What's new
Van's Air Force

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

Power system architecture for EFI

always hot wires

... I have been considering running an AWG 6 "fat wire" to the switches for the EFI bus. According to Nuckolls he feels it does not require fusing.

Where is that EFI bus located?

Bob Nuckols' Aeroelectric Connection: "As a rule of thumb we try to avoid long runs of always hot wire fused at more than 7A. This is a crash-safety issue. If your E-Bus really needs a feeder protected at more than 7A, consider adding a power relay..." (rev 12A says in Appendix Z Page Z-6 under Figure Z-32)

In the Aeroelectric List I found: "Always hot wires are either crew controlled (relay or contactor) -OR- protected at low levels on the order of 7A fuse or 5A breaker. If you are attaching wires to the battery and they have lengths greater than the 6-inch rule of thumb then the design goal for low magnitude protection applies. In other words, a battery bus is located next to the battery. Any other location makes it something else."

Also an always hot wire can get you in trouble while working on the aircraft if you don't disconnect the negative battery terminal first.

The S704-1 20A relay sold by B&C has coil current of only 0.1A.

Bob Nuclols puts a "current limiter" in the alternator B lead so if there is a hard short at the alternator the current limiter will open. A fat wire shorted to aluminum sheetmetal will reportedly clear itself by burning the aluminum away like EDM machining but shorted to something thicker like alternator or engine case can conduct hundreds of amps. http://www.aeroelectric.com/articles/anl/anlvsjjs.html

Note also that Bob adds a starter contactor close to the battery contactor so there will not be an always-hot fat wire running to the starter.

Just my 2 cents as a Nuckols believer... flame on!
 
Last edited:
If your E-Bus really needs a feeder protected at more than 7A, consider adding a power relay..." (rev 12A says in Appendix Z Page Z-6 under Figure Z-32)

I may not be understanding this sentence properly but an EFI E-Bus will always require a lot more than 7 amps feeding it- like triple that. Was this statement directed at EFI architecture specifically or other electrical systems like avionics?
 
I may not be understanding this sentence properly but an EFI E-Bus will always require a lot more than 7 amps feeding it- like triple that. Was this statement directed at EFI architecture specifically or other electrical systems like avionics?

It's a generic statement... doesn't matter what's at the other end of the wire. He's saying don't run 20A from the battery bus to a switch on the panel; instead have a cockpit controllable relay close (6" or less) to the battery bus like the latest Z-13. An earlier Z-13 had a 7A fuse and no relay. The bus and relay is at the battery and the switch on the panel. https://drive.google.com/open?id=19vu8KCJ8UNvPf6yP5PMRef2DWwaqTbOp
 
Last edited:
It's a generic statement... doesn't matter what's at the other end of the wire. He's saying don't run 20A from the battery bus to a switch on the panel; instead have a cockpit controllable relay close (6" or less) to the battery bus like the latest Z-13. An earlier Z-13 had a 7A fuse and no relay. The bus and relay is at the battery and the switch on the panel. https://drive.google.com/open?id=19vu8KCJ8UNvPf6yP5PMRef2DWwaqTbOp

I tried to follow his rule in my draft Z-14 modified for EFI. https://drive.google.com/open?id=1rQ0whUjYb5haWeZaqdncufqJo0o-iQuq

Generic, ok, got that.
 
I may not be understanding this sentence properly but an EFI E-Bus will always require a lot more than 7 amps feeding it- like triple that. Was this statement directed at EFI architecture specifically or other electrical systems like avionics?

Ross, what is the total load ...dual ECU, two coils, two fuel pumps, six injectors, all together?
 
A round figure on a 540 would be 20 amps with both (standard) pumps on.

Depends on fuel pressure, pumps used (we offer 3 different pumps), rpm, MAP, number of injectors and coils you're driving.
 
A round figure on a 540 would be 20 amps with both (standard) pumps on.

Depends on fuel pressure, pumps used (we offer 3 different pumps), rpm, MAP, number of injectors and coils you're driving.

Got it.

Returning to the very simple backup you described...

My RV6A also has no alternator connection to the backup battery- just a 30 amp ATO fuse and toggle switch to the engine bus. I isolate that bus from the primary battery and alternator by turning off the master.

Operating description here:

http://www.sdsefi.com/air45.htm

So the basic architecture looks like this?

55mgqg.jpg
 
That's correct Dan.

I must admit I haven't read this page for over a decade but it's still pretty much valid. The backup battery is now an 18 amp hour AGM one and I charge it the 1st of every month on the ground. Load test every annual. Voltage is verified prior to engine start, on my checklist.
 
Thanks, John. Yes, in my myopic quest for the SIMPLE power design, I neglected to address the always hot wires. (approx. 20A max load per Ross).

So the only way I can reliably keep panel switches for the two EFI power feeds is to have relays of some sort. This does add two relays and connections to the simple concept.

Is there any way to avoid the relays? Reading the Nuckols' comment, it appears ill advised. Still, they can be independently tested, and the power paths to the EFI bus are redundant.

And is there a "most" desirable location for the EFI bus?

I appreciate the critical comments. Really.
 
Much cleaner diagram(s), Dan. I have been considering running an AWG 6 "fat wire" to the switches for the EFI bus. According to Nuckolls he feels it does not require fusing. Eliminates another component and connections.

http://www.vansairforce.com/communi...4703&highlight=fat+wires+fuse+Nuckolls&page=2

Any complaints with that?

Can I take a stab at that? I think the other responses were good, but I don't think the basic philosophy is clear yet. The following is my interpretation of what Nuckolls is saying.

It's not a universal truth that 'fat' wires don't need fusing (protecting). If it's a very short run (say, 6" or so), the danger to the wire is so low that you can make a 'sure bet' that you can get away with no protection. But if it's an extended run, there are more opportunities for faults. Now there are various ways to protect, and a contactor (relay) is one legit way to protect. It just requires a human in the loop to trigger the protection. That's what we do with a battery & master contactor in the back, and the fat feeder running to the front of the plane.

For my E-dependent engine bus, I wanted the engine to be as independent from the rest of the plane *as possible*, as we're accustomed to with a magneto ignition. My power paths:

battery>fusible link >fat wire>high current switch>engine bus<>high current bus tie switch<>main a/c bus<fat wire<master contactor<battery.

I'm contemplating a high current diode wired across the bus tie switch, pointing toward the engine bus. This will allow uninterrupted power to the engine bus if the engine power switch fails. If the master contactor fails, the airframe would go dark (EFIS has its own battery backup) until the bus tie switch is closed.

The bus tie switch provides one-switch redundant power feed to both the engine bus (from the main bus), and to the main bus (from the engine bus).

This is a two identical alternators, single battery system.

FWIW...

Charlie
 
SNIP...

This is a two identical alternators, single battery system.

FWIW...

Charlie

I offer that assuming you have not abused it, the battery(s) is the most reliable element of your electrical power system. That reliability however ends at the battery terminals.

I also offer that the second alternator in a single battery system protects from only one risk, the loss of the primary alternator. While this may be the single most likely risk, it is not the majority of risk. Twin airplanes with two alternators have lost all electrical power (both engines still running and both alternators still working). There are lessons to be learned from these failures.

This translates to designs with two batteries and a single alternator provide opportunity for a far more robust power distribution design when compared to two alternators and a single battery designs. Add a B&C 20 amp standby alternator on the vacuum pad to take the design to the next level. Just be very cautious on how elements are connected and buss faults isolated.

Carl
 
So, given all the responses......(very informative)

Ross....adapting Charlie's power path (using two batteries):

battery1>fuse>fat wire>high current switch>engine bus<>high current switch<fat wire<fuse<battery2.

How much would I need to rate the fuse(s) and switch(s) for an EFI system? I am thinking 30amps? (I realize the fuse protects the wire, but let's assume we go with fat wire for safety).

I think the rub here is whether the power wire to the EFI bus is switched/fused/relay controlled. Welcome any additional thoughts.
 
No vac pad on this engine; not a Lyc. I'm also not limited by the limited capacity of a low output 'backup' alternator, so my bus systems can be simpler. (either alternator can carry the full load of the a/c.) In addition, I'm able to carry only an additional ~10 lbs of alternator instead of 15-20 lbs of extra battery, and after one alternator failure, I still have unlimited electrons instead of those left in the batteries. Fuel is still the limiting factor of any single flight, instead of battery capacity.

Your point is well taken that the area around the battery can potentially be a single point of failure. But one alternator directly feeds the battery; the other feeds the load side of the master contactor. Even if the 'battery area' develops a fault, one of the alternators will still feed the system. (With a long background in electronics, I'm not in the camp that thinks an alternator will go insane without a battery attached.)

And, as you point out, even twins with redundant everything (not just alternators) can and have gone dark. It's not hard to make the jump to bad (certified) architecture as the cause, since even operator error should have a really hard time making an entire twin go dark.

Ultimately, we have to pick our poison. I've been tempted to add a 2nd battery, but the added weight, and more importantly, the added complexity, has driven me to this compromise, rather than the 1 alt/2bat compromise, or the much higher complexity of the 2/2 system.

Charlie
 
So, given all the responses......(very informative)

Ross....adapting Charlie's power path (using two batteries):

battery1>fuse>fat wire>high current switch>engine bus<>high current switch<fat wire<fuse<battery2.

How much would I need to rate the fuse(s) and switch(s) for an EFI system? I am thinking 30amps? (I realize the fuse protects the wire, but let's assume we go with fat wire for safety).

I think the rub here is whether the power wire to the EFI bus is switched/fused/relay controlled. Welcome any additional thoughts.

My airplane uses a 30 amp ATO fuse and a 25 amp switch. If you intend to run both pumps at the same time, you should use a switch (or relay or contactor if you prefer) rated for 30 amps. The surge current for the pumps is pretty high, being driven by DC motors.

In our experience, 2 batteries and 1 alternator is preferable to 2 alternators and one battery. We see many people going 2 batteries and 2 alternators (one driven off the accessory pads) on RVs these days as they may have more room with one or more of the mags gone.

My RV10 was fitted with dual batteries and dual alternators as the airports are further apart up here than in the US.

Below is a typical RV10 installation with B&C backup alternator that more people are considering or fitting these days.

 
Hi Ross,

I tried to detail my thought process that led me to the 2alt/1bat configuration. However, I'm not flying and you've been flying auto-style engine control for quite a while. Can you summarize or refresh my memory as to why the 1alt/2bat system is better? I know we're all trying for 'best', and while I've tried to think of all the variables, I don't doubt that I've missed a few.

Thanks,

Charlie
 
Batteries can't overvoltage and the AGM ones we've used hold a good charge for many months (designed for UPS applications). They have been completely reliable. No moving parts in a battery. They are not expensive either.

Disadvantage is more weight (AGM anyway) and the AH capacity is your limiting factor for duration compared to an alternator.

If you do lithium, you can save weight but are subject to their unique operating characteristics which might not inspire total confidence as a backup power source. The AGMs will actually fire coils and run pumps down to 9V or so. Not sure on the lithium but I know one guy who did not have such great luck with one as a backup. You'd want to test one under load and see where it signs off.

2nd alternator needs a regulator and preferably a crowbar. Starts getting expensive there.

If I was doing a -10 again, especially with a lot of glass and for IFR, I'd install 2 batteries and 2 alternators again. The pad driven alternators are pretty light for Lycs.

With one battery, if it goes wonky (say shorted cell or something like that) and you have to disconnect it, how will your alternator regulators reference voltage to control output? Should test for that too.

Test all this stuff on the ground and see what it does. You don't want to find out in a real emergency that something doesn't work the same as you thought it would- especially on a dark and stormy night IFR.
 
Last edited:
My 6-year-old PC680 failed open with an internal broken weld. You can read about it here:
http://www.matronics.com/forums/viewtopic.php?p=458468&sid=e9e8d09021e1ecd6ac4fa1960b22b54f
At first it was an intermittent condition. The first sign of a problem happened when the throttle was closed for landing. The EFIS (no backup battery) rebooted because the Rotax alternator does not generate at idle and the battery internal weld opened up.
Like Charlie said, the alternator will not go insane without a battery. But the battery is needed to stabilize the alternator. I have conducted an experiment by disconnecting the battery while flying. Without a battery, the Rotax electrical system voltage varied plus or minus 3/4 of a volt or 1.5 volts total.
Based on this experiment, the Rotax charging system needs a battery for stability. While the electrical loads worked OK, high and low voltage alarms can be annoying. A pilot would not want that distraction while flying IFR.
Although rare, batteries can fail suddenly without warning.
 
Hi Ross; I appreciate your thoughts.

If I were flying IFR with an E-dependent engine, I'd probably go whole hog with 2/2, as well. And it sounds like we agree on the quality/reliability of AGM batteries (and the not-yet-provens of lithium), too.

On the subject of alternator choice, I don't fear controllable IR alternators, and OV modules are very inexpensive these days. I have less than $250 in a pair including OV modules, so, less money than one pad mounted backup alt (which I don't want anyway; each of mine can support the whole plane). Alternator regulators don't look to the battery the battery for their voltage reference; otherwise, how would they know what voltage to supply a battery in some state of discharge? Any regulator will have its own internal voltage reference; typically involving a zener diode.

And believe me, I will be testing on the ground. :)

With what I've seen so far, it really does seem to come down to an individual's comfort zone. I will continue trying to maintain an open mind if/when more more info becomes available.

Thanks again,

Charlie
 
As I always say guys, use what you want and test it well on the ground first.

Don't assume because one brand or type does something connected in one way that it will behave the same way when combined with other parts or in other ways.

It would be instructional for folks to shoot some video of their battery/ regulator/ alternator experiments and share the links.

As for the Odyssey batteries, yes we've read of a few failures here on VAF but I'm thinking there are thousands of these installed for decades now. Maybe we should start a poll and see what the real percentage failure rate might be. I got 11 years out of my first (still cranking the engine and passing the annual load test), on the 5th year with my second. Mine are FW aft, not exposed to temperatures they were not designed for and not sitting on a trickle charger like some folks seem to be doing. Look at the bottom of this chart for operating temp specs. http://www.odysseybattery.com/extreme_battery_specs.aspx

Anything can fail but I'll bet a dime that Odyssey batteries are 100X more reliable than one brand of alternator commonly fitted to RVs...

You'll only find out how smart you were at the end of 10 or 20 years of flying. Share the good and the bad experiences so we can all learn and improve.

BTW, I use a Powersonic 18 AH battery for backup. http://www.power-sonic.com/images/powersonic/sla_batteries/ps_psg_series/12volt/PS12180NB.pdf
 
Last edited:
The AGMs will actually fire coils and run pumps down to 9V or so. Not sure on the lithium but I know one guy who did not have such great luck with one as a backup. You'd want to test one under load and see where it signs off.

What voltage you get down to is largely irrelevant. LiFePo batteries don't have a constant voltage drop with use like a lead-acid style one will; their output voltage will stay relatively flat until reaching their capacity limit, and then they drop fast.

lifepo4.jpg


Saying "my AGM will fire the coil down to 9V" doesn't mean you get more use out of it; it just means the battery behaves differently. EarthX does note in their manual that "The low charge level is very different from a lead acid battery, for a lithium battery is completely drained at approximately 11.5V". They recommend the "yellow" range (for alternator offline) at 13.5V and the red "low charge level" at 12.6V. That puts the "low capacity warning" about at the "knee" in the discharge graphs, and tells you that you've used ~80-90% of the capacity.


EarthX does publish time vs. voltage curves for varying discharge rates. They're very useful for planning purposes (use some margin, of course). As you say, you need to run a test to be sure.
 
Would be better to post a discharge curve of a 12V lithium of a usual capacity we fit to our aircraft for comparison purposes and have the temperature and amps curves like a proper chart has, not something generic like this.

I'd disagree that the voltage is irrelevant. The device you're powering may well care.
 


Better to use real discharge charts to compare products, not stuff from the marketing department.

You'll see the curves on this AGM look nothing like the one depicting a lead acid one in the lithium chart.
 
Last edited:
What voltage you get down to is largely irrelevant...

...Saying "my AGM will fire the coil down to 9V" doesn't mean you get more use out of it; it just means the battery behaves differently...

However, the Pb battery will give up its life if you need to drain it to nothing for some critical reason. I understand that the lithium chemistry will shut off to self protect. I'd hate to be in a situation where I'm at 11.4v one minute (thinking I'm fat with power) and then the battery shuts down two minutes later to "save itself".


I understand that correct sizing of any chemistry battery will give you the time needed before using devices go dark, but the unfamiliar behavior with the steep knee and auto shutdown is a concern. Since we are talking about use as an emergency power source, the last thing I need in a stressful time is adding to that stress wondering what the **** is the battery going to do next.
 
Further to the battery discussion. As Michael says, you better test your battery solution, no matter what the chemistry is, down to where the components you're driving sign off. If the battery shuts down before that level, you should rethink or resize your battery choice. The AGM is a known quantity. Different lithium batteries may have their protection circuits step in when you need to keep the EFI/EI stuff running.

Be aware that we're driving several devices that do not draw constant current such as the fuel pumps. Injectors and coils especially have high inrush current for 1-3 milliseconds. When the battery is well discharged, they may not be able to supply this current to operate these components.

Some components like certain MOSFETs may require charge pump circuits which become unstable at low voltages. I could go on.

What looks like will be a good idea on the surface, may not be in practice. If you stray from what we've used and tested to work over many years, please actually test it thoroughly before you fly with your solution. The demonstrated reliability we've established over 20+ years and a half million flight hours comes from building on a proven recipe. Other solutions can work, but can only be validated by actual use, not speculation or theory.

I've built plenty of things that I thought should have worked but didn't. Until you have a complete understanding of the problem, you may miss a few critical elements which lead to failure. That's how we learn, but in this instance, we want to be sure our solution gives us reliable running for the time required if we have a charging system issue in the air.
 
Last edited:
However, the Pb battery will give up its life if you need to drain it to nothing for some critical reason. I understand that the lithium chemistry will shut off to self protect. I'd hate to be in a situation where I'm at 11.4v one minute (thinking I'm fat with power) and then the battery shuts down two minutes later to "save itself".

Appears to be true for an EarthX. However, I assume not all lithium iron batteries have a built in disconnect.

Read Ross' account of failing an alternator way back when, and not catching it until electrical devices started doing wonky things. The declining lead-acid voltage made the shutdown progressive, all the way through engine misfire, then failure. In other words, there was some warning time. Is it a big deal? Look at the curves Ross posted above. At 18 amps, the drop from 11 to 9 volts looks to be about 6 minutes. I'd guess 11 volts would be about where a pilot should maybe forget the target airport and look for an open spot.

http://www.sdsefi.com/rv12.htm

Scroll down to the 2/7/05 date.

BTW, Ross has posted the account as long as I can remember, without regard for personal feelings or his business. It's a fine aviation tradition, intended to help keep others safe by passing along lessons learned the hard way. My compliments sir.

There are valuable lessons to be learned in two more recent incidents, both of which appear to entirely relevant to this discussion of EFI power architecture. Hopefully the owners would post, like Ross. If you know them personally, please ask.

https://app.ntsb.gov/pdfgenerator/R...D=20171031X10251&AKey=1&RType=Prelim&IType=LA

https://app.ntsb.gov/pdfgenerator/R...D=20161129X71536&AKey=1&RType=Prelim&IType=LA
 
Last edited:
Appears to be true for an EarthX. However, I assume not all lithium iron batteries have a built in disconnect...

I'm not sure either, but it is exactly this uncertainty that gives me comfort in the well known Pb chemistry. I'm not at all against the newer chemistries - I think they are one of the more exciting things to come to this industry - but I'm not comfortable with their apparent behavior in a deep discharge, long duration emergency scenario. Not yet, anyway.
 
The battery issues currently being discussed are what drove me to the 2 alt/1 bat choice for my installation. Not claiming to be right, or the only solution, but if your one alt fails, you are instantly in the realm of 'known unknown'. Any battery, regardless of chemistry, will have X energy storage capacity *when new*, and that capacity begins to decline at <?> rate as soon as it's put in service. Did my alt fail a day after the new battery was installed? A year? Just shy of the 2nd year's annual capacity check? Will the capacity decline in a known linear fashion, or could it decline exponentially?

Engineers' and designers' view of batteries, in both traditional automotive and aviation fields, is that batteries are *starting* batteries. The system is never intended to operate on battery power; the alternator is expected to supply all operational electrical energy.

That, and operational simplicity (with weight as a minor driver) took me to dual equal sized alternators. Loss of one alternator still leaves the system functioning as designed. Loss of the battery just leaves the system with a little more 'ripple' in the DC output of the alternator. Only with failure of both alternators in one flight (can anyone figure the odds of that?) would I get into the realm of 'known unknowns'.

Again, not saying other choices are wrong. Just what got me to my current philosophy in electrical system design.

Charlie
 
A large enough lithium battery will do the job just fine and certainly weigh less than a comparable AGM but you may have to do you own tests to pick the correct one.

With regards to my aircraft, engine electronics were drawing around 13.5 amps but with the other avionics and engine scavenge pump running, probably over 20. Assuming the powersonic chart is similar to the PC680 I was using as primary (only) battery, I had at most about 20 minutes duration until 9 volts. Factor in capacity degradation and possibly temperature and you'll likely have less than that. The charts clearly show the importance of shedding load and how nice it is to have a second battery when things go south.

On the alternator topic, others here have posted that alternators work just fine with the battery disconnected. That has not been my experience which dates back to the late '70s when I was rebuilding my own Delco units in Corvairs.

This morning I decided to do a test on a Denso IR unit off my EG33. I spun it up with the battery disconnected from the B terminal. Voltage continued to climb with rpm, exceeding 22 volts at the maximum I could turn it to with my die grinder which was only an estimated 2000-2500 rpm. I'll shoot some video when I get some time. Your alternator may perform differently but please don't think that ALL alternators are going to stay around 14V with the battery disconnected. Test your's.
 
Engineers' and designers' view of batteries, in both traditional automotive and aviation fields, is that batteries are *starting* batteries. The system is never intended to operate on battery power; the alternator is expected to supply all operational electrical energy.

Charlie

May not necessarily apply to backup batteries. The Powersonic AGMs are not designed for starter motor loads, rather relatively low current draws for extended periods which is what we need in this case.
 
My statement:Originally Posted by rv7charlie View Post
Engineers' and designers' view of batteries, in both traditional automotive and aviation fields, is that batteries are *starting* batteries. The system is never intended to operate on battery power; the alternator is expected to supply all operational electrical energy.

Charlie

May not necessarily apply to backup batteries. The Powersonic AGMs are not designed for starter motor loads, rather relatively low current draws for extended periods which is what we need in this case.
I'm well aware of different design goals for different batteries.

But I wasn't talking about batteries. I was talking about the designer's view of the automotive and aircraft electrical *systems*. *In the system*, the battery exists for starting. The alternator exists to supply electrical energy to the system while it operates. Replacing the energy used to start the engine is a very minute percentage of the energy produced by the alternator, and used by the system, during operation.

When you ran the test with your Denso, did you ensure that the alternator was able to sense the voltage on the B-lead? Did it still have any electrical loads on the B-lead, or was it completely disconnected?
 
I was stumbling around on one of my back-and-forth-to-the-house thumb drives, and ran across a diagram I worked up for another project some months ago. What the heck, let's have some fun.

Much of the EFI bus architecture posted so far has required diodes for isolation, or switched recovery (one source switched to the the EFI bus at a time, with pilot intervention to restore power). I've suggested auto-switching, and yes, it does add some additional components. However, it can result in a system requiring no (or very rare) pilot intervention, and no diode voltage loss.

This particular example requires two contactors and a relay (left), or three relays (right). I'd probably use 30~50 amp relays, as they wouldn't be expected to switch under significant load very often. There are two control switches, both low amp. They serve the traditional ON-OFF function, and provide a test function. It normally operates on the left side supply, and auto-switches to the right side.

209ncl4.jpg


Operation is pretty much like a pair of mag switches. I ran an initial wire-by-wire failure analysis. It was done quickly, and I reserve the right to review again, but first pass was pretty good. Found a fault possibility requiring a switch flip to correct. I'll give you a clue; it's an open. Go ahead, find it ;)

Dual independent sources are the same as post 147. This just changes how the EFI bus is supplied, meaning no diodes, and the switches don't need to pass 20 amps. Which is better? You decide.

hu0co9.jpg
 
Last edited:
If left contactor coil fails open, the engine stops.
The panicked pilot can not remember how he wired it long ago.
Flipping left switch off is not intuitive for another pilot who is not the builder.
 
Very good Joe!

Actually that would be two; the other possible open requiring a switch flip would be the short link between the left contactor (or relay) and the bus. My cursory failure list had not yet reached internal component fails, just connections, so good catch.
 
I?m curious as to why there seems to be an aversion to the use of diodes. I use them on the output of my vacuum pad alternator so a fault on one avionics buss does not take down the other. Easy enough to design and large current Schottky diodes are cheap.

Carl
 
A round figure on a 540 would be 20 amps with both (standard) pumps on.

Depends on fuel pressure, pumps used (we offer 3 different pumps), rpm, MAP, number of injectors and coils you're driving.

That's a significant current draw, and of course that's just to keep the engine running and doesn't include any avionics.

On a single alt system if the alt quit you'd better be looking for a place to land like NOW as the battery is going to deplete rather quickly.

In my experience the average pilot doesn't usually notice the alt quit until things start to fail, in this case it would include the engine.
 
I’m curious as to why there seems to be an aversion to the use of diodes.

I assume voltage drop, energy loss (heating), and concern about failing open, none of which is an absolute reason to use or not use them. As noted previously, nothing is perfect. Design choice is about balancing the compromises.

For example, in theory both a diode and the relay coil discussed above could fail open. The first thing to consider is not the likelihood of the component failure, but what happens if it does. Let's assume either open (diode or relay coil) has the same result; the pilot must flip a switch. The result is equal and there is a recovery, so move on to the next compromise on the list. Let's pick heating. The diode wastes some energy, while the relay wastes little....advantage relay. Voltage drop? Yes for the diode, no for the relay...advantage relay.

Comparing my own architecture examples (post 147 vs auto switching), the failed relay would require pilot intervention, while a failed diode would not...advantage diode. Switch failures offer roughly the same result. The overall design balance swings toward diode.

After looking for critical failures, we can move to likelihood of component failure. Relays and contactors are designed to switch significant loads, toggle switches less so. Frankly, that's about the only widely accepted reliability conclusion, and even that appears to be more belief than data. Still, it appears to move the balance back toward relay switching.

See what I mean about carefully considered compromise? A "correct" answer is unlikely. The best you can do is to carefully analyze and make a reasoned choice, not a gut call, at least until there is nothing else to go on. Key point is that we want to avoid critical failures (a result), the ones which put airplanes in the dirt.

Me? I'm fine with diodes given small loads and benign failure results, and I fly one in the power supply for an EI drawing about 1A. I'm not a big fan given large loads, but obviously they can work.
 
Last edited:
That's a significant current draw, and of course that's just to keep the engine running and doesn't include any avionics.

On a single alt system if the alt quit you'd better be looking for a place to land like NOW as the battery is going to deplete rather quickly.

In my experience the average pilot doesn't usually notice the alt quit until things start to fail, in this case it would include the engine.

Knocks down a lot running on one pump. Still, the much higher draw of EFI/EI (as compared to just EI) is why I'm for dual alternator, dual battery, and access to both batteries with the master off.
 
Last edited:
That's a significant current draw, and of course that's just to keep the engine running and doesn't include any avionics.

On a single alt system if the alt quit you'd better be looking for a place to land like NOW as the battery is going to deplete rather quickly.

In my experience the average pilot doesn't usually notice the alt quit until things start to fail, in this case it would include the engine.

Keep in mind that 20 amps is with 2 pumps running - a non standard condition. Dump one of those pumps as you would in normal flight and you're down to 13"ish" amps.

And yes, alternator failure in legacy airplanes has often gone unnoticed. But it's easy enough today to introduce some very obvious warning stimuli for the pilot. Some will still ignore them and crash, yes. But some should not be flying airplanes they built in their garage either.
 
Last edited:
That's a significant current draw, and of course that's just to keep the engine running and doesn't include any avionics.

On a single alt system if the alt quit you'd better be looking for a place to land like NOW as the battery is going to deplete rather quickly.

In my experience the average pilot doesn't usually notice the alt quit until things start to fail, in this case it would include the engine.

We recommend either a backup battery or alternator in ALL cases of using EFI/ EI. Two 18 AH batteries will give you around 1.2-1.5 hours of flight time running one pump (4 cylinder engine). You can gain a bit more reducing rpm and shutting off one coil pack.

If this is not sufficient for your missions, either a larger backup battery is needed or a second alternator.

See Dave Anders' Shorai dual battery setup on his RV-4: http://www.vansairforce.com/community/showthread.php?t=139328&page=15

We also recommend aural voltage warning rather than just visual. As you say, visual indications can be missed as I did in my incident.
 
I assume voltage drop, energy loss (heating), and concern about failing open, none of which is an absolute reason to use or not use them. As noted previously, nothing is perfect. Design choice is about balancing the compromises.

For example, in theory both a diode and the relay coil discussed above could fail open. The first thing to consider is not the likelihood of the component failure, but what happens if it does. Let's assume either open (diode or relay coil) has the same result; the pilot must flip a switch. The result is equal and there is a recovery, so move on to the next compromise on the list. Let's pick heating. The diode wastes some energy, while the relay wastes little....advantage relay. Voltage drop? Yes for the diode, no for the relay...advantage relay.

Comparing my own architecture examples (post 147 vs auto switching), the failed relay would require pilot intervention, while a failed diode would not...advantage diode. Switch failures offer roughly the same result. The overall design balance swings toward diode.UIKeyInputDownArrow

After looking for critical failures, we can move to likelihood of component failure. Relays and contactors are designed to switch significant loads, toggle switches less so. Frankly, that's about the only widely accepted reliability conclusion, and even that appears to be more belief than data. Still, it appears to move the balance back toward relay switching.

See what I mean about carefully considered compromise? A "correct" answer is unlikely. The best you can do is to carefully analyze and make a reasoned choice, not a gut call, at least until there is nothing else to go on. Key point is that we want to avoid critical failures (a result), the ones which put airplanes in the dirt.

Me? I'm fine with diodes given small loads and benign failure results, and I fly one in the power supply for an EI drawing about 1A. I'm not a big fan given large loads, but obviously they can work.

Dan,

Perfectly logical but I offer that keeping the engine running and/or continued IFR flight after a component failure, without pilot immediate action, is a trump card.

The one takeaway from this discussion for me is what I consider a huge current draw of EFII systems. This aspect should be included in the builder?s decision process.

Carl
 
I assume voltage drop, energy loss (heating), and concern about failing open, none of which is an absolute reason to use or not use them. As noted previously, nothing is perfect. Design choice is about balancing the compromises.

For example, in theory both a diode and the relay coil discussed above could fail open. The first thing to consider is not the likelihood of the component failure, but what happens if it does. Let's assume either open (diode or relay coil) has the same result; the pilot must flip a switch. The result is equal and there is a recovery, so move on to the next compromise on the list. Let's pick heating. The diode wastes some energy, while the relay wastes little....advantage relay. Voltage drop? Yes for the diode, no for the relay...advantage relay.

Comparing my own architecture examples (post 147 vs auto switching), the failed relay would require pilot intervention, while a failed diode would not...advantage diode. Switch failures offer roughly the same result. The overall design balance swings toward diode.

After looking for critical failures, we can move to likelihood of component failure. Relays and contactors are designed to switch significant loads, toggle switches less so. Frankly, that's about the only widely accepted reliability conclusion, and even that appears to be more belief than data. Still, it appears to move the balance back toward relay switching.

See what I mean about carefully considered compromise? A "correct" answer is unlikely. The best you can do is to carefully analyze and make a reasoned choice, not a gut call, at least until there is nothing else to go on. Key point is that we want to avoid critical failures (a result), the ones which put airplanes in the dirt.

Me? I'm fine with diodes given small loads and benign failure results, and I fly one in the power supply for an EI drawing about 1A. I'm not a big fan given large loads, but obviously they can work.

Good post Dan. Agree with your conclusions here.

One more point, relays require multiple connections, switches only 2. #1 "failure" we see in thousands of systems and millions of hours are connections, not relays themselves and not switches, especially if switches are only activated when a primary system goes down, which is almost never.

Many reviews are based on MTBF. Of course, all things can fail but if we have a long history of something which has had zero failures we have more confidence in that device than ones which have a higher incidence of failure. This again points to experience in the field. People without long experience have no useful MTBF data to support their views.

Anything drawing more than 20 amps and especially being switched frequently, I'd probably use a quality, brand name relay wired with proper non-insulated terminals and glue infused heat shrink. You've made it as good as you can at that point. You can never prevent all failures, only minimize and mitigate with good design and workmanship.

At the same time, we do want to evaluate what will happen if that super reliable part actually fails as you say and that is the main reason why many people want a backup power supply. We don't care much what fails in the primary system but we want to make sure that no failure there will also take down the backup. You must be able to fully isolate batteries and alternators between the 2 systems in all cases.

Hold current on a relay is less than a contactor, another advantage IMO.
 
Last edited:
You can never prevent all failures, only minimize and mitigate with good design and workmanship.

This is what the last 19 pages boil down to.

To some folks good design means minimum parts and potential failure points.

To others good design means double or triple redundancy with all the included parts and potential failure points.

Some want automatic backup function, others want to be in control of the backups.

Bottom line would seem to be -----educate yourself, and build what suits your mission, and your comfort level.

And, your pocket book;)
 
This is what the last 19 pages boil down to.

To some folks good design means minimum parts and potential failure points.

To others good design means double or triple redundancy with all the included parts and potential failure points.

Some want automatic backup function, others want to be in control of the backups.

Bottom line would seem to be -----educate yourself, and build what suits your mission, and your comfort level.

And, your pocket book;)

Absolutely right. One solution does not fit for all folks. This thread presents many different views and points. Highly educational though it will take some time to sort through what you want to do.
 
I have been following this thread since the beginning and was waiting for the obvious to filter out, but new ideas keep coming and a few older ones will not die. It is amazing how the thread has kept on topic and not gone negative. Testament to the main characters on the thread, the knowledge and experience has been fun to follow.

Not to be too critical, but the above circuit diagram (SORRY DAN) leaves fewer pilot choices and more questions of reliability than two switches and two diodes.

1. I prefer no full-time pass through diodes, but there is a place for them. The earlier diagrams with standard switches, diodes, and two batteries is simple and pilot friendly. Both can be left on, both batteries are isolated and can be monitored, one can be left in reserve if wanted, simple runup test. Many wins here.
2. The diagram never allows both batteries to run the EFI bus at the same time? The left has to be off to allow the right to work. If all else fails it would be nice to just leave them both on until depleted. You can still run one at a time for management or to gain a little recovery if desired. How much is arguable without testing.
3. If one is desiring auto switching is this the answer? At what voltage point will the relays power off? Many questions. Neat idea though.
4. With the new EFIS/engine monitors set up properly both batteries can be monitored at all times with near instant warning of an alternator failure and the lower voltage that goes with it. Vigilant engine gauge scanning is becoming a thing of the past. With both EFI bus switches on just: EFIS warning, verify which battery, isolate, cross feed if necessary, manage loads, done. Easy with a well thought out system. Will require switches, more than a couple.
5. Relays are not the greatest thing to come along that many here make them out to be. The main benefits of a relay are to control large current with small switches and to remotely control switching. These small 30 amp relays are no more reliable with their small contacts and push on terminals than a Carling switch. I will argue less reliable than a Honeywell TL switch with screw on terminals. They have their place, but I feel are increasingly becoming over used. I realize there are larger relays and contactors. Another discussion.
6. It has been stated on VAF in the past that some have never seen a relay failure. I have replaced "hundreds" of relays on cars (no exaggeration). Most on fuel pump, blower motor, AC clutch, and EFI circuits (in that order). There are several failure modes that will likely be seen here in the future, with relays now running fuel pumps continuously throughout the flight. Fuel pump/EFI battery bus relay maintenance will require careful periodic inspection and possibly require preventive replacement at yet to be determined intervals. We will see.
7. I do see the problems with long runs of the electrically hot 14-12 gauge cable required to run a bus from a rear mounted battery, the ability to properly protect them, and crash worthiness. Crashworthiness is the only big negative to manual switches powering the "EFI Bus" with a rear mounted battery, the only good reason for a EFI bus relay in the back or running power after the master relay. If the battery is anywhere up front I lean to no bus power supply relays.
8. Many here have been choosing fancy paddle switches and will go to great lengths to minimize the number of switches required. Installing electronic fuel injection is going to require more switches, better switches, and added complexity. This is not the place to overly simplify for the sake of lowering switch count. It is more complex.

If I was building today, electronic fuel injection would be something I would seriously consider, but I see many pitfalls and little being gained by its use. Are the rewards that great? Change for the sake of change? Still think it is pretty cool if you know what you are getting into.

I first installed fuel injected engines in my 72 2WD Blazer and wife's 72 Monte Carlo back in 1990. I have run an automotive shop for over 30 years and understand the technology and pitfalls. It is not that different, just much higher consequences if a failure occurs.

A few things to consider and just a start. Much, much more to discus.

George Meketa, RV8
 
Interesting and thought provoking post plus another perspective coming from long experience George. Thanks for posting.

Your experience doesn't mesh with my similar background in auto repair and maintenance on mainly Japanese products going back to the late '70s. Can I ask what brands you primarily work on, especially with all the relay replacements? I can't remember the last time I replaced a bad relay on a Toyota or Nissan. I can't ever recall in fact, even on my multiple 20+ year old winter beaters with many thousands of cycles on headlight and blower fan relays.

Every one of those cars also used push on connectors. Never saw a problem with those but I live in a dry climate. Things can be different in Florida for instance where long term corrosion is more prevalent on electrical connections.

I hope in aircraft, folks are using the kind of spade connectors that you basically can't pull off again without pliers.

Returning to the main topic in this thread, I had a long chat with my friend Rusty Crawford today who is the highest time Lycoming SDS user (around 1850 flight hours to date, some of that on Subaru). He said he'd forward his layout for review and posting here. Rusty does a lot of night and IFR flight up high. I consider him one of the smartest folks I've ever met. He's an electronics and mechanical guy by background and occupation so his thoughts carry a lot of weight with me. Will be well worth a look IMO. Rusty has helped us with beta testing and tons of good ideas to implement into SDS over the years. We specifically discussed his ability to continue IFR flight with the primary alternator down. Who wants to descend into crappy weather below when you can continue on to your original destination?
 
Last edited:
Perfectly logical but I offer that keeping the engine running and/or continued IFR flight after a component failure, without pilot immediate action, is a trump card.

In this I'm not hard to convince.

http://www.vansairforce.com/community/showpost.php?p=1260818&postcount=106

The auto-switched bus supply is a design exercise, an exploration to see if diodes and high current toggle switches can be eliminated. Only after a system is committed to paper can each wire and component be examined for the kind of critical faults which would end powered flight.

Allow an example. Ross made the point that he sees a lot of connection failures, and something like the auto-switched system has a lot of connections, thus the system is questionable. Here's reality...in deciding on an architecture, the number of connections doesn't really matter. What matters is the result of any possible connection failure. A bad crimp or a loose fast-on or a bad relay contact is simply an open. So put the system on paper, and work through it, making a written list for every wire. For each, what is the result of an open? What is the result of a short? Does any result make the fan go quiet? Or are all possible failures benign?

It's real hard to predict component reliability, at least at a high level of accuracy. So don't. Instead concentrate on designing for fault tolerance.
 
Last edited:
Slight deviation from the topic, but I see several references made throught this thread to separate switches for the various EFI components (pumps, coils, brain, etc). I dont see the logic in this, wondering if I'm missing something. In my mind, the engine can't run without the pump, brain and coils lit up, so why more than one switch for the engine control function?
 
Slight deviation from the topic, but I see several references made throught this thread to separate switches for the various EFI components (pumps, coils, brain, etc). I dont see the logic in this, wondering if I'm missing something. In my mind, the engine can't run without the pump, brain and coils lit up, so why more than one switch for the engine control function?

Power downstream from the EFI/EI bus? I recall Ross saying most of that was pre-wired, I assume per this diagram:

http://www.sdsefi.com/dualecu4.pdf

Set aside the two power supply switches, and what's left are two coil switches, two pump switches, and a LOP switch. A sixth, a switch to route injector grounds via relays, is not shown, as discussed in previous posts. ECUs and injector power supplies are not switched.

So how did you wire yours?
 
Last edited:
why more than one switch for the engine
If that one switch fails, the engine quits. On the other hand, if there are two
electrical supply paths and one fails, there is a back up path.
There are two fuel pumps. If one fails, there is a backup. Same for ignition
and etc. The objective is to keep the engine running no matter what single item fails.
 
Back
Top