What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Stupid fuel injection question

That would make sense. With there being a "vent" port on the regulator I can see where it would be sensing atmospheric pressures and not MAP. So, I guess it's back to the drawing board with regards to my plumbing plan. LOL!!

This next question may be aimed a little more at Ross and the Borla regulator he uses. In the Borla reference material it specifically says to install the regulator on the downstream/return side of the fuel rail. So, the question is, is it acceptable to have everything (fuel supply from pumps, fuel rail distribution lines, and fuel pressure regulator) all plumbed into a single fuel log/distribution block?

Mark

We plumb the reg on the downstream side of fuel block which is either rear baffle or engine spine mounted.



This photo shows the 6 cylinder spine mount block. Fuel pump connects to one end fitting, regulator to the other end.

We use the Borla regulators exclusively as they have a dual Viton diaphragm and are made in the US. Too many other brands are built offshore these days (even some famous US brands) with inferior materials and design to save money. This is something which demands best quality and not a place to save money in our view.
 
This next question may be aimed a little more at Ross and the Borla regulator he uses. In the Borla reference material it specifically says to install the regulator on the downstream/return side of the fuel rail. So, the question is, is it acceptable to have everything (fuel supply from pumps, fuel rail distribution lines, and fuel pressure regulator) all plumbed into a single fuel log/distribution block?

Mark

Sure, that would work fine. The reason Borla says to put it on the return side of a fuel rail loop is so that cool fuel is always circulating through the system - once the fuel goes through the loop and past the injector lines, it goes back to the tank. You don't ever have a situation where fuel is sitting stagnant in the block/rail/lines due to low fuel flow (like idle or a rapid power-off descent) and picking up heat.

If you have the fuel pump and immediately downstream have the Borla with its return line, and then a long line to go up to the engine and lines to the injectors, you can still have a situation where low fuel flow allows that fuel to get hot before going into the engine - it will still work, just less desirable.
 
Last edited:
It appears that in reality that the pressure in that fuel log/distribution block would remain constant regardless as to which port the regulator was tied into it...just as long as the return fuel is plumbed into the return side of the regulator. In other words, you wouldn?t want the regulator supplying fuel to the log/distribution block; only returning fuel needs to go through the regulator. But it looks like once fuel hits that distribution block that it wouldn?t matter which actual physical port is used for which purpose because the pressure should be constant within that block.....or am I missing something (again) LOL!!

Mark
 
Sure, that would work fine. The reason Borla says to put it on the return side of a fuel rail loop is so that cool fuel is always circulating through the system - once the fuel goes through the loop and past the injector lines, it goes back to the tank. You don't ever have a situation where fuel is sitting stagnant in the block/rail/lines due to low fuel flow (like idle or a rapid power-off descent) and picking up heat.

If you have the fuel pump and immediately downstream have the Borla with its return line, and then a long line to go up to the engine and lines to the injectors, you can still have a situation where low fuel flow allows that fuel to get hot before going into the engine - it will still work, just less desirable.

Roger that...thanks Greg...and Ross for the reply?s!!

Mark
 
Hi Mark,

If you're using Ross' system you need to plumb it to his specs.

Having said that, I can tell you that what you described earlier is often called a 'dead head' system with pump(s)>regulator(with bypass back to tanks)>distribution block>injectors. Some variation of this can be found on many cars over the last few decades, and some alternative engine guys that are running automotive style injection use this arrangement. Tracy Crook of Real World Solutions fame has been flying that architecture in his Mazda 20B powered RV-8 for many years and hundreds of hours.

Charlie
 
So I guess my new plumbing plan will be to mount a fuel distribution block on the firewall and basically have everything going into it (supply line, fuel rail distribution lines, and regulator). At that point come out of the regulator with the return line going back to the tanks. As noted in a previous post I'd still like to have a single dead-headed fuel rail on each side of the engine rather than individual lines going to each injector and also avoid a true "flow-through" circulating communication rail. I know that a circulating rail is optimal but in trying to simplify the plumbing and keep things neat and clean, I'll like to have two separate dead-headed rails that are being fed from that fuel distribution block.

Mark
 
Hi Mark,

If you're using Ross' system you need to plumb it to his specs.

Having said that, I can tell you that what you described earlier is often called a 'dead head' system with pump(s)>regulator(with bypass back to tanks)>distribution block>injectors. Some variation of this can be found on many cars over the last few decades, and some alternative engine guys that are running automotive style injection use this arrangement. Tracy Crook of Real World Solutions fame has been flying that architecture in his Mazda 20B powered RV-8 for many years and hundreds of hours.

Charlie

Thanks for the information Charlie!! Yes, that’s exactly what I was describing and hoping that I could use. I think that I may be better off in the long run following the standard EFI plumbing architecture. If I were using a throttle body where fuel was being injected into the intake manifold there rather than individual injector solenoids where constant pressures are more critical, than I’d probably stick to my plan...but I think that I’ll probably stick with more of the proven method.

Mark
 
Last edited:
One thing I should have mentioned is that the system I described still has a manifold pressure line to control the regulator. The regulator is on the cold side of the firewall, and a single fuel line goes through the firewall to the distribution block, with a manifold pressure line from the engine through the firewall to the control port on the regulator.

FWIW, I'd consider what you're now describing as least desirable to me, because it means 4 separate flexible lines crossing from firewall to the engine. With the distribution block on the engine (regardless of regulator location), only a single flexible supply line is needed, and possibly (again, depending on regulator location/plumbing) a flex return line.

Charlie
(again, if using Ross' system, follow his installation instructions...)
 
And again, keep in mind that with loop-style fuel rail, hot-starts are a non-issue since the heat-soaked fuel and vapor goes back through the regulator and into the tanks as soon as you turn on the fuel pump - within a very few seconds you've got cool fuel at the entrance to the injector lines, which minimizes how much hot fuel you can deal with to only a few seconds of runtime.

A dead-head system may still have hot-start symptoms to some degree.
 
You can mount the reg anywhere downstream of the pump but as has been discussed before, in the event you run a tank dry you'd like the reg placed right after the block or rail to minimize the volume of air which has to be processed through the injectors before it starts running again. This is a main reason we don't like to see the reg inside the cabin near the pump and why we use -3 lines to feed the injectors.

Injectors on a rail is better in this regard however you have other issues with increased surface area to gather heat if mounted below the hot cylinders and close by the exhaust pipes and the difficulty in fitting them on the top of a Lycoming.

There are advantages and disadvantages to almost any mechanical design.

The dead head layout and components we use has been proven over many years in temperatures up to 118F ambient. There haven't been any issues with hot start or vapor lock using either the thread-in top mounts or the weld-in bosses on the intake tubes. The top mount injectors run much cooler under flight conditions obviously and even after hot shutdown as they're not in close proximity of some 800F exhaust pipes. We know this from instrumentation, not guesses. Thanks to Dave Anders for running back to back tests on both types of mounts.

We feel for best results, people using our systems should follow our recommendations and use our supplied hardware. It has more flight time on it than anything else flying EFI. Trying unproven ideas may lead to tears in some cases.
 
One thing I should have mentioned is that the system I described still has a manifold pressure line to control the regulator. The regulator is on the cold side of the firewall, and a single fuel line goes through the firewall to the distribution block, with a manifold pressure line from the engine through the firewall to the control port on the regulator.

FWIW, I'd consider what you're now describing as least desirable to me, because it means 4 separate flexible lines crossing from firewall to the engine. With the distribution block on the engine (regardless of regulator location), only a single flexible supply line is needed, and possibly (again, depending on regulator location/plumbing) a flex return line.

Charlie
(again, if using Ross' system, follow his installation instructions...)


Hey Charlie, picture this layout. One main fuel line exiting the cockpit and tied into the firewall mounted distribution block. The distribution block will have four items attached to it. 1. One fuel inlet line 2. One fuel outlet/rail line (no spaghetti). 3. The fuel pressure regulator. 4. The fuel pressure sensor. The one fuel outlet line will then go forward toward the engine. Once in a happy location that rail feed line will have either a "T" or a "Y" installed. There will be one individual fuel rail on each side of the engine that ties its perspective injectors together (not individual lines going to each injector). Each rail will deadhead at its front injector and then tie into the previously mentioned "T" or "Y" at the back. So basically, the plan is to have a very simple plumbing layout with only a single feed line coming from the distribution block which then splits off to feed a single fuel rail on each side of the engine. The fuel block itself is where everything is happening...(inlet, outlet, mixing, pressure regulation/return, and pressure monitoring)....but the plumbing itself will hopefully be a single line that only divides once to feed the two rails. Your thoughts?

Mark
 
And again, keep in mind that with loop-style fuel rail, hot-starts are a non-issue since the heat-soaked fuel and vapor goes back through the regulator and into the tanks as soon as you turn on the fuel pump - within a very few seconds you've got cool fuel at the entrance to the injector lines, which minimizes how much hot fuel you can deal with to only a few seconds of runtime.

A dead-head system may still have hot-start symptoms to some degree.

Hey Greg,

I agree, a flow-through type of plumbing system will provide that constant fresh flow of fuel to the injectors at all times but as Ross mentioned, as that fuel is making its journey around the loop it has the potential to pick up heat. With the amount of fuel that's flowing and the speed at which it's flowing, it's doubtful that it really has enough time to pick up enough heat to really make any difference at all. However, with a fuel distribution block which feeds individual fuel lines or rails, the fuel will always be cold at the distribution block and hopefully the fuel won't have time to heat up enough to make a difference as it feeds the injectors. Does anyone know how much pressure is required to keep the fuel from boiling and turning into vapor? I think in these EFI systems with the fuel pumps always being fed by cold fuel, it's probably highly unlikely that these EFI systems can suffer from a typical vapor lock scenario as can be experienced by a standard aircraft fuel injection or carburetor system due to the fact that the injectors always have that constant high pressure fuel being delivered to them. I think that keeping that fuel in a liquid state as it feeds the injectors is more important than the actual temperature of the fuel itself....(or am I missing something here)? So, regardless if it's a deadhead or flow-through type of plumbing system, I think we're still much better off with an EFI system than a standard aircraft fuel system when considering vapor lock issues. Your thoughts?

Mark
 
Hey Charlie, picture this layout. One main fuel line exiting the cockpit and tied into the firewall mounted distribution block. The distribution block will have four items attached to it. 1. One fuel inlet line 2. One fuel outlet/rail line (no spaghetti). 3. The fuel pressure regulator. 4. The fuel pressure sensor. The one fuel outlet line will then go forward toward the engine. Once in a happy location that rail feed line will have either a "T" or a "Y" installed. There will be one individual fuel rail on each side of the engine that ties its perspective injectors together (not individual lines going to each injector). Each rail will deadhead at its front injector and then tie into the previously mentioned "T" or "Y" at the back. So basically, the plan is to have a very simple plumbing layout with only a single feed line coming from the distribution block which then splits off to feed a single fuel rail on each side of the engine. The fuel block itself is where everything is happening...(inlet, outlet, mixing, pressure regulation/return, and pressure monitoring)....but the plumbing itself will hopefully be a single line that only divides once to feed the two rails. Your thoughts?

Mark

I think we have a 'failure to communicate'. :)

If you're not splitting the fuel line until it gets to the engine by my definition I'd say that the distribution block is on the engine.

If I were using Ross' system, I'd probably just follow his recommended layout. It's likely to be as simple (or simpler) than what you're describing, and if I had any issues, he can more easily help troubleshoot.

I don't see any practical difference between what you're describing, and having that 'distribution block' immediately after the fuel pumps. EXCEPT: more of the fuel is exposed to the heat of the engine compartment. If heating the 40 psi fuel is a concern, then I would either keep as much as possible out of the engine compartment, or use the flow-through system Ross describes.

To Airguy's recent point about hot starts with a deadhead system: There seems to be conflicting info out there. It's never an issue with modern cars, where virtually all have deadhead systems. It's not been an issue with a/c installations I've seen. But IIRC, Ross has said they've had issues in their testing if a tank is run dry.
 
Once the fuel leaves the pump under pressure, there is almost no likelihood of vapor lock due to the 40-50 psi it's under. Additionally, the return line purges everything up to the -3 feed lines which are stainless and don't transfer heat very well.

We've had no reports of vapor lock with our EFI in aircraft but people do need to pay attention to the inlet plumbing to the pumps as we've discussed before- preferably no 90 degree fittings, smooth bends and mount the pumps on the floor in an RV. Our biggest concern is vapor lock at the pump inlet on a hot day, high altitude and on mogas.
 
Once the fuel leaves the pump under pressure, there is almost no likelihood of vapor lock due to the 40-50 psi it's under. Additionally, the return line purges everything up to the -3 feed lines which are stainless and don't transfer heat very well.

We've had no reports of vapor lock with our EFI in aircraft but people do need to pay attention to the inlet plumbing to the pumps as we've discussed before- preferably no 90 degree fittings, smooth bends and mount the pumps on the floor in an RV. Our biggest concern is vapor lock at the pump inlet on a hot day, high altitude and on mogas.

Yes, my pumps will be mounted on the floorboard just forward of the fuel selector. The main fuel line will turn up and run up the firewall (on the cockpit side) and then exit the firewall up as high as possible. This location is where I'll have the fuel distribution block and fuel pressure regulator mounted as previously described. (You know, the part where Charlie and I are failing to communicate). LOL!! Then I'll continue on to the actual engine/injector plumbing from there. So, even if running MOGAS and with the pumps mounted on the floorboard on the cockpit side of the firewall, and even with a deadhead type of plumbing system, hopefully I won't have any vapor lock issues due to the continuous 40-50 psi fuel line pressures.

Mark
 
Three thoughts favoring mounting the manifold on either the top of the engine or back of the baffles-
- although it would look neat to have injector lines running forward from the firewall to each injector, I think of the 'clutter factor' of all those lines behind the engine possibly causing restricted maintenance access later.
- the shorter the length of the individual injector runners, the better I like it.
- the fewer connections between the 'solid' airframe and the 'shaky' engine, the better I like it, even if it means 2 -6 braided lines run discretely between the firewall & top of the engine, positioned to not block access.
 
Three thoughts favoring mounting the manifold on either the top of the engine or back of the baffles-
- although it would look neat to have injector lines running forward from the firewall to each injector, I think of the 'clutter factor' of all those lines behind the engine possibly causing restricted maintenance access later.
- the shorter the length of the individual injector runners, the better I like it.
- the fewer connections between the 'solid' airframe and the 'shaky' engine, the better I like it, even if it means 2 -6 braided lines run discretely between the firewall & top of the engine, positioned to not block access.

Well said Ralph, My system started out early on with a -6 line snaking its way all around my engine to feed the injectors. All the fittings and complexity got a little ridiculous, and I switched to Ross's nice looking, simple fuel manifold on the firewall.

I later looked at all the fittings and connections mounted to the engine as potential failure points. In terms of vibration, I also like the idea of only the injector boss and one small light fitting being exposed to engine vibrations.

A totally uneducated guess by me would say that the theoretical benefits of having a "loop" style system instead of "deadhead" are minimal. I think there are several good successful examples including the Chevy LS and honda K series.
 
Once the fuel leaves the pump under pressure, there is almost no likelihood of vapor lock due to the 40-50 psi it's under. Additionally, the return line purges everything up to the -3 feed lines which are stainless and don't transfer heat very well.

Ross - what has been your experience with a heatsoak where the gas in the injector lines has flashed to vapor, and then you hit the fuel pumps for starting, pressurizing those lines to the 40-ish psi? Do the vapor bubbles collapse into liquid (mostly or completely enough that it makes no difference), or is the quantity simply small enough in that -3 line that the result gets lost in the stumble of the engine starting in the first three revolutions? The injectors (and engine) really don't care what temperature the fuel is - as long as it's liquid.

Or does the pressure in those lines NEVER drop low enough to allow the fuel to boil during shutdown? Or at least slowly enough that it doesn't bleed down before the engine is cool?
 
Ross - what has been your experience with a heatsoak where the gas in the injector lines has flashed to vapor, and then you hit the fuel pumps for starting, pressurizing those lines to the 40-ish psi? Do the vapor bubbles collapse into liquid (mostly or completely enough that it makes no difference), or is the quantity simply small enough in that -3 line that the result gets lost in the stumble of the engine starting in the first three revolutions? The injectors (and engine) really don't care what temperature the fuel is - as long as it's liquid.

Or does the pressure in those lines NEVER drop low enough to allow the fuel to boil during shutdown? Or at least slowly enough that it doesn't bleed down before the engine is cool?

It's not hot enough here in Calgary (max 38C) for us to have first hand experience with this so we rely on guys like Dave Anders and other users in hot climates in the Southern US, Australia etc. Nobody has ever reported any hot start issues with the dead head design. Dave purposely let the engine heat soak about 5 minutes before trying a hot start at 115F+. He said a few blades and it started and idled smoothly right away.

I think you could still get it in the lines or injector with mogas as the boiling point is lower than 100LL.

The regulator should maintain pressure in the system for hours or days. If not, anything that's vapor should become liquid again as soon as the pump starts and builds pressure to the usual 45-50 psi (about 1 second usually).

As you surmise, the -3 lines only have a few ccs of volume and anything that was boiling will be purged out in a few seconds of cranking and idle.

From the Dave Anders SDS EFI thread-

"I love this system, you can full power or lean with no problems and it starts like a car, even on hot starts"

"Thermocouples on the injector housing showed a max temp of 174F after 10 minutes with the heads at about 280F on shutdown so the injectors don't get anywhere near as hot as the heads"

"I don’t see vapor locking ever being a problem, because when I extremely heat soaked everything on the ground to 360 to 370 CHTs (the hottest I’ve ever seen) with an oil temp of 232 and the #4 injector reaching 324 after shut down, I let it sit for perhaps 4-5 mins and it started perfectly after 1 1/2-2 blades and idled smooth at 600 rpm. The fuel manifold seems to hold about 140-135 degrees."
 
Last edited:
Southeast U.S.

Ross - what has been your experience with a heatsoak where the gas in the injector lines has flashed to vapor, and then you hit the fuel pumps for starting, pressurizing those lines to the 40-ish psi? Do the vapor bubbles collapse into liquid (mostly or completely enough that it makes no difference), or is the quantity simply small enough in that -3 line that the result gets lost in the stumble of the engine starting in the first three revolutions? The injectors (and engine) really don't care what temperature the fuel is - as long as it's liquid.

Or does the pressure in those lines NEVER drop low enough to allow the fuel to boil during shutdown? Or at least slowly enough that it doesn't bleed down before the engine is cool?

I elected to wait for Ross since your question was to him, specifically.

I will add to his response to say I have had zero hot start issues in Tennessee. I turn on the fuel pump, wait for the pressure to build then push the start button. It just starts, no coughing no waiting for the air to clear out of the lines. My fuel block is located on the rear baffle with four SS lines going to the injectors.
14mgl5u.jpg
 
I elected to wait for Ross since your question was to him, specifically.

I will add to his response to say I have had zero hot start issues in Tennessee. I turn on the fuel pump, wait for the pressure to build then push the start button. It just starts, no coughing no waiting for the air to clear out of the lines. My fuel block is located on the rear baffle with four SS lines going to the injectors.
14mgl5u.jpg

Very clean installation!!
 
Since this has turned into a discussion about fuel rail design, here's what I'm doing. The rail is just a loop of .035 SS hardline with 6 bungs welded on. There is hardline from the rear baffles down to the fuel pump blockoff plate; supply and return flex hose from there to firewall. Regulator is on the cold side of the firewall.

15gozz5.jpg


Was going for "light", but in reality its probably pretty close to the fuel block Ross offers.

I'm not advocating the approach by any means, just showing an alternative view.
 
Last edited:
Very nice work Michael...I wish I trusted my welding skills enough to try something like that!! :)

Mark
 
Since this has turned into a discussion about fuel rail design, here's what I'm doing. The rail is just a loop of .035 SS hardline with 6 bungs welded on. There is hardline from the rear baffles down to the fuel pump blockoff plate; supply and return flex hose from there to firewall. Regulator is on the cold side of the firewall.

15gozz5.jpg


Was going for "light", but in reality its probably pretty close to the fuel block Ross offers.

I'm not advocating the approach by any means, just showing an alternative view.
I'm curious about benefits of this design vs putting a Tee at each injector top and running the fuel line around in a loop.
 
I'm curious about benefits of this design vs putting a Tee at each injector top and running the fuel line around in a loop.

Short answer: My preference to minimize MMOI. I'm trying to minimize mass hanging off the injector body. Originally I intended to have an "in" and "out" fitting on these bodies just so I could flush them with cool fuel given my hot climate. As it turns out, many others flew this setup before I did and had no issues with fuel boiling even in long individual lines. So I downshifted into plan B and got as close to a true rail system as I could without hanging the mass on the bodies.

I see now that there is ample evidence Ross has made a good choice with the high strength stainless steel injector bases he supplies, but as an early adopter I was conservative.
 
Short answer: My preference to minimize MMOI. I'm trying to minimize mass hanging off the injector body. So I downshifted into plan B and got as close to a true rail system as I could without hanging the mass on the bodies.

I see now that there is ample evidence Ross has made a good choice with the high strength stainless steel injector bases he supplies, but as an early adopter I was conservative.

While on this topic, we had design evaluations concerning the strength of the thread-in injector bases and minimizing the mass of the parts to reduce stresses on the bases. Clearly aluminum was a poor choice here from a strength and fatigue standpoint. After a lot of consideration and searching, I went with a high strength steel alloy here which is 4 times stronger and has 7 times the fatigue resistance at the temperatures involved, of aluminum parts. Having these parts break and spray fuel all over a hot engine wouldn't make for a good day. We also use a narrow pattern injector so the spray does not impinge on the bore of the base as well.
 
Last edited:
injector mounts

We have only seen two instances where injector bodies were stressed to a failure. This is with close to 500 installations in the field over 12 years of operation.

In both cases the injectors were not supported and were hung on the side of the engine supporting their own weight and the attached fuel rails. Our installation manual clearly states to support the fuel rail with Adel clamps.

Any mechanic can tell you that mechanical injector systems (and yes SDS injector systems) need to be supported as well, or lines could/will eventually crack.

Be safe, use common aircraft installation methods (no free hanging parts on engines). No one wants fuel spraying inside an engine cowl.

Robert
EFII
 
While on this topic, we had design evaluations concerning the strength of the thread-in injector bases and minimizing the mass of the parts to reduce stresses on the bases. Clearly aluminum was a poor choice here from a strength and fatigue standpoint. After a lot of consideration and searching, I went with a high strength steel alloy here which is 2+ times stronger and has 4+ times the fatigue resistance at the temperatures involved, of the aluminum parts used by one competitor. Having these parts break and spray fuel all over a hot engine wouldn't make for a good day. We also use a narrow pattern injector so the spray does not impinge on the bore of the base.
what is the thinnest wall thickness on these housings?

Full disclaimer: I play around with my own stuff for my own use, so I'm curious.
 
15gozz5.jpg



Just an observation? that sure looks like a lot of unsupported line to be as important as it is. Sure looks like a huge failure potential.
 
We have only seen two instances where injector bodies were stressed to a failure. This is with close to 500 installations in the field over 12 years of operation...

In this case we are discussing the direct mount bosses that thread into the primer/injector ports, not the weld in types.
 
...Just an observation— that sure looks like a lot of unsupported line to be as important as it is. Sure looks like a huge failure potential.


I appreciate that observation but in this case the picture isn't worth 1000 words. So here's a few more: The clamp spacing on the main rail is consistent with typical engineering guidance. The one exception being the run that crosses over the engine. Just out of frame at the split is a clamp which stretches the unsupported run a touch longer than generally accepted. It's a watch item.
 
what is the thinnest wall thickness on these housings?

Full disclaimer: I play around with my own stuff for my own use, so I'm curious.

Approx .080 which is why you don't want to use aluminum alloys here, not to mention the heat transfer.
 
We have only seen two instances where injector bodies were stressed to a failure. This is with close to 500 installations in the field over 12 years of operation.

Interesting. The number of sales doesn't tell us much; this is EAB, where projects drag out for years before completion. Two fails in how many hours?

In both cases the injectors were not supported and were hung on the side of the engine supporting their own weight and the attached fuel rails. Our installation manual clearly states to support the fuel rail with Adel clamps.

A quick tour of the customer installation photos on the EFII website says very few of the of the welded port installations got supporting adels. As a practical matter, the fuel rail hoses were simply too far away from any mounting point. Typical installation:

EFII%20Injector%20Rail.jpg


The new primer port mounts are much closer to the heads, so an adel can be installed:

System%2032%20Port%20Mount%20Injector.jpg


However, note that you're still using stiff hoses to link injectors installed in separate heads, and those heads vibrate independently. That's distinctly different when compared to injectors with independent branch feed lines from a central fuel block.

I also note that the primer port location means the injectors are largely horizontal, vs vertical for the SDS injector port location. The large amplitude torque reaction shake of a Lycoming would load the horizontal injector mounts much more highly than a vertical mount.
 
Here is a photo of our thread-in injector mounts for Lycoming Parallel Valve engines.



It's a 3 piece design minus the stainless cap screws holding it together. Left to right:

1. 6061-T6 top cap, threaded 1/8 NPT female for AN3 fuel fitting, bore seals to upper injector O-ring
2. 7075-T6 clamping flange
3. High strength stainless steel (proprietary) threaded 1/8 NPT male to screw into factory injector port, bore seals to lower injector O-ring

Socket cap screws are Locktited into into the flange

The chart below shows the relative engineering properties of the material we use for our base adapter vs. 6061-T6 aluminum:



You'll notice that our material has over 4 times the tensile, shear and yield strength, 7 times the fatigue strength and 1/10 of the thermal conductivity (we don't want heat flowing from the head to the injector).

These are highly loaded parts because of the vibration levels, so only the strongest materials should be used.

Additionally, we use a stainless (not plastic) body injector with a narrow spray pattern to insure that the fuel does not impinge on the ID of the adapter. Due to the small adapter bore, most traditional injector spray angles means the fuel will hit the bore before entering the port which is obviously not desirable.

Very careful thought and analysis goes into each components we design and this is especially so on critical components such as these. My 30 year background in machining, materials, welding and fabrication comes in handy here. Reliability has to be our #1 design criteria on aviation products.
 
Last edited:
Here is a photo of our thread-in injector mounts for Lycoming Parallel Valve engines.

It's a 3 piece design minus the stainless cap screws holding it together. Left to right:

Ross, any chance you could take one apart for us? It appears to be very different from the injector mounts in the EFII photo...a lot more than just a stainless steel nipple.
 
The new primer port mounts are much closer to the heads, so an adel can be installed:

System%2032%20Port%20Mount%20Injector.jpg


However, note that you're still using stiff hoses to link injectors installed in separate heads, and those heads vibrate independently...

Certainly noteworthy, but worse than "vibration" is the amount of independent movement each head has relative to each other. We know that installing a one piece baffle spanning 2 or 3 heads will crack in short order because the rigid baffle can't restrain the movement of the heads. Engineering practices specifically caution that hose assemblies need to have some slack in them so they don't become a "structural element" between two assemblies - In this case the hose is certainly a tension element, and thanks to the adel, it's a fairly effective compression element as well. When the heads move apart due to thermal or operating stresses, the hose is not going to stretch. The reaction is going to be bending through the 1/8NPT nipple. That's all from pictures - there may be another method to isolate the injectors from these loads but it's not apparent. Anyone have any more info on this?
 
injector mounting

The teflon hoses linking cylinders are not stiff hoses.
Hard line is a "stiff hose".
We have had no issues what-so-ever with properly supported installs. The referenced injector failure was after many years of operation with an inadequately supported install.

"Support your fuel rails and anything else bolted to your engine -properly" is the lesson here.

Robert
 
What's the PN of the hose stock (Aeroquip 303, etc) that you use?

How many examples of your direct mount injectors are flying, and what's the approximate cum hours?
 
Makes for an interesting comparison.

The EFII injector mount starts with a length of ordinary aluminum tubing. The ID is threaded at both ends, probably with an NPT pipe tap. Then a section is cut from the side to clear the injector's electrical connector.

The engine end of the tube can be screwed on a welded bung, as in this photo...

EFII%20Injector%20Mount%20Body.jpg


....or on a machined adapter for the 1/8" NPT primer port. Apparently the material is 6061.

EFII%20Port%20Mount%20Injectors.jpg


A fuel feed fitting is screwed into the outboard end of the tube. As you see above, it can be a tee, or something similar to an AN914.

The inboard injector o-ring seals against the inner bore of the welded bung or port adapter. The outboard o-ring seals against the ID of the screwed-in fitting.

Given the cutaway section, the outboard end of the aluminum tube cannot support the screwed-in fitting without spreading and splitting. So, EFII installs an Oetiker clamp as a hoop reinforcement. The Oetiker is a spring clamp normally used to squeeze elastomer hoses. It is intended to be crimped into place with a jaw tool to a specified crimp force (see the Oetiker catalog for details), compressing the elastomer hose, then springing back to a final diametrical dimension. Here, because it's crimped on a metal tube, there can be no significant compression, meaning springback doesn't allow the finished diameter to provide any significant clamp force. Thus the thread interface between the tube and the feed fitting is loose enough to wiggle and shake; it cannot be tightened. The fitting floats in the end of the tube, and depends on the injector's o-ring to absorb the relative motion.

Robert, is the above description correct, and those two broken mounts you mentioned....where did they break?
 
Last edited:
...Engineering practices specifically caution that hose assemblies need to have some slack in them so they don't become a "structural element" between two assemblies - In this case the hose is certainly a tension element, and thanks to the adel, it's a fairly effective compression element as well. When the heads move apart due to thermal or operating stresses, the hose is not going to stretch. The reaction is going to be bending through the 1/8NPT nipple.

The teflon hoses linking cylinders are not stiff hoses. Hard line is a "stiff hose".
We have had no issues what-so-ever with properly supported installs. The referenced injector failure was after many years of operation with an inadequately supported install.
"Support your fuel rails and anything else bolted to your engine -properly" is the lesson here. Robert

Both arguments have merit, although the parties are not talking about the same loads. Mike is discussing tension/compression due to fore and aft movement of the heads. Robert's adel clamp addition is an attempt to provide support in response to a vertical movement of the heads, the effect of torque reaction (block shake counter to prop rotation), or piston side loads within the cylinders.

The modulus of elasticity for teflon is only 0.5 Gpa, roughly 140x less stiff than alumimum, and fore/aft cylinder displacement would be very small. Given those factors I would tend to ignore the tension/compression concern.

However, Robert can't have to both ways. If we buy into the argument that the teflon hose is not stiff (and the above modulus says it's not), then the adel clamp cannot support the injector bodies. At best, the adel can only support some percentage of the hose mass. It's an improvement, but the feed fittings are still free to vibrate in the threads of the aluminum tube injector mount, and the injector mount itself is still cantilevered from an 1/8" NPT aluminum nipple.

Previously I asked Robert to confirm where the injector mounts had broken, and here's why. Although the cut-thread NPT nipple is being fingered as a potential failure point, I suspect the broken mounts pre-dated those nipples, and being installed on welded steel bungs, had to have been failures of the aluminum tube body.
 
Last edited:
Safety Issue

Two of the Reno RV racers installed SDS top injector mounts on short notice after a failure of one of the mounts being discussed here which resulted in a fuel fire (it was fortunately put out quickly on the ground).

We`re pleased to report that both planes were quickly re-tuned with the SDS hardware and injectors and made the races this year.
 
Great articles!

My FM200 is closed loop. I have a large meat servo with its own CPU (apparently running DOS) monitoring exhaust temperature, with additional aural and tactile inputs.

Seriously, both Bendix style constant flow injection and the current electronic offerings are open loop, and there is no reason for any significant quantity of "wasted fuel" with either kind.

As Ross said, the typical constant flow system with .028" restrictors doesn't meter as well as an EFI system at low flow rates. At idle, the constant flow delivery is more dribble than squirt. Atomization is less than ideal, so vaporization is poor, thus cycle to cycle variation tends to be worse. In cruise, fuel divider accuracy drops off below roughly 6 GPH. That's not a big deal for most of us, as we tend to cruise above that, even LOP. Smaller engines can substitute .022" restrictors for an improvement. The EFI will generally deliver a smaller GAMI spread at very small fuel flows, as Dave Anders has demonstrated.

Dave, you should find all the basics in these two articles:

https://www.danhorton.net/Articles/Bendix-and-Beyond (1).pdf

https://www.danhorton.net/Articles/Fuel and Fire.pdf

Dan,

I enjoyed both articles. It helped fill in a lot of blanks in my ignorance about fuel injection. I will try to make it to one of the AFP classes.

Thanks for all you do,

Jeff
 
Back
Top