Glowgeek

An excellent essay/consolidation for tuning gassers Bert. All my Saito FG gassers are tuned that way, on the ground and in flight.

I'm hoping to find the "ugly sister" today by reducing or eliminating the fuel channel slowdown. If it works, a long standing issue will be resolved with the click of a button.

1967brutus

Let's hope and pray... It should make a difference no doubt.

And remember, the "ugly" sister still is cute by all standards, and better than having no gal at all...

John_M_

Why did you slow down the fuel channel in the first place?... that would equate to opening the throttle ( introducing more air ), and then leaning out the mixture... I can understand the throttle servo slow down, but as more or less air is introduced by opening / closing the throttle , the fuel mixture needs to respond to the throttle... and fuel mixture ( air / fuel ratio ) with the fuel channel curve.

The Furtaba 18sz will allow multiple channel mixes.... the throttle stick can be setup as just a " stick position" reference only... and then the "first slave" channel to be the channel that controls the throttle servo... and the "second slave" channel to be the channel that controls the fuel mix curve ... you can either slave the fuel mix channel directly to the "stick position" channel, or to the "first slave" throttle channel... so if you put a slow down on the "first slave" throttle channel, and the "second slave" fuel mix channel is slaved to it, they will follow each other, "second slave" fuel mix channel following the "first slave" throttle channel...... or you can slave the "second slave" fuel mix channel directly to the "stick position" channel, so the two slaved channels ( throttle & fuel ) will follow the 'stick position" channel... so the slow down only affects the "first slave" throttle channel... and the "second slave" fuel mix channel follows the "stick position" channel... ... also if you put a slow down on the "stick position" channel, it effects the two slave channels ( throttle & fuel )... I found it interesting that the Futaba radio could be setup this way... but doing so you have to forfeit a channel output for the "stick position" channel

John_M_

First ever Walbro / STHIL Hybrid carb... Started coring out the hi/lo needles in the walbro WT-452... needs to go deeper, and then I need to plug and re-drill the fuel feed channels from the metering chamber into the solenoid chamber... and then select a suitable orifice / plug for the lo speed fuel circuit in the bottom of the solenoid chamber, haven't figured out how I'm going to go about that just yet.

https://ibb.co/9H7b6vp

Glowgeek

All good, ready to mount and fly as soon as I figure out why it's down on power. Unrelated to the project here.

Thanks to everyone who has carried me along through this project. Especially Bert and Dave of course.

A special thanks HAS to go out Chris, who beat the "dead horse" back to life. Thanks Chris for updating me on the settings for fuel channel speed.

The Saito 180 I started this project with, what?, 10 months ago?, is mounted on the test stand and ready to fire up the next time I can get away from life.

Happy banana dancing,

Lonnie

Cat 1

My Feedback: (1)

SOOOO good to hear... That just goes to prove how much control this system has - The engines just don't like having to wait for fuel even if its just a short delay.

This also proves how hard it is to diagnose things from a distance sometimes - Would have been nice to be 2000 miles closer together to troubleshoot this one.

1967brutus

Basically, because initially I simply did not know any better. The throttle delay was easy enough to figure out, because that was allready discovered when I was still grinding carbs. And given that the LS needle assembly moves with the throttle, I simply assumed that when fuel is controlled separately, the fuel channel should move in sync with the throttle channel. It turned out not to be true.

There were not exactly instruction manuals included with the solenoids...

Glowgeek

I'm really surprised at MYSELF in not recognizing that the solution to my particular problem was to increase the fuel channel speed beyond the ".2 seconds faster" limit I had read about early on. I didn't recognize what caused the cough/die problem until later in the game, actually not until I slaved the fuel channel to the stick a couple of weeks ago. I thought I saw a bit of improvement increasing the fuel channel speed by that .2 seconds but didn't put 2 and 2 together entirely.

I've monitored the RCG thread as well as this one every day, getting notifications and reading every single post, or so I thought. I must have missed something along the way and as a result tortured some of you with a relentless barrage of questions. For that, I apologize.

1967brutus

don't... Because those questions at the very least kept all our brains collectively engaged. No telling what ideas originated based on your questions.

Glowgeek

Had half an hour to sneak away while the wife was napping. I left at 5:00 and typing now at 5:23. Started the 180, curve flattened, dialed in and it's ready to mount and fly in less than 30 min. And it freak'n rips from 1500 to 9k in .7 sec, no problems.

More happy banana dancing!!!!!!

John_M_

Right, but even with the older multi point EFi systems, when opening the throttle quickly, the ECu will double up on the injector pulse to enrich the fuel mixture to improve throttle response, that and the fuel rail pressure increases with the drop in manifold pressure.

So when you first setup the mixture curve... are you guys moving the throttle stick to a point in the fuel mix curve... and then adjusting that mixture point in the curve to where the engine runs reasonably well at that throttle position & mixture curve position... and then move the throttle stick to the next point, set that curve point to where the engine runs reasonably well, and then just keep moving through each throttle position / curve point adjusting each point... and then fine tune those points as needed once you get it running well enough to fly the model... Futaba defauts with 9 mix points, with the addition of up to 8 more, total of 17 points, each point mixed to a throttle stick position.

Cat 1

My Feedback: (1)

Thats my basic process John... My Graupner only has 7 point curve but there is a "curve" function that smooths the curve and the points can be moved in both axis making the curve infinitely adjustable. I tend to have a knob set during testing and setup so I can play and find where the happy spot is - It's a bit harder to tell the sweet spot on gas than glow and this what I struggled most with at the start because the reaction is a bit different. Once I have the curve established I set the knob to be a 5% either way trim and start flying - this way I can safely fly and play with the mixture.

Cat 1

My Feedback: (1)

didn't describe the "knob" that well - It's mixed to the curve channel and when in the centre you are "on the curve" - During initial testing I leave it at "full throw" to be able to drive the mixture where needed to find the curve point locations - After that I cut the "travel" back to 5% each was so it acts like a trim to narrow down on the sweet spot and I won't risk a shutdown if I get "Ham Handed".

Cat 1

My Feedback: (1)

Dave - Once I get the system figured out I will get with you as you can probably greatly improve my "code" and I think it would be very doable to include it into the Mixture controller as its fairly "light" on processing needs - I'm using the opto to clean up the Ignition output a bit - it seems to have some "spurious" signal content that plays bad with the Arduino input- not sure if its emanating for the HV side of the Box but its hard to see on a scope. It plays havoc with many forms of "tach code". The opto seemed to help. - And the opto also allows for the "higher voltage" boxes with less issues.

So far here is what I have - Simple measurement of RPM from the ignition Tach out.. a bit of math and it outputs an "analogue" PWM to drive a small mosfet to turn the little pump. The math contains a "exponential curve" function so manipulating one variable changes the "curve" and another sets Max RPM (where the pump is at 100% duty cycle) . Arduino and Pump are powered off a 5V receiver Channel and I'm planning on taking the output pulse of that channel to have a "prime" function where you can Momentarily pump the tank because real planes don't have chokes - Just primers

1967brutus

guessing that you now will not send us a vid of the engine anymore.... Send us a vid of you doing that banana dance....

Great to hear, Lonnie! I am happy for you. And 1,5 to 9k IN 0,7 sec, thaat sounds like absolutely normal values!

John_M_

The Futaba shifts the curve on both axis as well... also has the smooth function, and certainly does affect the PWM signal on the oscilloscope... its a fixed algorithm to smooth between points... I will probably thin the number of points down to what is necessary... I'll have to take a look and see if I can program a dial knob to trim the mixture, add it as a conditional mix possibly, its not as flexible as OpenTX, and may only shift the entire curve... would be nice to trim the active point, with the ability to "set and save".


So my plan is to get the engine running on the STHIL carb... get an established throttle / mixture setup, and then replace the STIHL with the modified walbro carb, and see how the two related to each other... STIHL's carb is nothing more than a regular diaphragm carb, with the solenoid valve in place of the two needles, nothing real fancy happening there... but the STIHL carb is half as much larger in size over the walbro... just too much buggering around to make it fit under the cowl without having to rework everything forward of the firewall... it kinda ticked me off, three sixteenth of an inch minimum, but the cascading effects of moving everything forward just upset the entire appearance of the model... I rather spend the time modifying the walbro, than the AC.

1967brutus

Yup... that is the general idea. Of course, first you need to find "rich but runable" needle settngs. Important step, don't leave that out. Either disconnect the driver module from the receiver and set the HS needle rich, or set a horizontal fuel curve at 80% (or +60) and tune the HS needle like you normally would
It takes a few times going stepwise up and down the throttle range and to this day I am still not sure why the first go is not final, but it isn't.
I usually need between 5 and 10 times up and down the range before the fuel curve is there. Tune the LS needle as needed, going by what shape of fuel curve you end up with.
The slowdown can be set after the curve is done, but since it is going to be somewhere between 0,4 and 1 sec anyway, I usually start at 0,6~0,7 right from the beginning. Not that it makes any difference in the finding of the fuel curve.

EDIT: Leave out that part with the needles, it does not apply when setting up a Stihl or modified Walbro, I guess...

1967brutus

That sure sounds like you have a fairly workable system WRT tuning with two tunable parameters. I am curious how it will work out.

Raleighcopter

That could be easily added to the controller: increase flow to solenoid on increasing mixture during throttle moves. I could even add a channel so the "electronic acceleration pump" can be adjusted from the transmitter. This would allow coupling the mixture to the throttle channel output so that the mixture is always synchronized with the throttle position while also adding better control over acceleration. It would mean you would first tune the mixture curve for static throttle and then go back and add throttle slowdown and additional fuel for acceleration

What does everyone think?

John_M_

Actually, I have plans to put back the needles in the hi/lo speed fuel circuits, so I can adjust the base mixture... so I only plan to sink the solenoid down deep enough to engage the o-ring seals... that will raise the floor of the solenoid chamber enough so I can drill and tap from the side of the hi/lo speed circuits, and install two tapered set screws... the STIHL carb has fixed orifices at the bottom of the solenoid chamber, and they came up with the base mixture orifice sizes for the hi/lo fuel circuits, and programing to control the mixture, BUT their firmware incorporates some temperature feedback. so it has some adaptability base on tempuratue ... Once I figure out the base mixture settings, I can machine a version 2 and install fixed orifices at the bottom of the solenoid chamber, and deeper so the solenoid will fully insert down to the shoulder.

John_M_

You would only use that for sudden abrupt throttle changes, based on a given jump in the throttle stick channel pulse width... and then just shift the entire fuel curve enriched trim as Cat describes, but only for a few μs in time... and or rise the tank pressure if Cat's air pump system lends itself worthy.

1967brutus

All it would take is to de-energize the solenoid during throttle movements above a certain rate of change. that would cause a momentary "flooding" at quick throttle changes, which might even eliminate the need to lead the throttle with the fuel, possibly eliminate the need to minimize the V-shaped curve,, AND allow for shorter throttle slowdowns.
If I think about it, the accelerator pump on my old Ducati also is a rather crude contraption without any adjustability, and it works fairly well.

How to execute technically, I have no clue, because due to the unpredictability of the fuel curve shape, the trigger needs to be the throttle signal, not the fuel curve signal. So the controller would (IMHO) a 2nd imput from the throttle channel, for no other reason than to detecht changes in throttle setting.
It might also close off the fuel at rapid decelleration, that would make the engine follow the throttle downaward a bit quicker, but I do not think that will result in any user-noticable improvement.
I also do not yet have an idea of the needed adjustable parameters (I can take a wild guess but I would need a bit more "braintime" to think it over" ). But it should not be all that critical, I think.

It might be as simple as setting a fuel curve, NOT using a throttle slow down, and have an adjustable rate-of-rise. adjust the rate of rise such that the engine accepts rapid throttle movements. It might need a "dwell time" (time the solenoid remains fully open after the throttle has stopped moving)
It might work, it might not. It would take a prototype controller to confirm.

It is, at the least, an interesting possible feature. Probably a can of worms to get set-up, but interesting.

Raleighcopter

Here's my thoughts so far....

The throttle slowdown serves to keep the signal to the throttle servo synchronized with the servo position. It prevents the signal from changing faster than the servo can move. I think it would need to be equal to the servo sweep time provided by the servo manufacturer.

The mixture gets slaved to the throttle channel output. This synchronizes the mixture to the throttle position and separates the fuel needed for acceleration from the fuel needed for a steady state operation at a given rpm.

We add fuel to that base mixture fuel amount and we do it based on throttle speed and only on increasing throttle. How much to add is a guess at this point but a single additional channel could be used to define the acceleration fuel amount as well as reading the throttle position.

I'm just spitballing at this point but it seems doable and it sounds like it would be more easily tuned.

1967brutus

I can to a point follow the rest of your post, but are absolutely confused by this... I for now fail to see the need to sync the signal rate of change with the servo's maximum speed of movement in the context of the rest of the post....

Raleighcopter

I probably should have said the minimum throttle slowdown should be the servo speed rating.

I want to account for the separate components that all add up to the total instantaneous fuel requirement. The obvious first component is the fuel required at steady state (constant rpm). I want my base mixture curve to corrspond to this and I want the throttle position to exactly track with this base mixture curve.

Next, I want to add the fuel required for acceleration to that base mixture only when accelerating. I do not want the base mixture to include the acceleration fuel via having the mixture lead the throttle because I won't be able to model that in software. It also negates the requirement of a mixture curve with a positive slope.

By the way, with a broken airbleed screw, there's not much that can be done to change the mixture curve.