# 100LL and O2 sensors; your thoughts

Discussion in 'General Auto Conversion Discussion' started by pfarber, Oct 13, 2019.

1. Nov 1, 2019

### dtnelson

#### Member

Joined:
Apr 30, 2019
Messages:
19
15
Hi pfarber - thanks again for a good discussion. And thank you for correcting me regarding the number of engines certified for auto fuel!

Unfortunately, detonation is still a real thing. Even with our relatively low compression engines. Why? Because, all things are not always perfect, especially in the experimental world.

A cylinder is running hotter than you'd like, but you're flying anyway. You inadvertently leave the carb heat on during a go around. You have everything needed for detonation.

You have a clogged air bleed nozzle (fuel injector). You have everything you need for detonation.

A spark plug fails, causing a hot spot. You have everything you need for detonation.

You are running lean at high power settings. You may have everything you need for detonation.

Detonation certainly does happen in aircraft engines. See https://www.lycoming.com/node/17607, or https://www.airmarkoverhaul.com/detonation-pre-ignition/ or https://www.savvyanalysis.com/articles/detonation-and-pre-ignition, or (my favorite)

https://www.avweb.com/features/pelicans-perch-43detonation-myths/

Which makes several of pfarber's points, and many more. It's a good read, and a very good discussion of detonation (and pre-ignition).

Given that I race my airplane, and regularly run at fairly high power settings, yup, I should be concerned, or at least aware, of detonation. Even with my "low" compression IO-360 and 100LL. I'd argue you should be too.

Which (if you'll allow), get's us back to the original discussion - 100LL and O2 sensors (and from there, a closed loop system). Given the real world things that can (and do) happen to our engines, to really get the kind of performance that's been discussed here, detonation has to be considered and avoided. That's why I argued (and argue) that a knock sensor is part of the solution. The problem is, our engines are so "noisy" that getting a knock sensor to work is a very difficult problem. Given the small market (as has been pointed out), getting people other than experimenters to invest in development of a true closed loop system, with all of the appropriate safeguards, is unlikely.

Thanks again -

Dave

2. Nov 1, 2019

### blane.c

#### Well-Known MemberHBA Supporter

Joined:
Jun 27, 2015
Messages:
3,315
574
Location:
capital district NY
Getting anywhere in aviation is like herding cats.

3. Nov 1, 2019

### pfarber

#### Well-Known Member

Joined:
Feb 21, 2019
Messages:
303
19
Location:
Pennsylvania
But the reasons for no ethanol can easily be remedied. If you told a Flight school that burns 100's of gallons a day that they could STC their planes to run regular pump fuel do you not think they would just at the chance? Your ROI would be weeks. Adding a few $100 to the cost of an annual to replace some rubber fuel parts is not onerous. Not compared to some of the GA wide AD's that were forced on owners that cost$1000's out of pocket. And once 100LL is dead, the cost of gas, and flying, almost drops 30% or more. Want more pilots? Want pilots with more time? Reduce the cost per hour of flight.

The FAA rammed ADS-B down our throats, they can do the same to phase out 100LL.

In the past airports had different grades of fuel onsite. I remember having 100 (green) and 100LL (blue) at the airport. There wasn't a huge issue with people pumping the wrong grade.

4. Nov 1, 2019

### pfarber

#### Well-Known Member

Joined:
Feb 21, 2019
Messages:
303
19
Location:
Pennsylvania
Dude, if you have nothing CONSTRUCTIVE to contribute, then please don't post.

5. Nov 1, 2019

### pfarber

#### Well-Known Member

Joined:
Feb 21, 2019
Messages:
303
19
Location:
Pennsylvania
"With the given that the fuel is the correct choice for the engine, for a pilot the number one cause of detonation is excessive leaning at high power settings."

I would think that a simply sensor like used by current ECUs to detect the vibration of detonation, and alert the pilot/or engine management system that something is wrong is a simple answer. Lycoming, in the document you cited also mentions:

"Light detonation may not show any indication in the cabin of the aircraft. Moderate to severe detonation could be noticed as engine roughness, vibration or loss of power and eventually engine damage. "

So there is no way for a pilot to know that he's to lean, as you'll probably never hear the knock.

Octane is A solution, but in 2019 there are OTHER ways to detect and prevent detonation. Car makers have been doing it for years. The only reason Lyc/Cont aren't doing anything about it is because either its 1. not that great an issue, or two 2. if you blow a cylinder, they make money.

6. Nov 1, 2019

### Topaz

#### Super ModeratorStaff Member

Joined:
Jul 30, 2005
Messages:
13,840
5,466
Location:
Orange County, California
Or (and more likely, IMHO) "3. The FAA makes it so horrifically expensive and time-consuming to make even a minor change to an aviation-certified engine such that, unless it's something that makes the 'upgraded' engine hugely more profitable, there's no business case to be made for doing it."

Especially with the strong cultural push towards electrics for light aviation in the last five years, there really isn't any incentive for the certified-aircraft-engine manufacturers to spend research and development money, certification money, marketing money, and all the associated overhead, with developing an engine with things like knock-detection and so on. There aren't really any new type-certificated designs on the horizon to use such an engine, and all the existing type-certificated designs would need an STC to use one. Which STC has to be paid for by the aircraft manufacturer. Once you add it all up, there's just no economic justification for doing it.

gtae07 likes this.
7. Nov 1, 2019

### TFF

#### Well-Known Member

Joined:
Apr 28, 2010
Messages:
11,721
3,320
Location:
Memphis, TN
Flight schools can use the auto fuel. Most flight schools ride in the infrastructure of the local airport. Having own fuel in the type of storage the airport requires is generally too costly. Cheaper to pass it to the student than buy 3000 gallon batches. Buying the fuel tank probably requires permits and comp for fuel sales will make the airport resist fuel competition.

In the past they did have multiple fuels. When fuel was cheap and people tended to bend rules less. I think today given $3 unleaded and$4 100ll and the plane is supposed to have 100ll, the unleaded is going in and hope; even if it’s not right. The FAA is just as lawyered up and will always give the supper safe answer. Remember they protect the public not you and your toy that can hurt the public. You gripe about it, but I see them grounding everything if it becomes too much a hassle. They don’t have people for enforcement so they can’t check on the system like the could 30 years ago. Airlines would love for us to go away.

8. Nov 2, 2019

### pfarber

Joined:
Feb 21, 2019
Messages:
303
19
Location:
Pennsylvania
I've never seen vetted information for the cost to certify a modification. I mean they do it all the time. The people that STC'd MOGAS did it for dozens of AC and sell if for $300. So certain things can't be THAT expensive. The FAA certification document for engines is not that difficult, its just about 200 hours on a dyno. So the FAA's cost can't be that much. Now how much the engineering costs? Well that's an R&D cost, not a certification cost. So take a externally mounted knock sensor. Its epoxied to an engine case, with a wire and light to the EMC or instrument panel. That's just not going to be a lot of money. So what's the problem? Well, Lyc/Cont don't care about your fuel costs. If anything they'll make more money as more engines need overhauls because people would actually FLY the planes since it magically just got 30% cheaper. Look at all the small companies that have STC'd engine accessories, there are now EI modules, MOGAS, exhaust, HP boosts by swapping jugs etc etc etc. There are tons of STCs for engines... and they didn't cost the millions most people think it would cost. 9. Nov 3, 2019 ### TFF ### TFF #### Well-Known Member Joined: Apr 28, 2010 Messages: 11,721 Likes Received: 3,320 Location: Memphis, TN Depends on the STC. Fuel. Big deal. I forget which auto fuel STC but every plane on the STC checked. Like in three days. They had a crew and a fuel bowser. They had planes fly in all day. They wold put fuel in, start, power run, get engine hot, and do a heat soaked start. Mooneys vapor locked, that’s why you don’t see them on the list. I know the helicopters I work on have the option on the table for composite blades. It will cost$2M to get the first set and qualifications done.

STC for an eyeball vent not very much but more than you think, a couple of thousand.

I have a friend who bought an STC design for extended range fuel tanks. It would cost a hundred thousand today to get done. They could only sell about ten if they make some. I think the whole design buy was $5000. These legacy stcs are cheap if you can know what to do with them. Doing one today is not cheap. Lots of performance enhancers are generally crossover where there is a certified version out there and they convinced someone to turn the old spec into new. You can’t do that much today. The FAA inspectors now will refer you to a DER. This is all homebuilt stuff. No legal requirements to put any type of fuel in. You can run auto gas. Do your own mods and get flying. 10. Nov 3, 2019 ### pfarber ### pfarber #### Well-Known Member Joined: Feb 21, 2019 Messages: 303 Likes Received: 19 Location: Pennsylvania But again, this is all speculation. That$2MM includes what? I'm sure the accountants have already determined that they will make money, otherwise they wouldn't put up the initial costs.

The 'would cost hundred of thousands today'... citation needed

Look at all the STC glass cockpits that graduated from E/AB to certified?

Yes, the costs are variable, does anyone have a REAL numbers for comparison? I wouldn't even know where to start to look.

11. Nov 3, 2019

### TFF

Joined:
Apr 28, 2010
Messages:
11,721
3,320
Location:
Memphis, TN
The little Garmin stuff is a good deal certified but the homebuilt stuff is way better for the price. You want big screens in a certified plane and it’s $60k easy. Those little retro fit Garmins and Dynons are great at 5k for a Cherokee. The FAA did do something extraordinary. Dynon kind of got the shaft though. They supplied the reliable data over what, 10+ years, then Garmin swoops in with a unit 1/3 the size same cost. Garmin was able to use its muscle and probably had the little unit stillborn and never released it, because until recently Garmin could have given a flip about homebuilts they want certified. There are some STCs for 390s in place of 360s that popped up pretty easy. LED lights. You want to put a Lycoming in place of a Continental, you have to turn over a ready plane to the FAA and they fly it around 100 hrs on your fuel dime. That’s all after you have a stack of reliability reports that they require. DER salary, tooling, and parts you wear out are all on you. As for the blades, they decided it was too expensive. Yes easy 10% performance increase. It would take 60-70 aircraft to break even in just the development costs. 12. Nov 12, 2019 ### pfarber ### pfarber #### Well-Known Member Joined: Feb 21, 2019 Messages: 303 Likes Received: 19 Location: Pennsylvania We may be seeing this differently. CERTIFICATION costs are what it would cost to get the STC. R&D costs are what a company spends to bring an idea to market. The company would need at rough estimate very early one before tooling was committed. So where are 'they' getting the numbers? Is there a 1-800 certification quote number we can call?? My point is that the cost to CERTIFY is not that big. The R&D is wildly dependent, depending on what it is. 13. Nov 13, 2019 ### TFF ### TFF #### Well-Known Member Joined: Apr 28, 2010 Messages: 11,721 Likes Received: 3,320 Location: Memphis, TN To me the cost of R&D is part the cost to certify. Why else are you doing it? Until then, that last page of paper delivered to the FAA is worthless, and the FAA would laugh if there was not a stack of paper under it. It’s all your problem if you want an STC. Idea, design, build, test, deliver to FAA for their tests, then receive STC. If financially viable or sellable has nothing to do with actual STC. That’s business. 14. Nov 13, 2019 ### gtae07 ### gtae07 #### Well-Known Member Joined: Dec 13, 2012 Messages: 1,819 Likes Received: 1,087 Location: Georgia Certification costs are baked into the product from the very beginning. If you intent to certify anything other than the most basic item you need to be thinking about it before you start your design. The FAA is going to require more documentation than you would even think possible to create. I work on the certified side of aviation (at an OEM) and it amazes me every day that we progressed beyond the 707. The paperwork and testing and documentation burden seems to get exponentially larger with each new airframe. Last edited: Nov 13, 2019 15. Nov 13, 2019 ### dtnelson ### dtnelson #### Member Joined: Apr 30, 2019 Messages: 19 Likes Received: 15 Unfortunately, certification is not the issue for adding a knock sensor. The issue is engineering... You can think of a knock sensor like a little microphone & signal processor stuck on the side of your engine. The noise it's listening too, and the noise it's listening for, both must be characterized, and it's going to be a bit different for each engine. It's not like you can just take a knock sensor from your Buick and epoxy it onto the side of your Lycoming... If it where as simple as that, it'd already be a thing in the experimental world... AdrianS likes this. 16. Nov 14, 2019 ### AdrianS ### AdrianS #### Well-Known Member Joined: Jul 5, 2014 Messages: 403 Likes Received: 250 Location: Australia Knock sensors are better combined with crank / cam position information: that way you can "window" the result : only look for knock at certain points in the cycle. Sensor position is also important. These things can be sorted by trial and error- but you have to try make the engine knock, so you know you're detecting it. Preferably at full load. It's easy for Toyota... 17. Nov 15, 2019 ### pfarber ### pfarber #### Well-Known Member Joined: Feb 21, 2019 Messages: 303 Likes Received: 19 Location: Pennsylvania No. Because you can design a widget for E/AB and never certify. So the two costs are completely separate. Things like FI, EI, EFIS etc can all be designed to meet FAA requirements but never leave the E/AB space. In fact, most start in the E/AB arena THEN get certified. I get that you see them as interconnected, and in the pre-engineering/market survey stage there will need to be an educated guess as to the cost. Again, read the AC for certifying an engine. Its less than 200 hours on a test stand. Once you submit a test plan to the FAA and they sign off on it, your costs are minimal. Now if that test shows that your engine DOESN'T MEET THE REQUIREMENTS, then that's not a certification cost, that's your R&D team not designing a suitable product. Your widget has to do X, it didn't. If your widget takes$5MM to design to meet the FAA test regimen, THAT NOT THE COST OF CERTIFICATION. That's simply a design parameter.

18. Nov 15, 2019

### gtae07

#### Well-Known Member

Joined:
Dec 13, 2012
Messages:
1,819
1,087
Location:
Georgia
I'll say again... oh, wait, you're serious? HAHAHAHAHA!

Certifying an engine is orders of magnitude more work than just 200 hours of testing on a test stand. There may be hundreds or even thousands of hours of testing (all-up or component-level) conducted to meet other requirements in the regs that aren't included in that 200 hour limit.

There's a lot more to certifying something (especially a type-certified product like an engine or aircraft) than just submitting a test plan to the FAA and running tests; there's a whole lot more to it than just testing the final as-built product. If you showed up on the FAA's metaphorical doorstep with a finished product and some test results and asked for your cert documents they'd laugh in your face.

See, the FAA isn't just concerned about the performance of the final product. They want to know what you're planning to make, how you're planning to design, certify, and test the product... and they want to know before you get started. You need to settle with the FAA what certification basis you're using for your product, cover any and all special conditions or issue papers that might come up, come up with a certification plan, and so on. The FAA wants to know how you're going to show compliance to everything (and by which method--test, analysis, similarity, etc) and how you're going to manage all the various aspects of configuration control, documentation, traceability, etc... and that's before you get started.

Then, you can sit down and start designing your airplane. Keep in mind that throughout the entire design, you're going to have to keep records of how every single part and assembly meets the myriad regulations for design and construction. The FAA wants to know. For all your official "for score" testing, which is a whole lot more than just 200 hours of run time, you need a DER or equivalent to witness in person. You'll need stacks of reports and analysis, all of which need to be reviewed. You need to document every process and material you're using and show the FAA that those processes and materials meet safe, repeatable, suitable standards.

And even once all that's done, and you've turned in the truckloads of paperwork that's all been signed off by your DERs, and all your thousands of hours of testing are completed and the FAA signs off your design... now you get to enter the fun world of production certification, where the FAA more or less approves your production line (and everything that goes with it, like your quality manual and procedures, your work flow, your parts tracking process, your MRB process... on and on and on). That's a whole separate ball of wax that takes piles of paperwork on its own.

My point is, final all-up testing (e.g. your 200 hour example) is just a tiny piece in a very large puzzle. So many people think it's simply a matter of showing up with some test records one day and saying "look, I meet Part 23!" and it's not. If it were that easy, we wouldn't see so many Experimental-to-Certified projects taking so long, changing so much, or never making it to a finished product at all.

You can design to the standards all you want. Anyone can do that. The trick with actually certifying something vs. just designing to the top-level standards is basically record-keeping, documentation, and testing--and knowing what the standards actually mean, in the FAA's eyes. Learning how to play the game and do things "the FAA way" is where all the expense in certification is generated.

In short, it's not good enough to meet the FAA's standards. You have to show that you meet the standards, you have to show how you meet the standards, you have to show how you're going to show that you meet the standards... and you have to show the FAA all of this the way the FAA wants you to show them. Learning how to do all that is what bites so many experimental-to-certified programs (and really, even airliner designs from new companies).

And this is just the FL300 view. The amount of paperwork, administrative work, and other work that isn't directly designing something or testing the final product is absolutely mind-blowing. The old saw about the weight of the paperwork equaling the weight of the airplane isn't an exaggeration--it's an absurdly low estimate. Maybe 60 years ago under CAR 3 this was feasible... but not any more. All in the name of "safety", of course.

Last edited: Nov 15, 2019
mcrae0104, TFF and BJC like this.
19. Nov 15, 2019

### BJC

#### Well-Known MemberHBA Supporter

Joined:
Oct 7, 2013
Messages:
9,714
6,505
Location:
97FL, Florida, USA
Gee, gt, it sounds as if you actually work in the industry as an engineer and have relevant experience in dealing with the FAA.

BJC

gtae07 likes this.
20. Nov 15, 2019

### pfarber

#### Well-Known Member

Joined:
Feb 21, 2019
Messages:
303
19
Location:
Pennsylvania
That's a lot of words to try and mix oil and water.

Other than people that have certified an engine or airframe, none of us know the true cost.

My point is that the certification costs cannot be defined as anything more than the cost of the testing to get the TC/STC.

I spend $8k on flight training, but the tests to get a pilots license is$500+$150 (practical test + written test). How much does it cost to get a PPL cert (aka certification cost)?$650.

How much does it cost to learn to fly (aka R&D) \$8000.

Does that clarify it??