Any questions you have before you begin buying, building and installing.
#43270
Greetings. I've spent some time getting familiar with the Speeduino world, but will admit that this post is a blatant attempt to take a shortcut and get quick answers rather than do all the reading I know I should do. I promise I'll do the reading later, but getting some guidance out of the gate will be a big help. Please forgive the newbie questions.

The project: A B&S V-twin engine (627 to 810cc). Right now I'm planning only to do EFI and to stick with the OEM "Magnetron" ignition (fixed timing), but I'd like to have the option to perhaps fit a more advanced ignition later. In this application, reliability is more important than fuel economy. Manual mixture adjustment in real time (based on EGT, a wideband O2 sensor, etc) would be acceptable. To keep things simple, I'd think batch injection will be just fine.

1) Would the NO2C board be the right choice? Space is not a giant issue, so if there are advantages to going with a larger Speeduino board (better cooling, easier assembly, more flexibility, etc) I could accommodate a larger board. But, if there's no downside, then the the smaller NO2c board would seem best.

2) MAP sensing: The induction cycle for these V-twins is obviously very uneven, so having an injector for each cylinder will be great. In the carbureted versions, the induction and exhaust runs are kept entirely separate for each cylinder to help avoid having uneven induction pressures/fuel amounts into the cylinders. In a Speeduino installation, would each cylinder get its own MAP sensor? Some people doing tuning on these engines have trouble getting good MAP readings on individual cylinder MAPs, I'm guessing this could be because of the long lag time between induction cycles (i.e. long periods with zero flow, then a relatively short flow period). Will I need to use a special MAP sensor or employ some sort of data smoothing/averaging to get good readings for use in computing fuel requirements?

3) RPM sensing: To keep things simple, is it likely I can just use the OEM magnetron trigger magnet on the flywheel to trigger whatever sensor I'll use for RPM (Hall effect, probably)? This won't give information on what the valves are doing, but just engine speed. Good enough for batch injection?

4) Throttle bodies: Any known good (inexpensive) dual-throat candidates in the 28-30mm range? Some of the motorcycle throttle bodies for 250cc-400cc 4 stroke engines look like good candidates.

5) Fuel pressure: The fuel flow requirements of these engines are pretty modest (2.5 GPH / 10 LPH max). Is it likely I'll be able to use a fairly low fuel pressure (7-15 psi / 0.5-1 bar) in my fuel delivery line? The alternators in these engines aren't big, and I'd prefer to use a pump that draws less current.

6) Speeduino power filtering/EMI characteristics: I haven't done much analysis on how "clean" the power is from this engine/electrical system (particularly during starting), and I don't know much yet about the shielding of the ignition system components. My guess is that clean power and low EMI might not have been high priority issues in the design of these industrial engines. Is the Speeduino hardware already tolerant of voltage excursions, etc, or should I consider taking extra steps?

7) Any known similar projects I should be reading about, or folks who frequent these boards and who spend their time fitting EFI/electronic ignition to these small industrial engines?

Thanks much for your assistance.

Mark
#43280
Have a search on here as there is a lot of info from others who have done similar engines.

Although your fuel demand is low ,don't try and run injectors at low pressure, they won't work correctly. You can get small sized pumps and if power is an issue you could even pwm the supply to the pump too.
You need to devise a better trigger wheel setup for even just fuel only.
#43284
dazq wrote:
Mon Jun 01, 2020 7:07 am

You need to devise a better trigger wheel setup for even just fuel only.
Just a single input once per cycle will actually be enough for fuel only.

Doesn't need to be accurate.

These pumps use 12W and should be enough for what you are doing.

https://www.aliexpress.com/item/3278964 ... hweb201603_

For MAP just have a normal map sensor connected to both runners via vacuum tube and a T.

But the above suggestion is best.

You claim you don't want to do the reading but i'm afraid there isn't really any way around it unless you have someone else do the job for you.

Read through the wiki.

What you are doing is simple in Speeduino terms. The hard part will actually be building everything around it.
#43300
dazq, thanks for the reply.
dazq wrote:
Mon Jun 01, 2020 7:07 am
Have a search on here as there is a lot of info from others who have done similar engines.
Yes, I'll keep learning from what they've posted. There's no great honor in being the >second< person to make the same mistake. :D
dazq wrote:
Mon Jun 01, 2020 7:07 am
You need to devise a better trigger wheel setup for even just fuel only.
I don't know how flexible Speeduino is (I have more reading to do), but with batch injection and modest RPMs (4000 RPM max), I would think I could get by with injection pulses at once per crank revolution (twice per induction event), and the timing wouldn't be very critical. The injectors don't need to be at the TB, I could place them close to the port where wall temps are higher and condensation/wall wetting would be less of an issue.

Thanks again.
#43303
theonewithin wrote:
Mon Jun 01, 2020 11:41 am
Just a single input once per cycle will actually be enough for fuel only.
Doesn't need to be accurate.
Per my response to dazq, those were my thoughts, too. But, I really don't know enough (yet) to have a credible opinion.
theonewithin wrote:
Mon Jun 01, 2020 11:41 am
These pumps use 12W and should be enough for what you are doing.
https://www.aliexpress.com/item/3278964 ... hweb201603_
Thanks for the link. I've also read that the injector "ON" time can be too short, that a lower pressure and longer pulse width can give better results. With the small fuel requirement of these engines, I was concerned that using automobile-type pressures and automobile-type injectors could result in problems.
theonewithin wrote:
Mon Jun 01, 2020 11:41 am
For MAP just have a normal map sensor connected to both runners via vacuum tube and a T.
I'll have to think through how that would work. On the face of it, it would seem that, with a "T," during the intake pull in one runner (low pressure), the MAP sensor wouldn't see that pressure reduction because the reduced pressure would just prompt flow through the "T" from the other runner. If we try to overcome this with check valves to each side, the pressures seen by the sensor have no way to go back to the ambient pressures (i.e. the outside air pressure (approx) that would be present in the induction runners between the throttle and the intake valve for the approx 75% of the time that nothing is happening there).
theonewithin wrote:
Mon Jun 01, 2020 11:41 am
But the above suggestion is best.
Sorry, I'm not clear on what suggestion you are referring to.
theonewithin wrote:
Mon Jun 01, 2020 11:41 am
You claim you don't want to do the reading but i'm afraid there isn't really any way around it unless you have someone else do the job for you.

Read through the wiki.
Sorry if I gave the wrong impression. I'm definitely willing to read the documentation, but getting a little perspective on this particular project will help me put that information into context. For example, the wiki tells us what type of TPS to use, but I didn't see an option to do away with the TPS entirely (run a little rich, accept a little lag on acceleration. In this application, that has worked well for others when fitting EFI ).
theonewithin wrote:
Mon Jun 01, 2020 11:41 am
What you are doing is simple in Speeduino terms.
I agree. When I see some of the other projects people are doing, I think Speeduino may be overkill for this one. I want things to be very simple: a sensor that isn't present can't fail. A software subroutine that isn't present can't cause a crash. I don't need variable valve timing, I don't want closed loop, etc. Just "good enough" open loop operation to make power at all RPMs (maybe based on RPM and MAP alone, perhaps a "START" switch to give 5 seconds of enrichment for starting, thus eliminating the need for an engine temp sensor?). Manual trim knobs to allow the user to adjust mixture to each side (by use of the EGT reading, or by RPM drop) would be good enough--I think!

Thanks again!
#43309
Yes although a simple trigger for fuel only will suffice, you mentioned you will be adding ignition control later so hence my suggestion to investigate a better trigger source now, that way you won't need to change that later when you do upgrade :-)
#43317
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
Per my response to dazq, those were my thoughts, too. But, I really don't know enough (yet) to have a credible opinion.
No problem, and as you venture forward you will find your own opinions of what will and can work, and what can't - for your project and goals. No better teacher than experience, and I assume you have already read the threads here on 1 and 2-cylinder "industrial" engines for ideas? Try a search for "cart" or "golfcart", such as this early thread. I talked to him more recently, and it's still running so well, he doesn't need the turbocharger I have for it. :lol: Speeduino now has many more options and refinements than back then, but it's still a simple example.
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
…I've also read that the injector "ON" time can be too short, that a lower pressure and longer pulse width can give better results. With the small fuel requirement of these engines, I was concerned that using automobile-type pressures and automobile-type injectors could result in problems.
That is a legitimate concern. Size your injectors well, and use standard pressures when possible. There are many more small injectors available today for econo cars, bikes and scooters, etc. That said, we have run injectors down to very low pressures with good performance for certain uses.
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
…On the face of it, it would seem that, with a "T," during the intake pull in one runner (low pressure), the MAP sensor wouldn't see that pressure reduction because the reduced pressure would just prompt flow through the "T" from the other runner.
You would think so, but for the air to flow from one runner to the other through the Tee fitting, the pressure must reduce. One method used for many years was a plenum Tee, effectively a fat Tee with the volume of your thumb inside. The plenum allowed the pressure to reduce across the volume before drawing on the opposite inlet, affecting the signal outlet, while also providing some averaging or signal smoothing. Don't go crazy big. Just some options.
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
… getting a little perspective on this particular project will help me put that information into context. For example, the wiki tells us what type of TPS to use, but I didn't see an option to do away with the TPS entirely (run a little rich, accept a little lag on acceleration. In this application, that has worked well for others when fitting EFI ).
Yes, there are many other options or workarounds that may be used or in combination. I do not suggest running outside of efficient Lambda for this. Instead use other tuning techniques to cover holes in AE, such as VE cell accel enrichment, which is helpful even when AE seems to be OK.
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
…I want things to be very simple: a sensor that isn't present can't fail. A software subroutine that isn't present can't cause a crash. I don't need variable valve timing, I don't want closed loop, etc. Just "good enough" open loop operation to make power at all RPMs (maybe based on RPM and MAP alone, perhaps a "START" switch to give 5 seconds of enrichment for starting, thus eliminating the need for an engine temp sensor?). Manual trim knobs to allow the user to adjust mixture to each side (by use of the EGT reading, or by RPM drop) would be good enough--I think!
A sensor that isn't present, can't give you information. ;) There is always the urge to simplify, and simplification is good to a limited extent. There is a cost to simplifying a simple system. Typically, you will find that tuning becomes far more difficult and results incrementally less effective when deleting sensors or basic functions. I would not suggest deleting any of the basics unless your system can not use it anyway.

Along those lines, Speeduino code is constructed to work as a whole, so if you don't use something, just ignore it. It won't mind, and is safer than hacking body parts out. Cold-start and warmup settings really are "dial-in" once you see how they work. Same for fueling, with specific cell or entire table changes with a click. We are back at the beginning, where your new hands-on experience will find what you need, and you don't yet know what to expect in operation, tuning, or techniques until you get there. Rock on!
Image
David
#43338
David,
Thanks a lot for your insightful comments. More on that below.
For context, this B&S engine is intended for aircraft use. Like any other application, the use affects the priorities of the project.
1) In-service reliability is priority one. The engine has to keep making power. Fuel economy, ease of use, engine longevity, etc are important, but secondary to reliability.
2) Manual operator adjustments in use are acceptable. Most pilots learn to fly on simple airplanes with primitive carburetors. They are familiar with techniques of manual mixture adjustment (using EGT or RPM, or by ear). Reliance on manual mixture control may not be the >best< approach for a modern EFI system, but I’m just saying that, unlike a daily driver or motorcycle, it is acceptable to me (and probably most users). And it has advantages: When CHTs are near the upper limit on climbout, it can be very handy to run a little rich to protect the exhaust valves, etc.
3) Fuel: It would be >convenient< to be able to burn leaded fuel. 100LL fuel is still the most widely available fuel for spark-ignition engines at US airports. Premium unleaded “pump gas” is becoming more widely available, but it isn’t as ubiquitous as 100LL. I’m willing to lug cans of gasoline to the airport to fill up my plane (we have only 100LL on site), but on a cross country flight it would be nice to be able to use 100LL when necessary. So, I’d like to run open loop, but can use a lambda sensor when setting up the fuel map.
PSIG wrote:
Wed Jun 03, 2020 2:51 am
Try a search for "cart" or "golfcart", such as this early thread.
Thanks, my keyword searches for applicable threads missed it. Yep, plenty of similarities to what I’m thinking of (air cooled, small engine), and some differences (uneven firing).
PSIG wrote:
Wed Jun 03, 2020 2:51 am
Size your injectors well, and use standard pressures when possible. There are many more small injectors available today for econo cars, bikes and scooters, etc. That said, we have run injectors down to very low pressures with good performance for certain uses.
That’s good to know. I’ll need to learn more about the various options for injector placement, too. Some folks flying behind these small engines use low-mounted twin carburetors with long induction runs to the top, which can lead to fuel condensing inside the walls of the relatively cool runners. If I use a motorcycle twin throttlebody, it would be easiest to use any existing injector bosses, MAP sensors, TPS, etc that are already built into the unit.
PSIG wrote:
Wed Jun 03, 2020 2:51 am
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
…On the face of it, it would seem that, with a "T," during the intake pull in one runner (low pressure), the MAP sensor wouldn't see that pressure reduction because the reduced pressure would just prompt flow through the "T" from the other runner.
You would think so, but for the air to flow from one runner to the other through the Tee fitting, the pressure must reduce. One method used for many years was a plenum Tee, effectively a fat Tee with the volume of your thumb inside. The plenum allowed the pressure to reduce across the volume before drawing on the opposite inlet, affecting the signal outlet, while also providing some averaging or signal smoothing. Don't go crazy big. Just some options.
Okay, a dumb question: If using a twin TB (so the butterfly will be open the same on both sides) and the engine RPM will obviously be the same for both cylinders, is there any reason to measure the MAP on both manifolds? Assuming the runner geometry is the same/virtually the same, won’t the MAP be the same? So, whatever the fuel requirements are for Cyl 1, we can just use the same injector pulsewidth for Cyl 2? If, during tuning, we find differences between Cyl 1 and Cyl 2, can Speeduino provide a separate fuel map for the two injectors, or some other trimming mechanism?
PSIG wrote:
Wed Jun 03, 2020 2:51 am
Vigilant1 wrote:
Tue Jun 02, 2020 2:36 pm
For example, the wiki tells us what type of TPS to use, but I didn't see an option to do away with the TPS entirely (run a little rich, accept a little lag on acceleration. In this application, that has worked well for others when fitting EFI ).
Yes, there are many other options or workarounds that may be used or in combination. I do not suggest running outside of efficient Lambda for this. Instead use other tuning techniques to cover holes in AE, such as VE cell accel enrichment, which is helpful even when AE seems to be OK.
Thanks. I haven’t learned enough to know if eliminating the TPS can be expected to improve >reliability< or not. It occurs to me that, in the case of loss of MAP information, it might be possible to keep the engine fed by reverting to an N-Alpha mode if a TPS is present. But, I don’t know if Speeduino allows user-selectable changes like that while in use.
PSIG wrote:
Wed Jun 03, 2020 2:51 am
There is always the urge to simplify, and simplification is good to a limited extent. There is a cost to simplifying a simple system. Typically, you will find that tuning becomes far more difficult and results incrementally less effective when deleting sensors or basic functions. I would not suggest deleting any of the basics unless your system can not use it anyway.

Along those lines, Speeduino code is constructed to work as a whole, so if you don't use something, just ignore it. It won't mind, and is safer than hacking body parts out.
Thanks, that’s good to know. To the degree I can, I’ll want to explore the way the code handles failed sensors (when they fail high, when they fail intermittently, when they provide no voltage at all, when they short to ground, etc). The OEM EFI software can be pretty opaque in this respect.

Thanks again!

Hardware: I don’t know if I should start with an NO2c board (this is a simple 2 cyl project), or if I’d be happier with a larger board. I’m concerned about the quality of the power the board will see. It’s not easy for me to tell if the various boards contain equivalent surge protection, filters on the sensor feeds, etc.
#43465
Vigilant1 wrote:
Wed Jun 03, 2020 9:32 pm
For context, this B&S engine is intended for aircraft use. Like any other application, the use affects the priorities of the project.
1) In-service reliability is priority one.

2) Manual operator adjustments in use are acceptable.

3) Fuel: It would be >convenient< to be able to burn leaded fuel.
Reliability; this will be based more on quality of components, assembly and installation, along with maintenance, than if it's EFI or what. Second, reliability of the engine's capability to run is directly impacted by how it is treated by the tune when run. The proper fueling, ignition timing, etc, will increase reliability of the mechanical functions, as partly has been demonstrated by modern vehicles. Finally, if you are not looking to use the advantages of the real-time data and self-adjustment EFI offers, then simplicity would be to use carburetors for equivalent efficiency.

Manual adjustment; consider the purpose of manual adjustment. If the system automatically adjusts to whatever you instruct, under the conditions you specify, then what is the purpose of manual adjustment? If you want full-rich when at full-power, then tune it for that. Unlike a carb with jets it will then do that for you, whenever at 100%, and still lean-out as you instruct it to do when throttled-back for cruise. That said, there are solutions for “knob” adjustments.

Fuel; 100LL is just fine, will result in a shorter sensor and spark plug lifespan than unleaded, and yet shorter if poorly tuned. I have burned lots of 100LL with WBO2 sensors. Considering a sensor is a consumable, and provides the self-adjustment that makes EFI attractive, I consume them. I look at it that I trash expensive spark plugs, oil and other stuff at a high rate in aircraft, and sensors are no different. At $50-$100 per-sensor, and a lifespan of perhaps 50-100 hours with 100LL, you're still only spending a fraction of the hourly fuel you'll burn or might save, or increased engine longevity at $0.50-$2/flight hour. That's me. Do whatever makes you giggle!
Vigilant1 wrote:
Wed Jun 03, 2020 9:32 pm
Okay, a dumb question: If using a twin TB (so the butterfly will be open the same on both sides) and the engine RPM will obviously be the same for both cylinders, is there any reason to measure the MAP on both manifolds?

If, during tuning, we find differences between Cyl 1 and Cyl 2, can Speeduino provide a separate fuel map for the two injectors, or some other trimming mechanism?
Yes, I would suggest both runners for MAP, simply because it gives double the signals to average each cycle, double the frequency (doubling updates) increasing accuracy for each of the two injections and cylinder firings.

Yes, with injectors on separate channels and a trigger signal providing crank position (sequential capability), individual cylinder fuel trim tables are optionally enabled.
Vigilant1 wrote:
Wed Jun 03, 2020 9:32 pm
I haven’t learned enough to know if eliminating the TPS can be expected to improve >reliability< or not. It occurs to me that, in the case of loss of MAP information, it might be possible to keep the engine fed by reverting to an N-Alpha mode if a TPS is present. But, I don’t know if Speeduino allows user-selectable changes like that while in use.
Rather than answer the TPS reliability question directly or whether to run dual ECMs for redundancy; I would suggest a step-back and re-think the use of experimental equipment, much of it not even designed to automotive standards, in your project. You either feel you can design and construct any part of your project to your purpose with enough inherent reliability, or you can't.

Air-frame, power-plant and control systems may not have to be medical or aviation-grade to be reliable enough for you in your application, but it is entirely your decision as to what is reliable enough, and how far down that experimental path you are willing to venture. Much of your project is a complete unknown, and home-made amateur-designed experimental EFI is only a small but important part of it. As my favorite flight instructor repeatedly stressed; “Loss of power is not an excuse to crash. Prepare, plan, practice and fly the plane with the knowledge that you may lose power at any moment.
Vigilant1 wrote:
Wed Jun 03, 2020 9:32 pm
… I’m concerned about the quality of the power the board will see. It’s not easy for me to tell if the various boards contain equivalent surge protection, filters on the sensor feeds, etc.
If you're concerned with electrical power, then improve it until you're not, or choose a different path if necessary. The various board designs each have schematics and component lists you can review. None of them are to any certification, grade or purpose. Then again, not much of your project is, by choice. They are all your choices. Certified components lose certification if used in non-certified applications. Catch-22. It's all part of Experimental-category aviation, and each of us must navigate it with our own judgment. 8-)

David
#43538
David,
Thanks for the reply. With an aircraft in the experimental category, it is often straightforward to tell if the construction is well done--or not. With a little more work, the key elements of the structure of the aircraft itself an be validated, or the record of other airplanes of the same design can be reviewed. Likewise with an engine. But (as you appreciate) it is quite a lift for a non-programmer, non EE to assess the suitability of any non-certified EFI system--hardware design, hardware installation, software failure analysis, etc. But if someone is going to try, it makes sense to ask about the relative power and I/O conditioning/filtering of the candidate boards, how the software handles errant sensor input, etc.
I wouldn't ask anyone to vouch for the suitability of Speeduino for aircraft use. Heck, even companies with a LOT of experience (and resources) have thrown in the towel on getting systems certified by the FAA.
Again, thanks.

Mark
Turbo dodge 2.5 minivan

Actually there is one critical issue. The stepper […]

Rb20det tps wiring

Have you referenced a factory service manual for p[…]

Thanks Dwight! yes, I have the speedy sensor grou[…]

UA4C: Fuelpump not working

...it is working now... What ended up being t[…]

Still can't find what you're looking for?