# Spark redundancy.

### Help Support HomeBuiltAirplanes.com:

#### 103

##### Well-Known Member
I guess I'll weigh in. As an automotive shop owner, I still see ignition failures of coils, plugs and connections. I still see vvt cam phaser failures, I see direct and port injector failures. None are immune. Some better than others. This is why I am adding redundant spark, all sensors, and also ECU's to my conversion. I am also ridding my engine of VVT as its not even close to necessary with the loads a propeller puts on the engine. If you can make something less likely to fall from the sky without reinventing the wheel I say go for it.
Chris
What longblock are you converting? Our very own Ross flies with a single FI and EI. Very fastidious about wiring integrity and clearly believes in his own system. That said he makes dual options avaible in a very clever ways to those who are not there yet. This short video was posted about 2 weeks ago.
I am not sure he has a way for redundant ECU & Sensors to drive a single set of plugs short of four MSD8210 combiners. Eithway subscribe to his channel even if you do not plan to use Fuel injection most post are educational and lay out the options for the mechanic in charge to select.

#### Chris Matheny

##### Well-Known Member
HBA Supporter
103, I a using a 2018 GM/Opel LE2 engine from a Chevy Cruze. There is a thread about it in the auto conversion Chevy section. I am in the middle of a shop renovation at the moment plus work so progress is slow at the moment.

#### rv6ejguy

##### Well-Known Member
As someone who owned an auto repair business at the start of the EFI era and then 26 years more in manufacturing EFI systems for automotive and aviation with over 10,000 controllers sold, and thousands of injectors sold, I've never seen an OEM injector failure.

Seen infrequent failure of sensors, coils (mostly COP), pumps (mostly crappy brands or ones with many thousands of hours). The good stuff is more reliable than the engine it's attached to in most cases. Yes, some brands are better than others and some are truly bad with systemic design issues making failure frequent or even certain. Don't use that stuff in aircraft...

#### Chris Matheny

##### Well-Known Member
HBA Supporter
I am in no way saying that these failures are common place but they do still happen. There are bad designs of everything out there and also good ones. I am just trying to use my background in automotive to make sure my engine is as reliable as the aircraft systems that I worked on the 12 previous years before I started my repair shop.

#### mcrae0104

##### Well-Known Member
HBA Supporter
Log Member
Have you driven a ford lately?
What a seemingly innocent time by comparison... Thanks for the momentary escape, pfarber.

#### slociviccoupe

##### Well-Known Member
Pfarber, i know better than to drive a ford. But yes i know what you are talking about on the fords ive fixed a few of them. Bad design, bad manufacturing with a multi piece spark plug, and bad coils in general. Like you also said research and pick your power wisely. Ive been rather lucky with the hondas im used to that have very good ignition systems and are generally very well built reliable engines.

Hopefully subaru im working in does just as well. Knowing nothing about subaru's coils and reliabilityill be running the honda/denso coil on plug. They fit perfectly with just s bung welded to the valve covercand a small plastic spacer turned on the lathe to space the rubber boot to seal against valve cover.

Yes coils do go bad from over saturation (dwell) or time.

In application i am facing the stock cam sensor known to be problematic. So replacing it with a hall effect sensor.

Hopefully staying with denso or ngk iridium plugs , denso coils and good name ecu i should be ok.

#### pictsidhe

##### Well-Known Member
I am in no way saying that these failures are common place but they do still happen. There are bad designs of everything out there and also good ones. I am just trying to use my background in automotive to make sure my engine is as reliable as the aircraft systems that I worked on the 12 previous years before I started my repair shop.
Chris, you likely have the skills to make this work. What happens a lot of the time is that people install poorly thought out and implemented kludge jobs, that decrease reliability. With the very high reliability of stock automotive spark systems and the common ineptitude of builders, there is a good reason to recommend leaving the stock system alone... Read enough HBA accident reports, you'll understand why..

#### rv7charlie

##### Well-Known Member
Again, 'stock' modern computer-based automotive engine control systems in a/c is a highly questionable, and potentially death defying, venture.

#### pfarber

##### Well-Known Member
HBA Supporter
I guess I'll weigh in. As an automotive shop owner, I still see ignition failures of coils, plugs and connections. I still see vvt cam phaser failures, I see direct and port injector failures. None are immune. Some better than others. This is why I am adding redundant spark, all sensors, and also ECU's to my conversion. I am also ridding my engine of VVT as its not even close to necessary with the loads a propeller puts on the engine. If you can make something less likely to fall from the sky without reinventing the wheel I say go for it.
Airplanes have an annual condition inspection. You can add whatever checks you like above the FAA's required inspections. Meter the coils, test them on stand etc. Heck you could do it every 100 hours if you like. ECUs will tell you the second there is a misfire then you can immediately land.

If you were going to run ONE coil and a distributor, I might think of a second spark source (aka dual mags). But COPs are EACH cylinder. You lose one, you still have 3/5/7 cylinders firing.

But a second spark on a single plug? No, would not even consider the effort.

#### pfarber

##### Well-Known Member
HBA Supporter
Again, 'stock' modern computer-based automotive engine control systems in a/c is a highly questionable, and potentially death defying, venture.
Please list the NTSB reviews where the ECU was the primary cause of failure.

There has to be at least one or two?

#### rv7charlie

As much as I respect the NTSB's ability to investigate air carrier crashes and train wrecks, I'd be downright shocked if they could figure out that a homebuilt a/c crash was caused by anything other than pilot error. In fact, unless there's a death or substantial (million$) property damage, they typically don't even investigate homebuilt crashes at all. I have seen numerous NTSB conclusions in homebuilt accident cases where it's blatantly obvious that the NTSB either ignored or had no understanding of factors directly related to the cases. I've also seen numerous cases where, when they recognized their lack of knowledge, they called as expert advisors the persons or companies who produced and sold the devices in question. This obviously calls into question the objectivity of the expert. I can't point to specific NTSB reviews where the ECU was blamed, but given my admitted bias, I've never bothered to look. And for every accident, there are uncounted but significant numbers of forced landings due to partial to total engine failure, that didn't result in any kind of accident. I know; I was one (with a Lyc). I have seen numerous 'internet accounts' (worth whatever value you're willing to assign) from people who've had factory ECUs go into 'limp mode' for various known and unknown reasons. Limp mode will typically allow driving a car home after a failure, but rarely supplies enough power to keep an a/c in the air. The whole problem with automotive ECUs is that we don't know what we don't know. We can't know all the 'hooks' the automotive engineers put in the software. We might hit a conflict in the software on the 1st takeoff, or in the 200th hour of operation. Everyone will do what they're comfortable with. I'd probably marginally more comfortable flying a generically 'older' (earlier) model ECU, simply because there were fewer things unrelated to the engine that had been coded into the ECU (door locks, alternator, break-in alarm systems, etc). If I could obtain a factory automotive controller that is purely 'engine specific' in its controls, with a clearly defined sensor/load set, I might well be enthusiastic about using it. For example, there are supposed to be some GM 'crate' engines that come complete with controller, that are intended to be 'drop-in' conversion engines in custom car builds and have no transmission or other interfaces. I'd likely be comfortable with something like that from a safety standpoint, but I'd be left without the ability to manually control mixture. Many attempts to work around mixture control in piston a/c engines, but none seem to be truly effective. Charlie #### 103 ##### Well-Known Member As much as I respect the NTSB's ability to investigate air carrier crashes and train wrecks, I'd be downright shocked if they could figure out that a homebuilt a/c crash was caused by anything other than pilot error. In fact, unless there's a death or substantial (million$) property damage, they typically don't even investigate homebuilt crashes at all. I have seen numerous NTSB conclusions in homebuilt accident cases where it's blatantly obvious that the NTSB either ignored or had no understanding of factors directly related to the cases. I've also seen numerous cases where, when they recognized their lack of knowledge, they called as expert advisors the persons or companies who produced and sold the devices in question. This obviously calls into question the objectivity of the expert.

I can't point to specific NTSB reviews where the ECU was blamed, but given my admitted bias, I've never bothered to look.

And for every accident, there are uncounted but significant numbers of forced landings due to partial to total engine failure, that didn't result in any kind of accident. I know; I was one (with a Lyc).

I have seen numerous 'internet accounts' (worth whatever value you're willing to assign) from people who've had factory ECUs go into 'limp mode' for various known and unknown reasons. Limp mode will typically allow driving a car home after a failure, but rarely supplies enough power to keep an a/c in the air. The whole problem with automotive ECUs is that we don't know what we don't know. We can't know all the 'hooks' the automotive engineers put in the software. We might hit a conflict in the software on the 1st takeoff, or in the 200th hour of operation.

Everyone will do what they're comfortable with. I'd probably marginally more comfortable flying a generically 'older' (earlier) model ECU, simply because there were fewer things unrelated to the engine that had been coded into the ECU (door locks, alternator, break-in alarm systems, etc).

If I could obtain a factory automotive controller that is purely 'engine specific' in its controls, with a clearly defined sensor/load set, I might well be enthusiastic about using it. For example, there are supposed to be some GM 'crate' engines that come complete with controller, that are intended to be 'drop-in' conversion engines in custom car builds and have no transmission or other interfaces. I'd likely be comfortable with something like that from a safety standpoint, but I'd be left without the ability to manually control mixture. Many attempts to work around mixture control in piston a/c engines, but none seem to be truly effective.

Charlie
Or Plug warning... You could buy a ECU that never had any limp home code added and likely 10% of the lines of code an Automotive ECU ever had. SDS EM5 Ref SDS EM-4: Aircraft why settle for marginal comfort when you can mitigate all of those unknowns leaving you only with system decisions and wiring integrity to manage.

Is you life worth more than +$1200 over the budget execution? No association other than repsect for the accumulated track record Ross has. #### rv7charlie ##### Well-Known Member 103, 'You talking to me?' ;-> SDS is certainly a good plan (unless you fly a rotary; no support there). I suspect that the allure of the original ECU is that it's there, and configured (they think). But with even the SDS, there's a ton of tuning ahead of you if you're flying anything other than an otherwise stock Lyc. Charlie 103 #### pfarber ##### Well-Known Member HBA Supporter I would not run a factory, stock ECU. Lots of things out of your control. 3rd party ECUs have much more flexability. At a minimum you should retain the CEL logic. People routinely fly and never check the gas, or oil gauges. A big honking red light would be invaluable. This is one thing that should be, at a minimum, required by a pilot flying an ECU. SDS fails in the this regard. A tiny LED is all they offer. I would like to see a remote idiot light. Megasuirt is quite configurable. Either monitor the sensors or not. And then you can either apply a 'mitigation' or just CEL is and ignore it. #### Daleandee ##### Well-Known Member I know better than to drive a ford. I've had great service from my little Ford Ranger 4 cylinder, 5 Speed. But judging from the wife's previous 2012 Ford Focus I'd agree with you. Some of the designs on that car came from folks that need to get the money back they spent on engineering classes. Yes coils do go bad from over saturation (dwell) or time. True and one reason why some ignition systems should not be left "on" unless the engine is running. I'm using the genuine Bosch blue coils. They have a good reputation for longevity. In application i am facing the stock cam sensor known to be problematic. So replacing it with a hall effect sensor. The WW Corvair system uses a hall effect sensor for primary and a set of points for the backup ignition. Works very well and no known failures (unless you reverse the polarity on installation). Hopefully staying with denso or ngk iridium plugs , denso coils and good name ecu i should be ok. I like the Denso Iridium plugs and have been running them for a few years now. I generally change them out at annual but this year was a clean, gap, and reinstall session as there wasn't many hours on them (thanks to Covid). Seems to me you have a good plan. #### pfarber ##### Well-Known Member HBA Supporter Or Plug warning... You could buy a ECU that never had any limp home code added and likely 10% of the lines of code an Automotive ECU ever had. SDS EM5 Ref SDS EM-4: Aircraft why settle for marginal comfort when you can mitigate all of those unknowns leaving you only with system decisions and wiring integrity to manage. Is you life worth more than +$1200 over the budget execution?

No association other than repsect for the accumulated track record Ross has.
Once you get into custom ECU having the ability to monitor, mitigate or ignore a bad input is better than just blindly saying 'all limp home is bad'. On a STOCK ECU, yes, you have little idea what will happen.

On a custom ECU you have all sorts of control to monitor, failover or switch to a backup.

A perfect example... your oil temp is reading high. Having just a CEL light (a tiny led on the SDS units) and no way to monitor/resolve means you're going to land ASAP. But a custom ECU (with say a CAN bus for monitoring) means that you could dial in and see that sensors exact reading. Is it 0v? 5v? jittering? Can you clear it? Can you switch in a secondary sensor?

In a stock car ECU none of that is possible. And depending on the custom ECU you use, your options are also dependent on the ECU. I've seen many touch screen CAN bus gauge clusters for custom ECUs. Why this is not a standard for AC is beyond me. Not to pick on SDS, it is a fine product, but there is no way I settle for a two line LCD for my engine. Can bus that display, allow for touch screen modification, and now you have an out of the box solution that works for me.

#### rv7charlie

##### Well-Known Member
That sounds like you are confusing/conflating ECUs with engine monitors. I seriously doubt that anyone who goes to the trouble of installing a full engine control system would then *not* install a comprehensive engine monitoring system. None of the dedicated a/c engine controllers I'm aware of make any claim to being engine monitors; they typically only give access to parameters that are already needed for the controller's operation (MAP, RPM, etc).

Charlie

#### pfarber

##### Well-Known Member
HBA Supporter
That sounds like you are confusing/conflating ECUs with engine monitors. I seriously doubt that anyone who goes to the trouble of installing a full engine control system would then *not* install a comprehensive engine monitoring system. None of the dedicated a/c engine controllers I'm aware of make any claim to being engine monitors; they typically only give access to parameters that are already needed for the controller's operation (MAP, RPM, etc).

Charlie
I think you are not aware that custom ECUs like Megasquirt have options that SDS and factory ECUs to not allow.

Custom ECUs monitor all critical engine parameters, (again, just as an example) I don't believe that SDS has a CAN BUS output to allow for custom data monitoring/gauges. Custom ECUs usually have more IO ports for things like EGT/CHT that can also trigger service lights or fail overs.

Please take a look at custom ECUs that are you there, you will see that they far exceed factory ECUs and some current 'bolt on' EFI/ECUs.

#### rv7charlie

##### Well-Known Member
I think you are not aware of what I'm aware of.

#### pfarber

##### Well-Known Member
HBA Supporter
I think you are not aware of what I'm aware of.
"confusing/conflating ECUs with engine monitors "

Well this statement vs what I have said makes me wonder if you do, also.

An ECU is BOTH. Why you think they are not would only be due to the manufacturers decisions. Again, most custom ECUs allow for full data display via CAN BUS, making them monitors, as well a controls.

Even in a car, the ECU is also an engine monitor. What do you think is controlling that gauge cluster??