For any discussion not specifically related to your project
#69768
I know this is going to be an issue. Just asking if anyone has worked around it before. I've been reading the forums and researching for days. I've spent a several hours in the code and still trying to get my head around it.

So I will be starting the V-twin engine by hand due to weight limitations. More specifically by flipping a propeller by hand. That means I need to get through intake to TDC without the speeduino losing it's crank position. Here are my assumptions from what I have figured so far.

1) The crank position calculation can not tell if the crank reverses any amount, it will count that as a forward pulse.

2) A slow down or stop could be interpreted as a missing tooth.

Minus any solutions you all have the only things I can think of are to trigger an injector pulse with the MAP sensor pressure drop, and add a separate crank sensor with one tooth that triggers just past TDC. My thinking is that should run well enough to get the crank angle figured out. No idea how much coding that would take yet.

Or maybe putting the missing tooth right before TDC as flipping the prop through compression how it's done. This only works if the system can sync up that fast. Again, not understanding the code well enough yet.
#69772
Only ignition is the issue (complicated by odd-fire), as the power-on fuel prime will get it to fire and spin enough for sync and run. This is used by other pilots I know using old code and distributor trigger. With missing-tooth it works for bikes and foot-cranking as the stroke is long enough to see enough of the pattern to sync and fire, but some still have to pre-position for the kick.

It can sync quickly, but is an issue needing a couple teeth to read as "normal" before a missing tooth, so it knows when it sees it. An alternative is Dual Wheel that needs no extra crank teeth to know where it is - but does require pre-positioning with a short rotation. Another is a self-triggering system that sparks on trigger for minimum timing, but allows more advance from the ECM, such as the GM HEI and Ford TFI distributor systems.

I might look into using one of those modules for firing only one of the coils as a bump-start. In other words, it trigger fires when it sees a single tooth at 10°BTC Cyl1, and fires it on the priming fuel. Hopefully that's enough to spin the crank sufficiently to sync and fire the second coil, continuing the process normally. See a short look into it here.
#69776
PSIG wrote: Sun Oct 06, 2024 9:15 pm Only ignition is the issue (complicated by odd-fire), as the power-on fuel prime will get it to fire and spin enough for sync and run. This is used by other pilots I know using old code and distributor trigger. With missing-tooth it works for bikes and foot-cranking as the stroke is long enough to see enough of the pattern to sync and fire, but some still have to pre-position for the kick.

It can sync quickly, but is an issue needing a couple teeth to read as "normal" before a missing tooth, so it knows when it sees it. An alternative is Dual Wheel that needs no extra crank teeth to know where it is - but does require pre-positioning with a short rotation. Another is a self-triggering system that sparks on trigger for minimum timing, but allows more advance from the ECM, such as the GM HEI and Ford TFI distributor systems.

I might look into using one of those modules for firing only one of the coils as a bump-start. In other words, it trigger fires when it sees a single tooth at 10°BTC Cyl1, and fires it on the priming fuel. Hopefully that's enough to spin the crank sufficiently to sync and fire the second coil, continuing the process normally. See a short look into it here.
I had figured it would always start it on one of the cylinders as the prop will not be in the correct place for the other cylinder. I never thought to search for a kick starter, that is a similar situation.

So if I can squirt fuel as part of a prime cycle can I simply turn the engine over a few times with the ignition disabled, bring the prop to the start of the compression stroke, then pull it through TDC and have the missing tooth on the crank at TDC? I should be able to pass quite a few teeth in one shot and then the missing tooth would resolve the crank position.

Seems like firing just after TDC would bring get the engine spinning enough to sync up. Firing before TDC with my hand on the prop is a bit scary to me.

I'm not sure if I'm going to run synchronous or just wasted spark yet. So a cam signal is still up in the air. There is lot's I'm looking to do, but not asking any of those questions until I get me head around things better.

I read that map sensors can give problems. Is that an issue with good automotive rated ones? Or just the $.1 ones from amazon?
Thanks again.
#69825
clytle374 wrote: Mon Oct 07, 2024 3:23 am So if I can squirt fuel as part of a prime cycle can I simply turn the engine over a few times with the ignition disabled, bring the prop to the start of the compression stroke, then pull it through TDC and have the missing tooth on the crank at TDC? I should be able to pass quite a few teeth in one shot and then the missing tooth would resolve the crank position.
That's the concept, and again, the Missing Tooth decoder requires some teeth ahead of the missing tooth to judge the missing tooth (and not that it's a slow standard gap). As it recognizes the missing tooth only as a gap that takes 1.5x the previous gaps, even a low battery can sometimes confuse missing tooth detection. This is one reason I'm suggesting a cycle sensor (cam speed), as it is a distinct signal that can't be confused with an uneven stroke through multiple teeth.

clytle374 wrote: Mon Oct 07, 2024 3:23 amSeems like firing just after TDC would bring get the engine spinning enough to sync up. Firing before TDC with my hand on the prop is a bit scary to me.
Indeed, ATC would be preferred to softly kick it in teh right direction. This was used on a Subaru-powered Kitfox for hand-propping from the open cockpit door, reaching forward. In this setup, the prop was positioned just after the compression of the target cylinder, and starting was by just flipping the prop tip backwards against compression enough to trigger the ATC spark. Bang. It was kicked back the other direction, and running in the correct direction under ECM timing control. Easy, safe, very consistent and reliable.

In order to use a setup similar to that, the spark was not sync'ed, but was direct fire on seeing a trigger. This is how the ignition modules I mentioned operate, firing on any signal at the physically set timing advance (retard). Once running and sync'ed, the ECM would advance timing to 'normal'.

clytle374 wrote: Mon Oct 07, 2024 3:23 amI'm not sure if I'm going to run synchronous or just wasted spark yet. So a cam signal is still up in the air. There is lot's I'm looking to do, but not asking any of those questions until I get me head around things better.
Odd-fire cannot run true waste-spark. Spark must be sequential or distributor. This is still possible using only a crank sensor, with two coils (one each cylinder), firing the waste spark only into the same cylinder instead of a sister cylinder. Not splitting hairs here for terms - there is a difference.

clytle374 wrote: Mon Oct 07, 2024 3:23 amI read that map sensors can give problems. Is that an issue with good automotive rated ones? Or just the $.1 ones from amazon?
I have had no issues with automotive MAP sensors. If in-doubt, use only OEM quality sensors, or the ones often used here (NXP/Freescale, Honeywell, etc). For inexpensive testing and experiments, sensors from collision wrecks at the wrecking yard are always good in my experience (they were running when wrecked ;)), and come with connectors for easy hookup. My local wrecker charges $5 for any sensor with connector. YMMV
#69830
PSIG wrote: Thu Oct 10, 2024 4:23 am As it recognizes the missing tooth only as a gap that takes 1.5x the previous gaps, even a low battery can sometimes confuse missing tooth detection. This is one reason I'm suggesting a cycle sensor (cam speed), as it is a distinct signal that can't be confused with an uneven stroke through multiple teeth.
This was my concern and question. 1.5x gives me a time frame, verses just knowing I'd have a problem at some point.

PSIG wrote: Thu Oct 10, 2024 4:23 amIn this setup, the prop was positioned just after the compression of the target cylinder, and starting was by just flipping the prop tip backwards against compression enough to trigger the ATC spark. Bang. It was kicked back the other direction, and running in the correct direction under ECM timing control. Easy, safe, very consistent and reliable.
I used to start RC 4 strokes the same way. They are technically diesels, so just getting to the right compression did it.

PSIG wrote: Thu Oct 10, 2024 4:23 amOdd-fire cannot run true waste-spark. Spark must be sequential or distributor. This is still possible using only a crank sensor, with two coils (one each cylinder), firing the waste spark only into the same cylinder instead of a sister cylinder. Not splitting hairs here for terms - there is a difference.
Gotcha. Yes firing each revolution was what I meant by wasted spark. Lack of terminology on my part.

PSIG wrote: Thu Oct 10, 2024 4:23 amI have had no issues with automotive MAP sensors. If in-doubt, use only OEM quality sensors, or the ones often used here (NXP/Freescale, Honeywell, etc). For inexpensive testing and experiments, sensors from collision wrecks at the wrecking yard are always good in my experience (they were running when wrecked ;)), and come with connectors for easy hookup. My local wrecker charges $5 for any sensor with connector. YMMV
Another issue on my part here. My brain may be missing a tooth. I meant HALL sensors, not MAP sensors. I've found some through digikey that have large temperature tolerances, ect. They would be a lot easier to mount than the hundreds of pictures I've seen from googling them to find one that would work.

I have read that 2:1 reduction drives are bad for propellers. But I can't find anything to back that up. If I could use an even 2:1 with a toothed belt and can used a sensor on the prop hub for the CAM signal. I'm looking into using the MAP as a cam sensor too. The response time of the MAP is .001 seconds, and I'm running separate intakes, so it should be a usable signal if properly filtered. And port close to the head, short hose, ect.

Thanks
#69845
clytle374 wrote: Thu Oct 10, 2024 6:16 amI meant HALL sensors, not MAP sensors. I've found some through digikey that have large temperature tolerances, ect. They would be a lot easier to mount than the hundreds of pictures I've seen from googling them to find one that would work.
Look for Hall sensors with automotive rating, for the proper type and back-bias magnet (to read ferrous targets) built-in. ;) For one example, I have used the little Allegro ATS617 successfully, while others have used similar versions well, and they can fit into a small bracketed or threaded nylon hose fitting or 3D printed 'sensor' body, to a mounted piece of PCB, or whatever gets it in position for a custom setup.

clytle374 wrote: Thu Oct 10, 2024 6:16 amI have read that 2:1 reduction drives are bad for propellers. But I can't find anything to back that up.
Yeah, they are known to have harmonic vibration issues. I've flown one, and the two effects were a wah..wah.. hum, not too bad but could become quite irritating. The other was a fairly short drive belt life on that one. The setup looked good, aligned and true at correct tension, but it ate belts every 500 hours or so, IIRC. Clean break with no signs before failure. Gear drives can sometimes be no better (and often worse for clattering at low rpm), although I ran a 2.40 2S and a 3.00 4S without issues. Then again, a 2.04 2S geared unit was fine. :?

clytle374 wrote: Thu Oct 10, 2024 6:16 amIf I could use an even 2:1 with a toothed belt and can used a sensor on the prop hub for the CAM signal. I'm looking into using the MAP as a cam sensor too.
True, but get creative with other sources of cam-speed signals. Examine the whole engine for everything that moves at half-speed, such as valves, pushrods, rockers, timing gears, etc. Anything with a groove, slot, motion, or blob of weld metal or steel rivet could be a signal source. I tuned on one with a junkyard Chrysler cam sensor epoxied through the valve cover just above a rocker arm for 1 signal per-cycle. Another with a small groove cut into the distributor/oil pump drive shaft. Find your sneaky signal location, and good luck!
#69848
PSIG wrote: Fri Oct 11, 2024 10:09 pm Look for Hall sensors with automotive rating, for the proper type and back-bias magnet (to read ferrous targets) built-in. ;) For one example, I have used the little Allegro ATS617 successfully, while others have used similar versions well, and they can fit into a small bracketed or threaded nylon hose fitting or 3D printed 'sensor' body, to a mounted piece of PCB, or whatever gets it in position for a custom setup.
Thanks for the part number, added it to my digikey list. I had one from infineon, but it was programmable. Got any fuel pump part number suggestions for 45-ish HP?

PSIG wrote: Fri Oct 11, 2024 10:09 pm Yeah, they are known to have harmonic vibration issues. I've flown one, and the two effects were a wah..wah.. hum, not too bad but could become quite irritating. The other was a fairly short drive belt life on that one. The setup looked good, aligned and true at correct tension, but it ate belts every 500 hours or so, IIRC. Clean break with no signs before failure. Gear drives can sometimes be no better (and often worse for clattering at low rpm), although I ran a 2.40 2S and a 3.00 4S without issues. Then again, a 2.04 2S geared unit was fine. :?
Okay, I'll eliminate the 2:1 option. I'd have to guess the clean break is the same argument against 2:1 gearboxes, it puts the load on the same tooth or set of teeth every power pulse.

PSIG wrote: Fri Oct 11, 2024 10:09 pmTrue, but get creative with other sources of cam-speed signals. Examine the whole engine for everything that moves at half-speed, such as valves, pushrods, rockers, timing gears, etc. Anything with a groove, slot, motion, or blob of weld metal or steel rivet could be a signal source. I tuned on one with a junkyard Chrysler cam sensor epoxied through the valve cover just above a rocker arm for 1 signal per-cycle. Another with a small groove cut into the distributor/oil pump drive shaft. Find your sneaky signal location, and good luck!
The intake valve was my next choice for the cam signal after the p̶r̶o̶p̶-̶s̶h̶a̶f̶t̶ and then MAP. I'm staying in US part 103 ultralight rules, so every ounce is a problem. Looking at the math I believe the MAP sensor can work to my intended 5000 rpm max. I'm still trying to learn the code well enough to see if I can devise anything clever in the code to make it work throughout the RPM range. That and tuning the intake runners might bite me there too. I should have an engine in November, so lots of guessing until then, a scope on sensor with a real running engine is the make or break there. I did add it to a fork of Ardu-Stim to simulate a theoretical MAP pulse from an intake event.
Screenshot_20241011_231528.png
Screenshot_20241011_231528.png (77.95 KiB) Viewed 772 times

I've got lot's of plans and ideas for this, I'll share as I go, don't want to reveal just how crazy I might be this early on. One is I plan to use 2 speeduinos. Both having their own sensors, with the injector and spark signals crossing over through the other speeduino and a set of normally closed relay contacts, using a switch like a left/right mag switch to energize the relay to take control of the other's spark and injection. So both are programmed to run the other cylinder, but aren't by default. Dual plugs are an option, but the heads are very expensive, and the coils are not light either. Looks like the LS1 coils are a good choice for a smart coil to keep the relays small.

Thank you for all the advice.
#69853
clytle374 wrote: Sat Oct 12, 2024 3:33 amLooking at the math I believe the MAP sensor can work to my intended 5000 rpm max. I'm still trying to learn the code well enough to see if I can devise anything clever in the code to make it work throughout the RPM range.
Your simulated MAP signal just needs a signal conditioner to create a square wave. Make your own with an op-amp or something, or a module such as the DSC should do for that. Goal is a clean low-high signal for clear timing point.

Second, Speeduino can either sync once and run on crank from there, or sync every cycle. As the MAP will be nearly useless at high throttle (tiny signal), I'm hoping the first one will do, or your valve or other sensor.

clytle374 wrote: Sat Oct 12, 2024 3:33 amI've got lot's of plans and ideas for this, I'll share as I go, don't want to reveal just how crazy I might be this early on.
Too late. :lol:

clytle374 wrote: Sat Oct 12, 2024 3:33 amOne is I plan to use 2 speeduinos.
What is your intended goal for 2 Speeduinos?

45hp only needs about 15 LPH, and I'm not up on latest tiny pumps, but cheap and easy would be a generic in-tank motorcycle pump, commonly advertised as "35 LPH" or similar. A tee to a common in-tank regulator such as a generic Toyota or similar type gets fixed 3-bar pressure set for simple and cheap. Scooter pumps and injectors are another option.
Toy_reg_side.jpg
Toy_reg_side.jpg (48.68 KiB) Viewed 716 times
#69854
PSIG wrote: Sat Oct 12, 2024 9:55 pm Your simulated MAP signal just needs a signal conditioner to create a square wave. Make your own with an op-amp or something, or a module such as the DSC should do for that. Goal is a clean low-high signal for clear timing point.

Second, Speeduino can either sync once and run on crank from there, or sync every cycle. As the MAP will be nearly useless at high throttle (tiny signal), I'm hoping the first one will do, or your valve or other sensor.
My thinking is to condition the signal in code. Best part, is no part. I realize that the signal will change drastically with throttle position. There is a setting for MAP as as cycle minimum, so I'm not sure what kind of sampling rate can be achieved, and what can be reliably extracted from the signal, how much CPU it would use, ect.

From what I've seen it looks like it will fall back if it looses CAM sync okay, having issues trying to sort all the details of that out. I saw a video of a guy using it on a motorcycle and if he failed the cam sensor it was still running. I don't like the idea of never being able to regain sync if it is lost, but I sometimes wonder if there is enough gain to even worry about the cam sensor. I'm probably missing proper terminology again. And a scope of the sensor of a running engine is the only true way to get an answer.
PSIG wrote: Sat Oct 12, 2024 9:55 pm What is your intended goal for 2 Speeduinos?
Redundancy. I live in the mountains. I can lose one of anything and worst case(coil, plug, injector) is single cylinder operation, it should fly on one cylinder. Fly, not climb..

PSIG wrote: Sat Oct 12, 2024 9:55 pm Too late. :lol:

Well, I guess since we have that out of the way. :lol: I'll explain the rest, I feel I'm competent with mechanical and electrical systems, I work in automation.

So I can swing a longer low pitch prop for climb performance, mountains again. But it will leave me spinning high RPM and guzzling my 5 gallons the whole time in cruse. IVO makes a flight adjustable prop. But if I set the prop to high pitch at high RPM I'm going to break the part 103 speed rules. FAA says it can be limited by other means as long it is not accessible in flight. So that leaves me with few options that I consider safe.

Firstly I need to be able to read the prop pitch. I called IVO and they said you can install stops and the motor will pull about about 8A when stalled. So I can have a high/low pitch that way. Or I can modify it in someway to read back the pitch, I have a few ideas, but need to get one in my hands first.

Secondly I now need to be able to limit the RPM of the engine when the prop is at high pitch. A rev limiter seems dumb and dangerous. The 2 throttle butterflys will share a common shaft with a TPS on each end for each controller. So I could use a servo stop, but then either failure could close my throttle. Or I could TBW so either controller could open the throttle against a common return spring. I saw this morning that the project is 100% against TBW. So I guess I'm on my own there.

More than one way to implement all of the above. I'm planning a year for the engine... I should probably find an instructor and my old log book and go the experimental/sport pilot option.
Thanks.

PS: if you want crazy how about writing an decoder to use to 2 overlapping HALL sensors on the crank wheel so the position is never lost even if the crank stops? Looking into it.
#69868
Re-posted this reply since I went back and realized that I have messed up the quoting and it made no sense.

PSIG wrote: Sat Oct 12, 2024 9:55 pm Your simulated MAP signal just needs a signal conditioner to create a square wave. Make your own with an op-amp or something, or a module such as the DSC should do for that. Goal is a clean low-high signal for clear timing point.

Second, Speeduino can either sync once and run on crank from there, or sync every cycle. As the MAP will be nearly useless at high throttle (tiny signal), I'm hoping the first one will do, or your valve or other sensor.
My thinking is to condition the signal in code. Best part, is no part. I realize that the signal will change drastically with throttle position. There is a setting for MAP as as cycle minimum, so I'm not sure what kind of sampling rate can be achieved, and what can be reliably extracted from the signal, how much CPU it would use, ect.

From what I've seen it looks like it will fall back if it looses CAM sync and still run, having issues trying to sort all the details of that out. I saw a video of a guy using it on a motorcycle and if he failed the cam sensor it was still running. I don't like the idea of never being able to regain sync if it is lost, but I sometimes wonder if there is enough gain to even worry about the cam sensor. I'm probably missing proper terminology again. And a scope of the sensor of a running engine is the only true way to get an answer.

PSIG wrote: Sat Oct 12, 2024 9:55 pm What is your intended goal for 2 Speeduinos?
Redundancy. I live in the mountains. I can lose one of anything and worst case(coil, plug, injector) is single cylinder operation, it should fly on one cylinder. Fly, not climb..

PSIG wrote: Sat Oct 12, 2024 9:55 pm Too late. :lol:

Well, I guess since we have that out of the way. :lol: I'll explain the rest, I feel I'm competent with mechanical and electrical systems, I work in automation.
So I can swing a longer low pitch prop for climb performance, mountains again. But it will leave me spinning high RPM and guzzling my 5 gallons the whole time in cruse. IVO makes a flight adjustable prop. But if I set the prop to high pitch at high RPM I'm going to break the part 103 speed rules. FAA says it can be limited by other means as long it is not accessible in flight. So that leaves me with few options that I consider safe.

Firstly I need to be able to read the prop pitch. I called IVO and they said you can install stops and the motor will pull about about 8A when stalled. So I can have a high/low pitch that way. Or I can modify it in someway to read back the pitch, I have a few ideas, but need to get one in my hands first.

Secondly I now need to be able to limit the RPM of the engine when the prop is at high pitch. A rev limiter seems dumb and dangerous. The 2 throttle butterflys will share a common shaft with a TPS on each end for each controller. So I could use a servo stop, but then either failure could close my throttle. Or I could TBW so either controller could open the throttle against a common return spring. I saw this morning that the project is 100% against TBW. So I guess I'm on my own there.

More than one way to implement all of the above. I'm planning a year for the engine... I should probably find an instructor and my old log book and go the experimental/sport pilot option.
Thanks.

PS: if you want crazy how about writing an decoder to use to 2 overlapping HALL sensors on the crank wheel so the position is never lost even if the crank stops? Looking into it.
speeduino.com/fw

The link to the Speeduino .ini file for RealDash s[…]

Volvo B20E

I think I may have found the issue now and I suspe[…]

Ktm 990 turbo

:ugeek:    Impressive. Would it be pos[…]

VQ35DE Build - V0.4X board

This is cool. I am definitely following along. I h[…]

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