the one my contact said was broken).

Coils vary a good bit, however, you typically see 2-8 ohms on the primary winding (between the two teminals you mentioned that feed power to the coil) and 10K to 20K ohms on the secondary winding (between output center terminal and ground). The fact that your secondary winding reads open likely points to a bad coil. Wasted spark coils are a bit different, so call LS and get advice on best way to measure the coils and the specs for those coils.

Larry
Larry
 
Last edited:
As alluded to previously, the LSE boxes do not have microprocessors and do not have software - Klaus advertises this as a selling point. Please see the attachment from LSE's website:

"Solid-State, Discrete logic (no microprocessor) for maximum reliability: No Software & No Software Updates"

Skylor

logic and software are mostly synonomous. Software is nothing more than logic that is read by a microprosessor and converted to binary instructions for the hardware to follow. Logic can also be burned onto a chip that does the processing of the logic and also provides those binary instructions. Either way it is the same thing once you get down to binary instructions at the hardware level. I agree that burning logic to a chip is safer than a microprocessor executing s/w instructions (several layers of flow and therefore failure points are eliminated), however, at the end of the day hardware at this complexity level doesn't do anything without instructions from something that can process logical constructs.

Larry
 
Last edited:
logic and software are mostly synonomous. Software is nothing more than logic that is read by a microprosessor and converted to binary instructions for the hardware to follow. Logic can also be burned onto a chip that does the processing of the logic and also provides those binary instructions. Either way it is the same thing once you get down to binary instructions at the hardware level. I agree that burning logic to a chip is safer than a microprocessor executing s/w instructions (several layers of flow and therefore failure points are eliminated), however, at the end of the day hardware at this complexity level doesn't do anything without instructions from something that can process logical constructs.

Larry

That’s not exactly true. It is possible and it used to be quite common to construct electronics purely from analog and discreet logic components that do not have any “burned” chips at all, and I think this is how the light speed boxes are constructed. These electronics are constructed from transistors, 74LS series discreet chips (which are themselves just arrays of transistors), 555 timers, RC and RL circuits, etc. Again, no programmable logic, PROMs, EPROMs, EEPROMs, flash, etc. Adjustable features of such systems are simply controlled by DIP switches or changing resistor values. Ignition systems are relatively simple and do not necessarily need any logic “instructions” to control their functions, including spark advance.

Skylor
 
That’s not exactly true. It is possible and it used to be quite common to construct electronics purely from analog and discreet logic components that do not have any “burned” chips at all, and I think this is how the light speed boxes are constructed. These electronics are constructed from transistors, 74LS series discreet chips (which are themselves just arrays of transistors), 555 timers, RC and RL circuits, etc. Again, no programmable logic, PROMs, EPROMs, EEPROMs, flash, etc. Adjustable features of such systems are simply controlled by DIP switches or changing resistor values. Ignition systems are relatively simple and do not necessarily need any logic “instructions” to control their functions, including spark advance.

Skylor

I didn't design it and you may be correct. However, it needs to sense a magnet pulse, relate that to TDC via an offset of degrees (inferred by time) and deal with error handling on that input, as well as a changing wave form as the RPMs change. It then needs a table that moves the timing advance based upon both RPM and data from a map sensor. Also needs an RPM threshhold to further reduce timing for starting RPMs, etc. It also needs to deal with coil dwell timing (to avoid burning up the coil at low RPMs) and numerous other factors. Just don't see how this is all done with timer circuits and transistors. The trigger pulse time interval is constantly changing with RPM, so basic timers don't help here; you need logic to translate interval spans to the delay of spark from the pulse. Degrees from trigger to TDC requires computation to know how fast the crank is spinning so that time can be used instead of degrees and that time is based upon computations based upon time between triggers, which is constantly changing with RPM. The computer cannot measure in degrees from trigger to TDC; Instead, it must use time to approximate degrees and this time is computed from the time between triggers and that time is based upon RPM.

No need to debate it. I was just stating that I didn't know the logic used for crank trigger sensing and error handling and I am pretty confident that meaningful logic is involved here, regardless of how it is done. I have installed several open source EFII systems and have a good grasp on the breadth of logic involved to make them work. That said, I am a software guy and cannot trully appreciate what can be done with the components you mention. I AM smart enough to know not to argue with a EE. One thing I know is that much logic and math is involved.

Maybe this is all done through transistors and that would explain why a failing coil could cause inaccurate RPM readings which would not occur in most EI systems, as they compute RPM based upon crank trigger counting and not coil triggering.

Lary
 
Last edited:
Well, replacing the coil pack that powers the #2 and #1 top spark plugs fixed it! I'm pretty happy about that! Thanks Van's Air Force for all your help!

Before replacement, we did a "spark test" of the coils by spinning the prop slowly by hand with the spark plug wires disconnected & fuel off, but with the coils still connected to the ignition modules and the ignition systems on. All of the other coils would send an arc between the two output towers of the coil packs at the appropriate time in the firing sequence, but the upper left coil output tower for the* #1 upper plug wouldn't arc to the #2 upper plug tower, it would arc instead from the #1 upper plug output tower*to one of the bolts that holds the coil pack bracket to the airframe. Aside from that, I'm not sure what went wrong.

Also, we noticed that the coil pack for the #1 and #2 bottom plugs was arcing through the case due to a hairline crack in both tower cases, and not from the tower tips, but changing that coil pack had no effect on the problem. The problem cleared up when I changed the coil pack powering the #1 and #2 top plugs alone.

So, yeah. On a lightspeed ignition, apparently some kind of breakdown inside a coil pack can cause the tach to read 20-30% high and swing erratically +/- 300 RPM with actual engine roughness not showing up until almost full throttle. Hopefully this helps someone else in the future.

Also, don't ban me but this was on a factory new Carbon Cub :p
 
Last edited:
Well, replacing the coil pack that powers the #2 and #1 top spark plugs fixed it! I'm pretty happy about that! Thanks Van's Air Force for all your help!

Before replacement, we did a "spark test" of the coils by spinning the prop slowly by hand with the spark plug wires disconnected & fuel off, but with the coils still connected to the ignition modules and the ignition systems on. All of the other coils would send an arc between the two output towers of the coil packs at the appropriate time in the firing sequence, but the upper left coil output tower for the* #1 upper plug wouldn't arc to the #2 upper plug tower, it would arc instead from the #1 upper plug output tower*to one of the bolts that holds the coil pack bracket to the airframe. Aside from that, I'm not sure what went wrong.

Also, we noticed that the coil pack for the #1 and #2 bottom plugs was arcing through the case due to a hairline crack in both tower cases, and not from the tower tips, but changing that coil pack had no effect on the problem. The problem cleared up when I changed the coil pack powering the #1 and #2 top plugs alone.

So, yeah. On a lightspeed ignition, apparently some kind of breakdown inside a coil pack can cause the tach to read 20-30% high and swing erratically +/- 300 RPM with actual engine roughness not showing up until almost full throttle. Hopefully this helps someone else in the future.

Also, don't ban me but this was on a factory new Carbon Cub :p

I'm glad you finally solved the problem! I suppose it's possible that the faulty coil was generating some EMI that was interfering with the tach signal.

Skylor
 
Back
Top