what about the os 3 needle heli carbs?
 

For some heli engines os has 3 needle carburetors
For example the 55hz "hyper" and there are also 91 and 105 "hz" versions.
Align had also there versions that are the same engines.

These have a dedicated midrange needle besides the high speed and idle setting.
Idle is adjusted with a eccentric i think its called screw that slightly rotates a jet inside the carb.

The lifespan of these engines (piston/liner sets) was not that long with some of the more aggressive 3d "pilots" so it might be able to find some of these carbs from trashed engines.
Those might work for some sizes of engines

As Dave said, you are completely free to experiment, and in itself, there is nothing wrong with that. It is an educational activity after all.

But please don't take this wrong... You are NOT "doing things differently"....
Virtually EVERYBODY that I have had the pleasure of helping, talking to, advising, but also LOTS of people that claimed to know "everything to know about converting engines" (without ever showing their personal results) the INVARIABLE suggestion of "I would use a Walbro" came along.
So basically, you are doing what everybody else is, or at least WOULD be doing if left to their own device...

The thing is: Walbro's WORK.... Above 15 cc they do. Below that, nope... Hit&miss affair, down to about 10 cc and below that, nope. Even OS Engines, after first using a Walbro-like carb on their 10 and 15 cc GT and GGT series, to the best of my knowledge now have cancelled the 10 cc gassers and made their own design carb for the 15 cc. I have worked on those engines, and was not impressed.
Try to find video's of home conversions using Walbro carbs on sub-10cc glow engines. You won't find them.

Everybody talks about it, some people try, nobody succeeds,

And that is what I was trying to get across.

Is it that the amount of fuel that has to be fed into the engine becomes so small that the carbs have to be made very accurate?
That ngh carb appears to be featuring some sort of membrane fuel pump driven by crankcase pressure.
I can feel play in the fit of the rotating barrel so my confidence in the precision of of the fuel metering part is "limited" ;)
"good" thing is that the engine itself looks to be a "bog standard" 2 stroke and even has the normal "pinch bar" style carb mounting hole in the crankcase.
So changing that carb with for exampe a super tigre should be possible .

os did try a "fuel injection" engine
 

Never saw one and likely really expensive. Also 31 cc so a lot bigger then what we are trying to do.
https://www.os-engines.co.jp/english...0sfi/index.htm

In a way i miss the days that os just would try crazy things like the wankel , and a 4 stroke with a form of turbo compressor.

Hate to say it, but... You guessed it, it won't work, and I know because I tried it.
The reason is simple: In its core, these 3-needle carbs are a twin needle carb with an additional fuel "channel" on them.
Read the manual and you will see that you start setting these up with the midrange needle closed, and only open that midrange, if midrange is too lean.

This means, that the combination "idle/main needle" already provides a fuel linearity that is "operable" for methanol, which means it is ALLREADY to rich for gasoline.
Since all you can do with the midrange needle is ADD fuel, you cannot lean out the midrange.

In general, dedicated helicopter carbs (for glow fuel) have a slightly better fuel linearity, and are marginally more suitable to use on aero-engines that are being operated on gasoline, but that difference is fairly marginal, and still way too rich in the midrange. That is with the midrange adjustment completely closed.

For anyone insisting on running glow carbs without modification, I stick to my advise: Try to find a Super Tigre carburettor. Not perfect, but your best option.

Yes... Just to give you an example: my .30 ASP fourstroke consumes about 3 ml of fuel per minute in flight. That is a horribly small amount of flow, and it needs to be metered with an accuracy better than 5% to be "runnable", and within 1% if you want to have a GOOD running engine.
At idle, the fuel opening between spraybar and idle needle, is a ringshaped opening with a width of about 0,02 mm....
And that is the problem. Testing at the factory, then calibrating manufacturing equipment to reproduce that kind of accuracy is prohibitive expensive.

Correct... Basically, what NGH did, was combine features of a glow carb (rotating barrel, twin needle) with those of a Walbro (pump membrane and fuel pressure regulator).
But they did so rather poorly, and the stupid thing is: Gasoline IS sensitive, but NOT THAT sensitive. At the "scale of things" where that engine is used, pumped and pressure regulated fuel supply is ABSOLUTELY NOT a necessity. I can prove that with Video documentary:


Both planes/engines have just muffler pressure and natural fuel draw,

Yup, if you can find one, that NGH most likely runs better with a SuperTigre carb.
MIND YOU: the manual says something like 20:1 or 25:1 for fuel/oil mixing ratio.
DON'T... Do NOT go leaner than 15:1, better 10:1. They look like decent engines, but they are not really...

Yup... the Wankel sure was a shame to go...

Fortunately, I have one, and you guessed it,that one TOO fell victim to my lust for modification.
"Those that know" said it would be impossible to convert one of them Wankels to gas, because of reasons of heat, as well as these engines "only can be lubricated by Castor Oil"....

Well, this one is running gas, and lubricated with the same synthetic 2-stroke oil I use in all my other engines...

I ran my OS 60 FP with a OS 46 fx carb. I had better results than with the 60 FP carb. I don’t know if it was because of the 2 needle carb or I’m getting better at tuning it. I was able to get 10500 rpm’s from a 13x6 prop which was better than the old carb. I struggled to get an idle below 2300. I then retarded the spark and got a better idle. I retarded it more and got an idle around 2000. My top end was only 8500. I advanced the spark a little and was able to get 9500 top end and an idle near 2000. I’m not sure it would idle at that all day but it seemed pretty consistent except when an air bubble went through. I noticed that when you change the timing you need to adjust the high end needle and as you have mentioned one or two clicks makes a big difference. This carb will lend towards Chris’s air bleed device. That will throw in another variable when it comes to tuning. At least I got a reasonable idle. I will need to work with the Evolution 61 and try retarding the spark to see if I can obtain a reasonable idle. I wonder if going to a smaller carb would be even better? I’ll have to see if I have a suitable smaller carb to try.

In theory the higher air velocity of a smaller throat should produce a better low pressure zone in the carb. Generally this makes for easier tuning with the trade off being less performance.

Xanaphyst is correct in that a smaller carb makes it easier to find decent adjustments.

Ideally,going back and forth between main needle and ignition timing, finding the best RPM, you should end up at near spot on 28 deg BTDC. Could be 27 or 29, but about that range.
So best is to just use a protractor and set timing at 28 degrees and be done with it.That is,when using an RcExl ignition with built in advance/retarding curve.

I seem to remember, yours does not have that. In that case, it is a bit of a compromise. You might want to consider getting one, it really improves low RPM behaviour.

But keep in mind, too much retardation, causes an increase in exhaust gas temperatures.
In fourstrokes, that can lead to the exhaust valve getting too hot, which not only is detrimental for life expectancy of that valve, but it can also cause the valve to start acting as a glowplug.

In twostrokes, increase in exhaustgas temperature can cause piston fretting at the topside of the exhaust port due to the overheated gasses burning off the oil film locally.

I would at all times pick top-end performance over idle, provided idle is low enough to bring the plane in for landing.

I originally set up both engines at 28 degrees BTDC. But it is easy to be off a degree or two. I’m not sure where it is now but that is easy to check. I may add an esp32 chip to control timing in my system. My brother is really good at that and will help me try it. I think I could make it as simple as anything below 3000 rpm’s and the timing goes to 10 degrees BTDC and anything above 3000 it goes to 28 degrees.

I don’t think I lost any high end performance with the .46 carb and it did seem easier to tune. I have some 30 size carbs that I might try. It would be interesting to see what happens.

The available CDIs have a built in curve and at low speeds is set to 28btdc. If you're implementing your own curve and not either adjusting for the cdi's built in curve or bypassing it by changing the cdi's schematic, you're probably doing it wrong.

just keep in mind the Rcexl ignition already takes care of all that. In my opinion the esp32 is a huge overkill in both required microcontroller performance, system complexity, and power resources.

If you really desire a homebrew ignition a simple low power PIC 12F683 will get the job done.

The CDI that I use is from Stirlingkit. It’s light weight and only $20. It doesn’t have a timing curve. I have used it for converting weed wacker motors and in planes with those motors and never had a problem with their reliability.

I also have three Saito CDI systems that are two wire systems. They use an inducer to signal the CDI not a hall sensor. I’m trying to sell them on EBay and nobody wants them so I took a crack at converting the unit to a hall sensor. It didn’t work but that was probably because I don’t know what I’m doing. I was following AIs instructions and will need to try again.

I appreciate the suggestion to use something other than the esp32. My brother is familiar with that and should be able to build something that can retard the spark at certain rpm’s.

I tried an OS 25FX carb on my OS 60. It’s surprisingly easy to swap these carbs and not have air leaks. My tank setup needs correcting. I was getting a lot of air in the line today making it difficult to tune and come to conclusions. Plus my tachometer died on me so I only got some initial readings. I was able to get 8500 or maybe 9000 for the top end. Idle was again a problem. I found that the engine ran better with a spark closer to 28 degrees BTDC. But I wasn’t able to really let the engine run in at different rpms. After I get my tank setup fixed I will try it again. Needle settings are just a few clicks here and there before you get too lean or too rich.

could you post images of the saito cdi?
 

Could you post close ups of the cdi and the sensors ? Depending on the parts used it might be possible to guess what signal the cdii expects to fire.

https://cimg3.ibsrv.net/gimg/www.rcu...d300af142.jpeg
https://cimg4.ibsrv.net/gimg/www.rcu...1fdc8e704.jpeg
https://cimg5.ibsrv.net/gimg/www.rcu...162b0773b.jpeg

Here are the photos. This is a two wire trigger mechanism. I understand it to be an induction type trigger. Hall sensors are three wire triggers.

My brother is working on an ignition advance circuit for my Stirlingkit CDI. It is actually a retard system because you program a delay from the top advance degree you start with. We will be starting with 28 degrees BYDC. He is going to start with a linear curve probably from 6 degrees at 1,000 rpm up to 28 degrees at about 7,000 rpm. It will stay at 28 degrees for any rpm above 7,000. He is going to have it connected to your phone so you can see the rpm’s and the timing number and he’s going to make it so you can
adjust the timing also by using your phone. I know this is like reinventing the wheel but it should be different because of your ability to change the timing curve.

On another note is do you think using a larger prop would help with getting a stable idle? I think I remember that converting to gas enables you to lug the engines without causing a problem.

If you can tune it, a heavier prop will idle slower because it has more inertia than a lighter prop.bigger props are heavier so they have more inertia. You need to balance that with top speed because you certainly need your plane to have enough power at full throttle, but it's always a balance.

Do you have a multi meter?
 

DO you have (or can borrow) a multi meter?
Use the continuity test(beeper sometimes combined with diode test) to do some checks.
The metal spark plug cap, the braiding and the ring at the end of the green wire all should be connected with each other.
So the meter should show beep when the lead "probes" are touching any combination of those 3 but the most important is from the cap to the ring.
Then next is to confirm the pickup probe ; the 2 contacts should also show continuity and if they do also do a resistance check.
If you can bribe a assistant then another test is possible.
One person holds the lead probes on the 2 contacts on the connector and another moves a magnet quickly past the sensor.
With the multi meter set to its lowest voltage setting (direct current is best) then it should show a voltage the moment the magnet passes the sensor.
That would confirm it is in fact a coil and that the coil is working.

Connecting a 3 wire hall sensor to this system would not work because a hall sensor needs power to work.
This is usually the red wire and that is missing here.

Once conformed it are coils and they are working AND the cap, braiding and ring make contact you could try to make a spark-plug spark outside of a engine.
Before you power the cdi measure if the outside of the plug, cap ect make contact and then power the cdi and move a magnet past the sensor to see if that triggers the cdi to fire,
That way you know at least that the system is working as designed,
After it should be possible to design a "in between" circuit that takes a hall sensor signal and then generate a voltage like a coil would do.

Also these saito sensors look like the os sensors that are designed to be mounted in the backplate and use the passing crankpin as signal.

Thanks. I may try a larger prop.

I have tested the Saito CDI and it generates a spark. I did try a circuit with a hall sensor and transistor but it didn’t work. I may have messed up my hall sensor but I have lots of them. I’m going to try again and do a better job of setting up the proposed hall sensor circuit. Thanks for the suggestions.

To convert a 3-wire Hall sensor to work with a 2-wire inductive (Variable Reluctance/VR) input on a CDI, you must transform a digital square wave into an analog-style pulse. This is significantly more complex than the reverse because most 2-wire CDIs (like older Saito units) expect an AC signal that crosses zero volts to trigger the ignition.
Conversion Options
1. Professional Signal Converter (Recommended)
The most reliable method is to use a dedicated signal converter designed for motorsport. These units are built to take a Hall effect signal and output a simulated inductive pulse that your CDI can understand.

• KMS Hall & Inductive Converter: This device from Kronenburg Management Systems (KMS) is specifically designed to convert Hall sensor signals into inductive signals for ECUs and ignition systems.
• Speedway / Motorsport Converters: Look for "Hall-to-VR" or "Digital-to-Analog tachometer converters" from specialized retailers like Summit Racing.

2. The Capacitor Coupling Method (DIY)
If you want to build a simple interface, you can try "AC coupling" the Hall signal. This tricks the CDI by turning the Hall sensor's 0-5V square wave into a brief positive and negative spike.
Components needed:

• 10k \(\Omega \) Resistor: Pull-up for the Hall sensor.
• 0.1µF to 1µF Capacitor (Non-polarized): Blocks DC and passes only the "edges" of the signal.

Wiring:

1. Hall Sensor Power: Connect Red to +5V and Black to Ground.
2. Pull-up: Connect the 10k \(\Omega \) resistor between +5V and the Hall Signal wire (White/Yellow).
3. The Converter: Connect one leg of the Capacitor to the Hall Signal wire.
4. To CDI:
   • Connect the other leg of the capacitor to one of the CDI's trigger inputs.
   • Connect the other CDI trigger input to Common Ground.

3. Important Considerations

• Signal Strength: Some 2-wire CDIs require a high-voltage pulse (several volts) to trigger. A 5V Hall signal through a capacitor might be too weak. If it doesn't fire, you may need a small transformer or an op-amp circuit to boost the spike.
• Timing Shift: Converting signals can introduce a tiny delay or shift the trigger point (rising edge vs. falling edge). Always re-check your ignition timing with a strobe light after making this conversion.
• Polarity: If the engine doesn't start, swap the two wires going into the CDI trigger input. Inductive inputs are sensitive to the direction of

This is a AI suggestion. It is not the same one I got earlier. I need to investigate this further.

Wouldn’t you just use a unipolar hall sensor so it’s expecting a single magnetic pole for the high / low ??

I’m not having much luck getting my Saito 72 dialed in on gas with an external air valve. I can’t seem to get any consistency at all.


https://cimg4.ibsrv.net/gimg/www.rcu...33625c09d.jpeg
https://cimg6.ibsrv.net/gimg/www.rcu...b982f635b.jpeg