jusr saw this in Utube. interesting , may be a good option for some here
https://www.youtube.com/watch?v=0ICzkgBMJx4
https://www.youtube.com/watch?v=0ICzkgBMJx4
new engine supplier: based on suzuki engines
- Thread starter Caltard13
- Start date
Help Support HomeBuiltAirplanes.com:
I have spent quite a bit of time with Mark the last years at OSH. He seems to know engines. He uses some factory pieces and some high perf aftermarket stuff, depending upon what works and has proven durable.
I spent quite a bit of time discussing his PSRU, his vibration control method, how he sizes the gears, etc. This powertrain engineer is skeptical of most PSRU because many of the design/build guys do not really understand vibration and have little idea if they have done a good enough job on the topic. Mark seems to know his stuff on the topic, and I have a pretty good impression of the guy.
His 260 hp engine is based on a road engine designed for big boost, and he is intending to flat rate it to 260 hp. It will fit inside the RV6 cowl and weight limits, so it would be great at 200 hp. Anything over that is a kick in the pants. He is doing a brand new PSRU specifically for the 260 with much bigger everything. He also said that at 150 hp it will run around 0.39 lb/hp/hr or 9.5-10 gph. I am bugging him for two alternators and remote oil and water coolers, as I want the power redundancy and intend to run belly scoop coolers. We shall see, but he is looking like he will do it.
In the meanwhile, has anyone reported on successful use of the smaller engines or any failures of note?
Billski
I spent quite a bit of time discussing his PSRU, his vibration control method, how he sizes the gears, etc. This powertrain engineer is skeptical of most PSRU because many of the design/build guys do not really understand vibration and have little idea if they have done a good enough job on the topic. Mark seems to know his stuff on the topic, and I have a pretty good impression of the guy.
His 260 hp engine is based on a road engine designed for big boost, and he is intending to flat rate it to 260 hp. It will fit inside the RV6 cowl and weight limits, so it would be great at 200 hp. Anything over that is a kick in the pants. He is doing a brand new PSRU specifically for the 260 with much bigger everything. He also said that at 150 hp it will run around 0.39 lb/hp/hr or 9.5-10 gph. I am bugging him for two alternators and remote oil and water coolers, as I want the power redundancy and intend to run belly scoop coolers. We shall see, but he is looking like he will do it.
In the meanwhile, has anyone reported on successful use of the smaller engines or any failures of note?
Billski
Aeromomentum has been the subject of quite a bit of discussion here. Do a search of this site and you'll find much more, if you are interested. They seem like good powerplants and Mark Kettering has provided information here that appears, to me, to be free of hype.
The use of the Microsquirt-based system for fuel and ignition is the only thing that gives me pause. That could be worked around.
just posted another video today, some weights, costs as well, and some justification for design choices
https://www.youtube.com/watch?v=jb9mUFfjupk
Vigilant, thanks for the heads up, I did eventually found that thread after i posted the original video.
https://www.youtube.com/watch?v=jb9mUFfjupk
Vigilant, thanks for the heads up, I did eventually found that thread after i posted the original video.
Last edited:
henryk
Well-Known Member
Thanks for the video. IMO, Mark Kettering does a good job here of explaining whey they chose to convert Suzuki's rather than some other brand/model of engine (weight, and the "right" mount of complexity for aircraft use). I'm surprised he didn't add the fact that these are still in fairly widespread production around the world, so parts are still available.
It is interesting that the Aeromomentum 1.0 liter (3 cylinder) appears to use a different style PSRU than the 1.3 and 1.5L engines. The intermediate gear on that 1.0L PSRU brings the prop hub up considerably higher than 2-gear PSRUs on the 1.3 and 1.5L engines. The "laid-down" versions of the 1.3 and 1.5 liter engines cost about $2k more than the vertical versions. I'll bet there are many planes that could use the vertical versions of these engines if the prop were moved up 6 inches through use of an intermediate gear in the PSRU (like the one on the 1.0L engine). That's just me eyeballing it, it might be hard to do that in practice.
Billski or anyone else, does anyone know what engine the 75 hp AM10 is based on? It appears to be different from the others.
What are your concerns with microsquirt?
I am not an expert in Megasquirt/Microsquirt, and so these are perceptions/anecdotal accounts, not experience-based facts:
From a previous post:
I want to run an open-loop EFI (i.e. no lambda sensor) so I can burn 100LL without destroying a lambda sensor and inviting more problems.
I want rock-solid hardware. That is, proven to work in a high-electronic noise environment, components selected for high temp use.
I want a system that has engineered-in capability to provide enough engine power to remain airborne even in the case of multiple sensor failures (and not just "null" failures, but erratic responses, out-of-range errors, etc).
I >don't< want a high-level processor running a fat program with lots of features and complex interdependencies.
These are attributes that I'd guess are not "mission critical" in the "tuner" community, where flexibility and ability to optimize are more important.
Mark Kettering cites the number of engines built and hours of operation in this video: https://www.youtube.com/watch?v=xWOj5_0DoQI
They begin to look like they have enough engines and hours for me to consider their engine and PSRU reliable. I still would prefer a dual alternator and SDS running the engine...
Billski
They begin to look like they have enough engines and hours for me to consider their engine and PSRU reliable. I still would prefer a dual alternator and SDS running the engine...
Billski
Monty
Well-Known Member
It seems fashionable in some circles to bash Mega/Microsquirt.....usually people who have never used it, those who are selling something else, or people who tried to use it and failed. Or people who for some reason listen to any of the former.
I've been running it on my vehicle for 5 years....Mostly to gain experience with the system and evaluate it in a non-critical application. During that time, I've experienced multiple sensor failures, and even an alternator failure. I drove for 30min with no alternator just on the battery-no problems. I've run it open loop, and closed loop. I've even run Avgas through it. So far the computer has been the least troublesome part of the whole thing. I'd run it in an airplane. The installation is as good as the installer. Megasquirt does many of the things you are talking about, and it can be as simple or complex as YOU make it. However, the onus is on YOU to make it work. None of the other systems out there (Haltech etc) are going to be any better, in this regard. They are also going to cost a LOT more and still have many single point failure modes and inter-dependencies. I'll take a "fat" program and computer with most of the "fat" disabled that is proven.... over a "skinny" unproven system ANY day.
I suggest SDS if you don't want to take on the responsibility of engineering your own system.
You'll still need to do a good job installing it.
Those who do not know how to use an O-Scope, properly crimp and solder, or read a schematic should stick to mags and a carb, or use a pre-engineered solution. I'm dead serious about that. People with no background in electronics have ZERO business trying to install a custom fuel injection system in their airplane. It borders on insanity, and the cost of failure is VERY HIGH. Most of the time I see people complaining about mega/microsquirt on cars it's due to this problem. I do have a background in these things, and I still wanted to test it on the ground first.....
If it doesn't work.....most people blame the hardware, not the installer/tuner.
Monty
It appears to be the Suzuki G10, which is of the same engine family as the G13 (AM13) and the G15 (AM15). The G10 was used in the Geo/Chevy Metro in the US, and (AFAIK) is still being manufactured overseas.
Here's a Wikipedia page on the Suzuki G engines.
rv6ejguy
Well-Known Member
What Monty says is often true. People blame the hardware first if they have a running problem. That being said, I've assisted a couple people over the years with MS trouble shooting, in both cases it turned out to be noisy crank sensor signals causing false triggering. They ran like crap. This was 5 and 7 years ago and things may have been improved since then. I am not sure the vintage of the ECUs or software but in these cases, the MS didn't appear to have any digital or hard filtering that worked properly. I've seen similar things on Haltech and TEC systems. The factory EFI worked just fine with these factory crank sensors being used. The TEC system used its own crank trigger setup but these failed frequently and often caused hard starting issues if the battery was a bit low.
Noise is a top concern on ECUs, especially those used in aircraft. Testing in a car is a first step but your car doesn't have lots of the RF equipment aircraft have nor other electronics. I'd do lots of bench testing with these bits close to the ECU, wiring and sensors to make sure nothing crashes or glitches. Won't be good to find out in the air.
Noise is a top concern on ECUs, especially those used in aircraft. Testing in a car is a first step but your car doesn't have lots of the RF equipment aircraft have nor other electronics. I'd do lots of bench testing with these bits close to the ECU, wiring and sensors to make sure nothing crashes or glitches. Won't be good to find out in the air.
The video mentions Bosch electronics and was apparently made before Aeromomentum starting using the Microsquirt system. Customers get an engine with the Microsquirt installed and the engine test run, which is nice. For a customer buying an Aeromomentum engine and wanting an SDS ignition/FI system, the optimum situation (IMO) would be for it to be installed and test run before it is delivered. With a newly manufactured engine, that seems like the best way to avoid a lot of finger pointing and head scratching.
It should be mentioned that AirTrikes has been adapting Suzuki engines for aircraft use for many years, too. They use the SPG series of PSRUs (from Russia?) and "experienced" engines from cars. They have a long track record and, as far as I can tell, happy customers.
Just to point it out: There are mumblings (informed? Uninformed?) about quality differences between the Suzuki engine components made in Japan and those made elsewhere. I don't know if there's anything to this or not, but some sellers of used engines promote a used Japanese-made engine as being possibly a better bet than a brand-new engine made of components manufactured elsewhere. I have no way to validate any of that, it could just be B.S.
Monty, thanks for the firsthand report. Maybe I didn't sufficiently caveat my post. I'm not a Megasquirt basher, I wish them success. But, as you point out, anybody counting on that system to keep their airplane aloft needs to go in with their eyes wide open. They need to not only know stuff, but "know what they don't know." And, IMO, that includes an understanding of the software architecture used in Megasquirt, which is the only way to appraise its stability and response to various sensor failures (none of us can hope to accumulate enough time in our own car to find the failure modes that may occur in an airplane, though it looks like you are doing your part!). As a user, is it possible to "engineer" a Megasquirt system? I don't mean install various bits and make sure they run, but look at the source code and truly understand what is happening inside the black box?
Mark
Last edited:
In my life as the transmission vibe guy at Ford, I can not tell you how many times I tried to make use of an existing sensor and found the sensor was absolute crap. So much so that the calibrators had to average the signal over a second or more to get a reasonable estimate on how fast a particular shaft was turning, much less know exactly when the tooth went by for my work at vibe sensing or spark timing. There have been a lot of crappy sensors used in cars because they could get by with them. I fear that a lot of these sensors may be convenient and available and then folks try to use them with other stuff. The result is incompatible setups.
You gotta have good sensors for things like crank position. Noisy sensors are just going to poison the rest of the system. Now who do we know who carries only unconditionally "good" sensors and hardware to go with it?
Billski
rv6ejguy
Well-Known Member
Yup, we've seen some really crappy (but expensive) sensors which failed in a few hours literally.
We build our own Hall Effect sensors in house because lots of OEM ones are not reliable enough IMO to use in aircraft. Never had one of these fail electrically, 600,000 flight hours, many millions of hours in automotive. Temp sensors, we supply Bosch or GM or we also make our own for aviation. Your choice. We do mod the GM ones with a drop of epoxy on the cage to thermistor as we've seen some fail eventually from vibration. No failures after doing that. MAP sensors are OTS GM type either genuine or from a long time aftermarket supplier. These all seem pretty robust but if they fail, SDS will continue to run the engine at high power and get you back to base. Only the crank sensor is critical with SDS.
rv6ejguy
Well-Known Member
You are maybe thinking of VR sensors with a flying reluctor. Hall Effect has neither a magnet nor wire wrapped around it. A flying magnet on the flywheel or crank pulley triggers it.
MS can use many types of sensors with varying characteristics, especially so with crank position sensors. Each sensor type would have to be evaluated and run (on an engine, not just the bench) to be sure the ECU could deal with the signal and any other noise properly. VR sensors output a nasty signal for digital devices which often requires a lot of processing to clean up and stabilize the waveform. They also put out a weak signal at low rpms (like cranking). Hall Effect works at zero rpm and outputs a nice square wave signal directly which is digital friendly.
Signal difference between Hall and VR. Middle photo is VR signal. Last photo shows waveform of missing tooth VR signal.
On our system, we chose to fit all engines with the same sensor that we supplied so there is no question it will work. That also allows us to write code to remove noise from the signal and design the filtering hardware as well. The disadvantage is you can't use OEM crank sensor setups. For Lycoming, Continental, Rotax, Jabiru and Subaru, we provide sensor and bracket and in most cases the magnet mounting system as well so it's all bolt-on and engineered to work together with no surprises or problems.
Here are a couple of setups with dual sensors feeding dual ECU boards and on a Subaru EG33 and the magnet disc and Hall mount for Rotax 9 series
A MS might work just fine on an airplane. Only time would tell, but they were never designed for aviation (you are specifically warned to only use them on ground-based vehicles in fact) and don't have many of the features we do like direct fuel flow output, in-flight individual cylinder trim and integrated fault warning and diagnostics with panel mounted HID.
We have more experience with EFI on Experimental aircraft than all other brands combined and we're delighted to be supplying EFI for three of the fastest Sport Class racers at Reno this year. We're the only EFI brand to have 2 Sport Gold Class championships there- 400+ mph laps last year.
I'm not bashing MS either, they are a very capable, low cost ECU (though actually more expensive, fully assembled than an SDS one). What I am saying is that there is almost no track record in aviation with them. That alone doesn't mean they couldn't work well but it would take multiple users with thousands of flight hours collectively to prove that. Anyone is welcome to experiment towards that goal.
Last edited:
2
